modelId
string | author
string | last_modified
timestamp[us, tz=UTC] | downloads
int64 | likes
int64 | library_name
string | tags
list | pipeline_tag
string | createdAt
timestamp[us, tz=UTC] | card
string |
|---|---|---|---|---|---|---|---|---|---|
responsibility-framing/predict-perception-bert-blame-none
|
responsibility-framing
| 2022-03-10T15:59:10Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T15:54:27Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: predict-perception-bert-blame-none
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# predict-perception-bert-blame-none
This model is a fine-tuned version of [dbmdz/bert-base-italian-xxl-cased](https://huggingface.co/dbmdz/bert-base-italian-xxl-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8646
- Rmse: 1.1072
- Rmse Blame::a Nessuno: 1.1072
- Mae: 0.8721
- Mae Blame::a Nessuno: 0.8721
- R2: 0.3083
- R2 Blame::a Nessuno: 0.3083
- Cos: 0.5652
- Pair: 0.0
- Rank: 0.5
- Neighbors: 0.5070
- Rsa: nan
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 20
- eval_batch_size: 8
- seed: 1996
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse | Rmse Blame::a Nessuno | Mae | Mae Blame::a Nessuno | R2 | R2 Blame::a Nessuno | Cos | Pair | Rank | Neighbors | Rsa |
|:-------------:|:-----:|:----:|:---------------:|:------:|:---------------------:|:------:|:--------------------:|:-------:|:-------------------:|:-------:|:----:|:----:|:---------:|:---:|
| 1.007 | 1.0 | 15 | 1.2585 | 1.3358 | 1.3358 | 1.1752 | 1.1752 | -0.0068 | -0.0068 | -0.0435 | 0.0 | 0.5 | 0.2970 | nan |
| 0.927 | 2.0 | 30 | 1.1310 | 1.2663 | 1.2663 | 1.0633 | 1.0633 | 0.0952 | 0.0952 | 0.4783 | 0.0 | 0.5 | 0.4012 | nan |
| 0.8376 | 3.0 | 45 | 1.0603 | 1.2261 | 1.2261 | 1.0574 | 1.0574 | 0.1518 | 0.1518 | 0.1304 | 0.0 | 0.5 | 0.2970 | nan |
| 0.7154 | 4.0 | 60 | 0.8347 | 1.0879 | 1.0879 | 0.8854 | 0.8854 | 0.3323 | 0.3323 | 0.6522 | 0.0 | 0.5 | 0.5209 | nan |
| 0.5766 | 5.0 | 75 | 0.7426 | 1.0261 | 1.0261 | 0.8340 | 0.8340 | 0.4059 | 0.4059 | 0.6522 | 0.0 | 0.5 | 0.5209 | nan |
| 0.4632 | 6.0 | 90 | 0.6671 | 0.9725 | 0.9725 | 0.7932 | 0.7932 | 0.4663 | 0.4663 | 0.6522 | 0.0 | 0.5 | 0.5209 | nan |
| 0.3854 | 7.0 | 105 | 0.6447 | 0.9561 | 0.9561 | 0.7424 | 0.7424 | 0.4842 | 0.4842 | 0.6522 | 0.0 | 0.5 | 0.4307 | nan |
| 0.3154 | 8.0 | 120 | 0.7198 | 1.0102 | 1.0102 | 0.8113 | 0.8113 | 0.4241 | 0.4241 | 0.6522 | 0.0 | 0.5 | 0.4307 | nan |
| 0.2637 | 9.0 | 135 | 0.7221 | 1.0118 | 1.0118 | 0.8319 | 0.8319 | 0.4223 | 0.4223 | 0.5652 | 0.0 | 0.5 | 0.4150 | nan |
| 0.1962 | 10.0 | 150 | 0.6999 | 0.9962 | 0.9962 | 0.7945 | 0.7945 | 0.4401 | 0.4401 | 0.4783 | 0.0 | 0.5 | 0.4056 | nan |
| 0.1784 | 11.0 | 165 | 0.7335 | 1.0198 | 1.0198 | 0.7969 | 0.7969 | 0.4132 | 0.4132 | 0.5652 | 0.0 | 0.5 | 0.4150 | nan |
| 0.1531 | 12.0 | 180 | 0.8277 | 1.0833 | 1.0833 | 0.8839 | 0.8839 | 0.3378 | 0.3378 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.1425 | 13.0 | 195 | 0.8644 | 1.1070 | 1.1070 | 0.8726 | 0.8726 | 0.3085 | 0.3085 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0921 | 14.0 | 210 | 0.8874 | 1.1217 | 1.1217 | 0.9024 | 0.9024 | 0.2900 | 0.2900 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.0913 | 15.0 | 225 | 0.8663 | 1.1083 | 1.1083 | 0.8914 | 0.8914 | 0.3070 | 0.3070 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.08 | 16.0 | 240 | 0.8678 | 1.1093 | 1.1093 | 0.8762 | 0.8762 | 0.3057 | 0.3057 | 0.6522 | 0.0 | 0.5 | 0.5931 | nan |
| 0.0725 | 17.0 | 255 | 0.8497 | 1.0976 | 1.0976 | 0.8868 | 0.8868 | 0.3202 | 0.3202 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.0696 | 18.0 | 270 | 0.8533 | 1.1000 | 1.1000 | 0.8796 | 0.8796 | 0.3173 | 0.3173 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0632 | 19.0 | 285 | 0.8563 | 1.1018 | 1.1018 | 0.8768 | 0.8768 | 0.3150 | 0.3150 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0511 | 20.0 | 300 | 0.8433 | 1.0935 | 1.0935 | 0.8684 | 0.8684 | 0.3254 | 0.3254 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0517 | 21.0 | 315 | 0.8449 | 1.0945 | 1.0945 | 0.8758 | 0.8758 | 0.3240 | 0.3240 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.0556 | 22.0 | 330 | 0.8305 | 1.0851 | 1.0851 | 0.8469 | 0.8469 | 0.3356 | 0.3356 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0457 | 23.0 | 345 | 0.8369 | 1.0893 | 1.0893 | 0.8555 | 0.8555 | 0.3305 | 0.3305 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0496 | 24.0 | 360 | 0.8441 | 1.0940 | 1.0940 | 0.8648 | 0.8648 | 0.3247 | 0.3247 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0467 | 25.0 | 375 | 0.8470 | 1.0959 | 1.0959 | 0.8633 | 0.8633 | 0.3224 | 0.3224 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0446 | 26.0 | 390 | 0.8562 | 1.1018 | 1.1018 | 0.8708 | 0.8708 | 0.3151 | 0.3151 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.0476 | 27.0 | 405 | 0.8600 | 1.1042 | 1.1042 | 0.8714 | 0.8714 | 0.3120 | 0.3120 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.042 | 28.0 | 420 | 0.8657 | 1.1079 | 1.1079 | 0.8763 | 0.8763 | 0.3074 | 0.3074 | 0.4783 | 0.0 | 0.5 | 0.4440 | nan |
| 0.0431 | 29.0 | 435 | 0.8654 | 1.1077 | 1.1077 | 0.8734 | 0.8734 | 0.3077 | 0.3077 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
| 0.0423 | 30.0 | 450 | 0.8646 | 1.1072 | 1.1072 | 0.8721 | 0.8721 | 0.3083 | 0.3083 | 0.5652 | 0.0 | 0.5 | 0.5070 | nan |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
|
responsibility-framing/predict-perception-bert-blame-object
|
responsibility-framing
| 2022-03-10T15:51:04Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T15:49:05Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: predict-perception-bert-blame-object
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# predict-perception-bert-blame-object
This model is a fine-tuned version of [dbmdz/bert-base-italian-xxl-cased](https://huggingface.co/dbmdz/bert-base-italian-xxl-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5837
- Rmse: 0.5589
- Rmse Blame::a Un oggetto: 0.5589
- Mae: 0.3862
- Mae Blame::a Un oggetto: 0.3862
- R2: 0.2884
- R2 Blame::a Un oggetto: 0.2884
- Cos: 0.3913
- Pair: 0.0
- Rank: 0.5
- Neighbors: 0.5024
- Rsa: nan
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 20
- eval_batch_size: 8
- seed: 1996
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse | Rmse Blame::a Un oggetto | Mae | Mae Blame::a Un oggetto | R2 | R2 Blame::a Un oggetto | Cos | Pair | Rank | Neighbors | Rsa |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------------------------:|:------:|:-----------------------:|:-------:|:----------------------:|:------:|:----:|:----:|:---------:|:---:|
| 1.0603 | 1.0 | 15 | 0.8503 | 0.6745 | 0.6745 | 0.4386 | 0.4386 | -0.0365 | -0.0365 | 0.1304 | 0.0 | 0.5 | 0.5197 | nan |
| 0.9662 | 2.0 | 30 | 0.8510 | 0.6748 | 0.6748 | 0.4548 | 0.4548 | -0.0374 | -0.0374 | 0.0435 | 0.0 | 0.5 | 0.4840 | nan |
| 0.9438 | 3.0 | 45 | 0.7622 | 0.6386 | 0.6386 | 0.4541 | 0.4541 | 0.0709 | 0.0709 | 0.0435 | 0.0 | 0.5 | 0.4635 | nan |
| 0.9096 | 4.0 | 60 | 0.8301 | 0.6665 | 0.6665 | 0.4305 | 0.4305 | -0.0119 | -0.0119 | 0.0435 | 0.0 | 0.5 | 0.3499 | nan |
| 0.8383 | 5.0 | 75 | 0.7306 | 0.6252 | 0.6252 | 0.3814 | 0.3814 | 0.1094 | 0.1094 | 0.3043 | 0.0 | 0.5 | 0.5098 | nan |
| 0.7828 | 6.0 | 90 | 0.7434 | 0.6307 | 0.6307 | 0.4005 | 0.4005 | 0.0937 | 0.0937 | 0.3043 | 0.0 | 0.5 | 0.4335 | nan |
| 0.7028 | 7.0 | 105 | 0.7218 | 0.6214 | 0.6214 | 0.4090 | 0.4090 | 0.1202 | 0.1202 | 0.3913 | 0.0 | 0.5 | 0.4470 | nan |
| 0.6661 | 8.0 | 120 | 0.7434 | 0.6307 | 0.6307 | 0.4042 | 0.4042 | 0.0938 | 0.0938 | 0.3913 | 0.0 | 0.5 | 0.4470 | nan |
| 0.578 | 9.0 | 135 | 0.7719 | 0.6426 | 0.6426 | 0.3975 | 0.3975 | 0.0591 | 0.0591 | 0.3913 | 0.0 | 0.5 | 0.4470 | nan |
| 0.544 | 10.0 | 150 | 0.7117 | 0.6171 | 0.6171 | 0.4126 | 0.4126 | 0.1324 | 0.1324 | 0.2174 | 0.0 | 0.5 | 0.3489 | nan |
| 0.4638 | 11.0 | 165 | 0.6683 | 0.5980 | 0.5980 | 0.3952 | 0.3952 | 0.1853 | 0.1853 | 0.3043 | 0.0 | 0.5 | 0.3989 | nan |
| 0.3998 | 12.0 | 180 | 0.6772 | 0.6019 | 0.6019 | 0.4201 | 0.4201 | 0.1745 | 0.1745 | 0.3043 | 0.0 | 0.5 | 0.3989 | nan |
| 0.3403 | 13.0 | 195 | 0.6576 | 0.5932 | 0.5932 | 0.4237 | 0.4237 | 0.1984 | 0.1984 | 0.2174 | 0.0 | 0.5 | 0.3491 | nan |
| 0.2839 | 14.0 | 210 | 0.6281 | 0.5797 | 0.5797 | 0.4208 | 0.4208 | 0.2344 | 0.2344 | 0.2174 | 0.0 | 0.5 | 0.3491 | nan |
| 0.2619 | 15.0 | 225 | 0.6254 | 0.5785 | 0.5785 | 0.3752 | 0.3752 | 0.2376 | 0.2376 | 0.3913 | 0.0 | 0.5 | 0.5756 | nan |
| 0.2175 | 16.0 | 240 | 0.6074 | 0.5701 | 0.5701 | 0.3985 | 0.3985 | 0.2596 | 0.2596 | 0.3043 | 0.0 | 0.5 | 0.4142 | nan |
| 0.1884 | 17.0 | 255 | 0.6045 | 0.5687 | 0.5687 | 0.4036 | 0.4036 | 0.2631 | 0.2631 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1797 | 18.0 | 270 | 0.6038 | 0.5684 | 0.5684 | 0.3914 | 0.3914 | 0.2640 | 0.2640 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1316 | 19.0 | 285 | 0.6199 | 0.5759 | 0.5759 | 0.4078 | 0.4078 | 0.2443 | 0.2443 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1429 | 20.0 | 300 | 0.6119 | 0.5722 | 0.5722 | 0.3954 | 0.3954 | 0.2540 | 0.2540 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1202 | 21.0 | 315 | 0.6193 | 0.5756 | 0.5756 | 0.3987 | 0.3987 | 0.2451 | 0.2451 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1159 | 22.0 | 330 | 0.6218 | 0.5768 | 0.5768 | 0.3995 | 0.3995 | 0.2420 | 0.2420 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.1027 | 23.0 | 345 | 0.6207 | 0.5763 | 0.5763 | 0.4100 | 0.4100 | 0.2433 | 0.2433 | 0.3043 | 0.0 | 0.5 | 0.4142 | nan |
| 0.1006 | 24.0 | 360 | 0.5646 | 0.5496 | 0.5496 | 0.3687 | 0.3687 | 0.3117 | 0.3117 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.0902 | 25.0 | 375 | 0.5582 | 0.5465 | 0.5465 | 0.3714 | 0.3714 | 0.3196 | 0.3196 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.0901 | 26.0 | 390 | 0.5650 | 0.5498 | 0.5498 | 0.3704 | 0.3704 | 0.3112 | 0.3112 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.0937 | 27.0 | 405 | 0.5713 | 0.5529 | 0.5529 | 0.3735 | 0.3735 | 0.3036 | 0.3036 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.0812 | 28.0 | 420 | 0.5773 | 0.5558 | 0.5558 | 0.3759 | 0.3759 | 0.2962 | 0.2962 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.0911 | 29.0 | 435 | 0.5818 | 0.5579 | 0.5579 | 0.3832 | 0.3832 | 0.2908 | 0.2908 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
| 0.082 | 30.0 | 450 | 0.5837 | 0.5589 | 0.5589 | 0.3862 | 0.3862 | 0.2884 | 0.2884 | 0.3913 | 0.0 | 0.5 | 0.5024 | nan |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
|
responsibility-framing/predict-perception-bert-blame-victim
|
responsibility-framing
| 2022-03-10T15:48:51Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T15:44:32Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: predict-perception-bert-blame-victim
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# predict-perception-bert-blame-victim
This model is a fine-tuned version of [dbmdz/bert-base-italian-xxl-cased](https://huggingface.co/dbmdz/bert-base-italian-xxl-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5075
- Rmse: 0.4599
- Rmse Blame::a La vittima: 0.4599
- Mae: 0.3607
- Mae Blame::a La vittima: 0.3607
- R2: -0.1848
- R2 Blame::a La vittima: -0.1848
- Cos: 0.2174
- Pair: 0.0
- Rank: 0.5
- Neighbors: 0.2924
- Rsa: nan
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 20
- eval_batch_size: 8
- seed: 1996
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse | Rmse Blame::a La vittima | Mae | Mae Blame::a La vittima | R2 | R2 Blame::a La vittima | Cos | Pair | Rank | Neighbors | Rsa |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------------------------:|:------:|:-----------------------:|:-------:|:----------------------:|:-------:|:----:|:----:|:---------:|:---:|
| 1.0264 | 1.0 | 15 | 0.4334 | 0.4250 | 0.4250 | 0.3666 | 0.3666 | -0.0119 | -0.0119 | 0.1304 | 0.0 | 0.5 | 0.2703 | nan |
| 0.9814 | 2.0 | 30 | 0.4505 | 0.4333 | 0.4333 | 0.3744 | 0.3744 | -0.0517 | -0.0517 | 0.2174 | 0.0 | 0.5 | 0.2751 | nan |
| 0.9283 | 3.0 | 45 | 0.4349 | 0.4257 | 0.4257 | 0.3627 | 0.3627 | -0.0152 | -0.0152 | 0.1304 | 0.0 | 0.5 | 0.2779 | nan |
| 0.8904 | 4.0 | 60 | 0.4662 | 0.4408 | 0.4408 | 0.3773 | 0.3773 | -0.0884 | -0.0884 | -0.0435 | 0.0 | 0.5 | 0.2681 | nan |
| 0.836 | 5.0 | 75 | 0.4188 | 0.4177 | 0.4177 | 0.3609 | 0.3609 | 0.0223 | 0.0223 | 0.2174 | 0.0 | 0.5 | 0.3051 | nan |
| 0.8293 | 6.0 | 90 | 0.4142 | 0.4155 | 0.4155 | 0.3512 | 0.3512 | 0.0330 | 0.0330 | 0.2174 | 0.0 | 0.5 | 0.3220 | nan |
| 0.7629 | 7.0 | 105 | 0.3837 | 0.3999 | 0.3999 | 0.3387 | 0.3387 | 0.1041 | 0.1041 | 0.2174 | 0.0 | 0.5 | 0.3051 | nan |
| 0.7266 | 8.0 | 120 | 0.3664 | 0.3907 | 0.3907 | 0.3250 | 0.3250 | 0.1446 | 0.1446 | 0.3043 | 0.0 | 0.5 | 0.3409 | nan |
| 0.6121 | 9.0 | 135 | 0.3718 | 0.3936 | 0.3936 | 0.3312 | 0.3312 | 0.1320 | 0.1320 | 0.3043 | 0.0 | 0.5 | 0.3983 | nan |
| 0.5694 | 10.0 | 150 | 0.3679 | 0.3915 | 0.3915 | 0.3197 | 0.3197 | 0.1411 | 0.1411 | 0.3913 | 0.0 | 0.5 | 0.3518 | nan |
| 0.4647 | 11.0 | 165 | 0.3868 | 0.4015 | 0.4015 | 0.3340 | 0.3340 | 0.0970 | 0.0970 | 0.2174 | 0.0 | 0.5 | 0.3285 | nan |
| 0.4212 | 12.0 | 180 | 0.3717 | 0.3936 | 0.3936 | 0.3188 | 0.3188 | 0.1322 | 0.1322 | 0.3913 | 0.0 | 0.5 | 0.3518 | nan |
| 0.3605 | 13.0 | 195 | 0.3437 | 0.3784 | 0.3784 | 0.3066 | 0.3066 | 0.1976 | 0.1976 | 0.3043 | 0.0 | 0.5 | 0.3423 | nan |
| 0.2759 | 14.0 | 210 | 0.3892 | 0.4027 | 0.4027 | 0.3230 | 0.3230 | 0.0914 | 0.0914 | 0.3913 | 0.0 | 0.5 | 0.3518 | nan |
| 0.2868 | 15.0 | 225 | 0.3720 | 0.3937 | 0.3937 | 0.3218 | 0.3218 | 0.1315 | 0.1315 | 0.3913 | 0.0 | 0.5 | 0.3440 | nan |
| 0.2467 | 16.0 | 240 | 0.3881 | 0.4022 | 0.4022 | 0.3291 | 0.3291 | 0.0939 | 0.0939 | 0.3043 | 0.0 | 0.5 | 0.3363 | nan |
| 0.2013 | 17.0 | 255 | 0.4121 | 0.4144 | 0.4144 | 0.3373 | 0.3373 | 0.0380 | 0.0380 | 0.3043 | 0.0 | 0.5 | 0.3363 | nan |
| 0.1966 | 18.0 | 270 | 0.4808 | 0.4476 | 0.4476 | 0.3506 | 0.3506 | -0.1224 | -0.1224 | 0.3913 | 0.0 | 0.5 | 0.3214 | nan |
| 0.177 | 19.0 | 285 | 0.4263 | 0.4215 | 0.4215 | 0.3398 | 0.3398 | 0.0046 | 0.0046 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1589 | 20.0 | 300 | 0.4274 | 0.4220 | 0.4220 | 0.3363 | 0.3363 | 0.0022 | 0.0022 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1488 | 21.0 | 315 | 0.4548 | 0.4353 | 0.4353 | 0.3431 | 0.3431 | -0.0618 | -0.0618 | 0.3043 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1428 | 22.0 | 330 | 0.4405 | 0.4285 | 0.4285 | 0.3417 | 0.3417 | -0.0285 | -0.0285 | 0.3043 | 0.0 | 0.5 | 0.3363 | nan |
| 0.1294 | 23.0 | 345 | 0.4955 | 0.4544 | 0.4544 | 0.3565 | 0.3565 | -0.1568 | -0.1568 | 0.3913 | 0.0 | 0.5 | 0.3440 | nan |
| 0.1291 | 24.0 | 360 | 0.4861 | 0.4501 | 0.4501 | 0.3529 | 0.3529 | -0.1348 | -0.1348 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1187 | 25.0 | 375 | 0.4752 | 0.4450 | 0.4450 | 0.3518 | 0.3518 | -0.1095 | -0.1095 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1141 | 26.0 | 390 | 0.5131 | 0.4624 | 0.4624 | 0.3598 | 0.3598 | -0.1978 | -0.1978 | 0.3043 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1094 | 27.0 | 405 | 0.4863 | 0.4502 | 0.4502 | 0.3547 | 0.3547 | -0.1353 | -0.1353 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.0925 | 28.0 | 420 | 0.4900 | 0.4519 | 0.4519 | 0.3564 | 0.3564 | -0.1439 | -0.1439 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.108 | 29.0 | 435 | 0.5019 | 0.4573 | 0.4573 | 0.3590 | 0.3590 | -0.1719 | -0.1719 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
| 0.1054 | 30.0 | 450 | 0.5075 | 0.4599 | 0.4599 | 0.3607 | 0.3607 | -0.1848 | -0.1848 | 0.2174 | 0.0 | 0.5 | 0.2924 | nan |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
|
responsibility-framing/predict-perception-bert-blame-assassin
|
responsibility-framing
| 2022-03-10T15:44:18Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T15:32:55Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: predict-perception-bert-blame-assassin
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# predict-perception-bert-blame-assassin
This model is a fine-tuned version of [dbmdz/bert-base-italian-xxl-cased](https://huggingface.co/dbmdz/bert-base-italian-xxl-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5128
- Rmse: 1.0287
- Rmse Blame::a L'assassino: 1.0287
- Mae: 0.8883
- Mae Blame::a L'assassino: 0.8883
- R2: 0.5883
- R2 Blame::a L'assassino: 0.5883
- Cos: 0.6522
- Pair: 0.0
- Rank: 0.5
- Neighbors: 0.5795
- Rsa: nan
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 20
- eval_batch_size: 8
- seed: 1996
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rmse | Rmse Blame::a L'assassino | Mae | Mae Blame::a L'assassino | R2 | R2 Blame::a L'assassino | Cos | Pair | Rank | Neighbors | Rsa |
|:-------------:|:-----:|:----:|:---------------:|:------:|:-------------------------:|:------:|:------------------------:|:------:|:-----------------------:|:------:|:----:|:----:|:---------:|:---:|
| 1.0184 | 1.0 | 15 | 1.2219 | 1.5879 | 1.5879 | 1.4308 | 1.4308 | 0.0191 | 0.0191 | 0.3913 | 0.0 | 0.5 | 0.3781 | nan |
| 0.9214 | 2.0 | 30 | 1.0927 | 1.5017 | 1.5017 | 1.3634 | 1.3634 | 0.1227 | 0.1227 | 0.5652 | 0.0 | 0.5 | 0.4512 | nan |
| 0.7809 | 3.0 | 45 | 0.8206 | 1.3013 | 1.3013 | 1.1808 | 1.1808 | 0.3412 | 0.3412 | 0.4783 | 0.0 | 0.5 | 0.3819 | nan |
| 0.6593 | 4.0 | 60 | 0.5894 | 1.1029 | 1.1029 | 1.0145 | 1.0145 | 0.5268 | 0.5268 | 0.7391 | 0.0 | 0.5 | 0.6408 | nan |
| 0.4672 | 5.0 | 75 | 0.4759 | 0.9910 | 0.9910 | 0.8868 | 0.8868 | 0.6180 | 0.6180 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.3356 | 6.0 | 90 | 0.4220 | 0.9332 | 0.9332 | 0.8083 | 0.8083 | 0.6612 | 0.6612 | 0.6522 | 0.0 | 0.5 | 0.4249 | nan |
| 0.2782 | 7.0 | 105 | 0.4477 | 0.9612 | 0.9612 | 0.8046 | 0.8046 | 0.6406 | 0.6406 | 0.6522 | 0.0 | 0.5 | 0.6101 | nan |
| 0.2075 | 8.0 | 120 | 0.4389 | 0.9518 | 0.9518 | 0.8050 | 0.8050 | 0.6476 | 0.6476 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.1725 | 9.0 | 135 | 0.4832 | 0.9985 | 0.9985 | 0.8356 | 0.8356 | 0.6121 | 0.6121 | 0.7391 | 0.0 | 0.5 | 0.6616 | nan |
| 0.1642 | 10.0 | 150 | 0.4368 | 0.9494 | 0.9494 | 0.8060 | 0.8060 | 0.6493 | 0.6493 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.1172 | 11.0 | 165 | 0.4538 | 0.9677 | 0.9677 | 0.8174 | 0.8174 | 0.6357 | 0.6357 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.104 | 12.0 | 180 | 0.4672 | 0.9819 | 0.9819 | 0.8384 | 0.8384 | 0.6249 | 0.6249 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0822 | 13.0 | 195 | 0.4401 | 0.9530 | 0.9530 | 0.8107 | 0.8107 | 0.6467 | 0.6467 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0755 | 14.0 | 210 | 0.4464 | 0.9598 | 0.9598 | 0.8251 | 0.8251 | 0.6416 | 0.6416 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0801 | 15.0 | 225 | 0.4834 | 0.9988 | 0.9988 | 0.8604 | 0.8604 | 0.6119 | 0.6119 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.053 | 16.0 | 240 | 0.4846 | 1.0001 | 1.0001 | 0.8651 | 0.8651 | 0.6109 | 0.6109 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0573 | 17.0 | 255 | 0.4970 | 1.0128 | 1.0128 | 0.8743 | 0.8743 | 0.6010 | 0.6010 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0571 | 18.0 | 270 | 0.4803 | 0.9956 | 0.9956 | 0.8503 | 0.8503 | 0.6144 | 0.6144 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0483 | 19.0 | 285 | 0.4936 | 1.0093 | 1.0093 | 0.8740 | 0.8740 | 0.6037 | 0.6037 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0414 | 20.0 | 300 | 0.5138 | 1.0297 | 1.0297 | 0.8943 | 0.8943 | 0.5875 | 0.5875 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0513 | 21.0 | 315 | 0.5240 | 1.0399 | 1.0399 | 0.9050 | 0.9050 | 0.5793 | 0.5793 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0499 | 22.0 | 330 | 0.5275 | 1.0434 | 1.0434 | 0.9048 | 0.9048 | 0.5765 | 0.5765 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0423 | 23.0 | 345 | 0.5350 | 1.0508 | 1.0508 | 0.8872 | 0.8872 | 0.5705 | 0.5705 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0447 | 24.0 | 360 | 0.4963 | 1.0120 | 1.0120 | 0.8754 | 0.8754 | 0.6016 | 0.6016 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0364 | 25.0 | 375 | 0.5009 | 1.0167 | 1.0167 | 0.8809 | 0.8809 | 0.5979 | 0.5979 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0412 | 26.0 | 390 | 0.5060 | 1.0219 | 1.0219 | 0.8781 | 0.8781 | 0.5938 | 0.5938 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0297 | 27.0 | 405 | 0.5027 | 1.0185 | 1.0185 | 0.8838 | 0.8838 | 0.5964 | 0.5964 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0416 | 28.0 | 420 | 0.5071 | 1.0230 | 1.0230 | 0.8867 | 0.8867 | 0.5929 | 0.5929 | 0.7391 | 0.0 | 0.5 | 0.4884 | nan |
| 0.0327 | 29.0 | 435 | 0.5124 | 1.0283 | 1.0283 | 0.8883 | 0.8883 | 0.5887 | 0.5887 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
| 0.0383 | 30.0 | 450 | 0.5128 | 1.0287 | 1.0287 | 0.8883 | 0.8883 | 0.5883 | 0.5883 | 0.6522 | 0.0 | 0.5 | 0.5795 | nan |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0
|
OrfeasTsk/bert-base-uncased-finetuned-nq-large-batch
|
OrfeasTsk
| 2022-03-10T14:09:50Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"question-answering",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-10T13:57:57Z |
{ 'max_seq_length': 384,
'batch_size': 24,
'learning_rate': {'val': 3e-5, 'schelduler': 'Linear'},
'max_clip_norm': None,
'epochs': 2
}
|
rocca/sims4-faces
|
rocca
| 2022-03-10T13:41:17Z | 0 | 0 | null |
[
"onnx",
"license:mit",
"region:us"
] | null | 2022-03-10T12:15:19Z |
---
license: mit
---
Datasets here: https://huggingface.co/datasets/rocca/sims4-faces
|
Kevincp560/bigbird-pegasus-large-bigpatent-finetuned-pubMed
|
Kevincp560
| 2022-03-10T13:11:37Z | 4 | 2 |
transformers
|
[
"transformers",
"pytorch",
"bigbird_pegasus",
"text2text-generation",
"generated_from_trainer",
"dataset:pub_med_summarization_dataset",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-10T10:58:00Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- pub_med_summarization_dataset
metrics:
- rouge
model-index:
- name: bigbird-pegasus-large-bigpatent-finetuned-pubMed
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: pub_med_summarization_dataset
type: pub_med_summarization_dataset
args: document
metrics:
- name: Rouge1
type: rouge
value: 45.0851
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bigbird-pegasus-large-bigpatent-finetuned-pubMed
This model is a fine-tuned version of [google/bigbird-pegasus-large-bigpatent](https://huggingface.co/google/bigbird-pegasus-large-bigpatent) on the pub_med_summarization_dataset dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5403
- Rouge1: 45.0851
- Rouge2: 19.5488
- Rougel: 27.391
- Rougelsum: 41.112
- Gen Len: 231.608
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 2.1198 | 1.0 | 500 | 1.6285 | 43.0579 | 18.1792 | 26.421 | 39.0769 | 214.924 |
| 1.6939 | 2.0 | 1000 | 1.5696 | 44.0679 | 18.9331 | 26.84 | 40.0684 | 222.814 |
| 1.6195 | 3.0 | 1500 | 1.5506 | 44.7352 | 19.3532 | 27.2418 | 40.7454 | 229.396 |
| 1.5798 | 4.0 | 2000 | 1.5403 | 45.0415 | 19.5019 | 27.2969 | 40.951 | 231.044 |
| 1.5592 | 5.0 | 2500 | 1.5403 | 45.0851 | 19.5488 | 27.391 | 41.112 | 231.608 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.9.1
- Datasets 1.18.4
- Tokenizers 0.11.6
|
Chijioke/autonlp-mono-625317956
|
Chijioke
| 2022-03-10T12:46:27Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"distilbert",
"text-classification",
"autonlp",
"en",
"dataset:Chijioke/autonlp-data-mono",
"co2_eq_emissions",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T12:45:12Z |
---
tags: autonlp
language: en
widget:
- text: "I love AutoNLP 🤗"
datasets:
- Chijioke/autonlp-data-mono
co2_eq_emissions: 1.1406456838043837
---
# Model Trained Using AutoNLP
- Problem type: Multi-class Classification
- Model ID: 625317956
- CO2 Emissions (in grams): 1.1406456838043837
## Validation Metrics
- Loss: 0.513037919998169
- Accuracy: 0.8982035928143712
- Macro F1: 0.7843756230226546
- Micro F1: 0.8982035928143712
- Weighted F1: 0.8891653474608059
- Macro Precision: 0.8210878091622635
- Micro Precision: 0.8982035928143712
- Weighted Precision: 0.8888857327766032
- Macro Recall: 0.7731018645485747
- Micro Recall: 0.8982035928143712
- Weighted Recall: 0.8982035928143712
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/Chijioke/autonlp-mono-625317956
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("Chijioke/autonlp-mono-625317956", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("Chijioke/autonlp-mono-625317956", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
gustavecortal/gpt-j-fr-covid-news
|
gustavecortal
| 2022-03-10T10:05:27Z | 6 | 1 |
transformers
|
[
"transformers",
"pytorch",
"gptj",
"text-generation",
"causal-lm",
"fr",
"dataset:gustavecortal/fr_covid_news",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-09T12:18:31Z |
---
language: fr
license: mit
tags:
- causal-lm
- fr
datasets:
- gustavecortal/fr_covid_news
---
### GPT-J COVID-19 French News with 8-bit weights
This is a version of Cedille's GPT-J ([fr-boris](https://huggingface.co/gustavecortal/fr-boris-8bit)) with 6 billion parameters fine-tuned on [COVID-19 French News dataset](https://huggingface.co/datasets/gustavecortal/fr_covid_news) to generate French headlines related to COVID-19.
You can generate the model in colab or equivalent desktop gpu (e.g. single 1080Ti) as the model has 8-bit weights. Inspired by [GPT-J 8bit](https://huggingface.co/hivemind/gpt-j-6B-8bit).
Here's how to run it: [](https://colab.research.google.com/drive/1lMja-CPc0vm5_-gXNXAWU-9c0nom7vZ9)
This model can be easily loaded using the `GPTJForCausalLM` functionality:
```python
from transformers import GPTJForCausalLM
model = GPTJForCausalLM.from_pretrained("gustavecortal/gpt-j-fr-covid-news")
```
Remember, you have to Monkey-Patch the model before loading it (see Colab above).
## One thousand AI-generated French headlines related to COVID-19
How not to be disoriented in a pandemic era when faced with an immense flow of information? [This page](https://gustavecortal.com/project/covid) features one thousand AI-generated French headlines related to COVID-19.
## fr-boris
Boris is a 6B parameter autoregressive language model based on the GPT-J architecture and trained using the [mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax) codebase.
Boris was trained on around 78B tokens of French text from the [C4](https://huggingface.co/datasets/c4) dataset.
## Links
* [Gustave Cortal](https://twitter.com/gustavecortal)
|
Splend1dchan/byt5small-glue-mnli
|
Splend1dchan
| 2022-03-10T08:40:27Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"t5",
"text2text-generation",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-07T15:57:56Z |
byt5 finetuned on MNLI dataset for 3 epochs, with lr=1e-4
valid matched acc = 0.80
|
cammy/bart-large-cnn-100-lit-evalMA
|
cammy
| 2022-03-10T07:49:09Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bart",
"text2text-generation",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-10T06:32:37Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: bart-large-cnn-100-lit-evalMA
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-large-cnn-100-lit-evalMA
This model is a fine-tuned version of [facebook/bart-large-cnn](https://huggingface.co/facebook/bart-large-cnn) on an unknown dataset.
It achieves the following results on the evaluation set:
- eval_loss: 2.1514
- eval_rouge1: 27.8026
- eval_rouge2: 11.2998
- eval_rougeL: 21.4708
- eval_rougeLsum: 24.6333
- eval_gen_len: 62.5
- eval_runtime: 25.6587
- eval_samples_per_second: 0.39
- eval_steps_per_second: 0.39
- epoch: 2.0
- step: 200
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.2
- Datasets 1.18.3
- Tokenizers 0.11.0
|
momo/MOTOD_pre_trained
|
momo
| 2022-03-10T07:29:29Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-09T02:23:19Z |
---
license: apache-2.0
---
|
amanm27/bert-base-uncased-wiki-sports-scouting
|
amanm27
| 2022-03-10T07:18:56Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-10T07:14:06Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-wiki-sports-scouting
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-wiki-sports-scouting
This model is a fine-tuned version of [amanm27/bert-base-uncased-wiki-sports](https://huggingface.co/amanm27/bert-base-uncased-wiki-sports) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4909
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 378 | 1.6816 |
| 1.9594 | 2.0 | 756 | 1.5421 |
| 1.66 | 3.0 | 1134 | 1.5022 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
amanm27/bert-base-uncased-sports-scouting
|
amanm27
| 2022-03-10T07:12:38Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-10T07:07:45Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-sports-scouting
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-sports-scouting
This model is a fine-tuned version of [amanm27/bert-base-uncased-sports](https://huggingface.co/amanm27/bert-base-uncased-sports) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5127
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 378 | 1.7194 |
| 2.0165 | 2.0 | 756 | 1.5709 |
| 1.6935 | 3.0 | 1134 | 1.5282 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
peterhsu/mt5-small-finetuned-amazon-en-zh_TW
|
peterhsu
| 2022-03-10T07:05:34Z | 22 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"mt5",
"text2text-generation",
"summarization",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-05T09:05:00Z |
---
license: apache-2.0
tags:
- summarization
- generated_from_trainer
metrics:
- rouge
model-index:
- name: mt5-small-finetuned-amazon-en-zh_TW
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mt5-small-finetuned-amazon-en-zh_TW
This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 3.2408
- Rouge1: 15.8831
- Rouge2: 7.1676
- Rougel: 15.5523
- Rougelsum: 15.4954
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5.6e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 7
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:------:|:-------:|:---------:|
| 7.5388 | 1.0 | 838 | 3.5888 | 12.6081 | 5.3611 | 12.3495 | 12.2926 |
| 4.0043 | 2.0 | 1676 | 3.4038 | 13.8517 | 6.3417 | 13.4755 | 13.4913 |
| 3.6776 | 3.0 | 2514 | 3.3294 | 15.1519 | 7.3842 | 14.8844 | 14.8458 |
| 3.4929 | 4.0 | 3352 | 3.2668 | 15.6067 | 7.4016 | 15.3715 | 15.2908 |
| 3.387 | 5.0 | 4190 | 3.2855 | 15.0546 | 7.3065 | 14.8271 | 14.7755 |
| 3.302 | 6.0 | 5028 | 3.2457 | 15.0213 | 6.6597 | 14.6131 | 14.5641 |
| 3.2806 | 7.0 | 5866 | 3.2408 | 15.8831 | 7.1676 | 15.5523 | 15.4954 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
amanm27/bert-base-uncased-sports
|
amanm27
| 2022-03-10T06:40:10Z | 4 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-10T06:32:29Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-sports
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-sports
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.0064
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.4926 | 1.0 | 912 | 2.1186 |
| 2.2168 | 2.0 | 1824 | 2.0392 |
| 2.1327 | 3.0 | 2736 | 2.0081 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
amanm27/bert-base-uncased-wiki
|
amanm27
| 2022-03-10T06:15:01Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-10T05:58:51Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-wiki
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-wiki
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.7509
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.9294 | 1.0 | 2319 | 1.7732 |
| 1.8219 | 2.0 | 4638 | 1.7363 |
| 1.7957 | 3.0 | 6957 | 1.7454 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
kyleinincubated/autonlp-cat33-624317932
|
kyleinincubated
| 2022-03-10T06:10:56Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"autonlp",
"zh",
"dataset:kyleinincubated/autonlp-data-cat33",
"co2_eq_emissions",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-10T06:09:35Z |
---
tags: autonlp
language: zh
widget:
- text: "I love AutoNLP 🤗"
datasets:
- kyleinincubated/autonlp-data-cat33
co2_eq_emissions: 1.2490471218570545
---
# Model Trained Using AutoNLP
- Problem type: Multi-class Classification
- Model ID: 624317932
- CO2 Emissions (in grams): 1.2490471218570545
## Validation Metrics
- Loss: 0.5579860806465149
- Accuracy: 0.8717391304347826
- Macro F1: 0.6625543939916455
- Micro F1: 0.8717391304347827
- Weighted F1: 0.8593303742671491
- Macro Precision: 0.7214757380849891
- Micro Precision: 0.8717391304347826
- Weighted Precision: 0.8629042654788023
- Macro Recall: 0.6540187758140144
- Micro Recall: 0.8717391304347826
- Weighted Recall: 0.8717391304347826
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/kyleinincubated/autonlp-cat33-624317932
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("kyleinincubated/autonlp-cat33-624317932", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("kyleinincubated/autonlp-cat33-624317932", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
franz96521/Recipe-Creator
|
franz96521
| 2022-03-10T00:52:18Z | 0 | 1 | null |
[
"region:us"
] | null | 2022-03-08T23:36:06Z |
data origin https://recipenlg.cs.put.poznan.pl/dataset
create environment
```
conda env create -v -f Recipe-Creator.yml
conda activate Recipe-Creator
```
|
amanm27/bert-base-uncased-scouting
|
amanm27
| 2022-03-10T00:40:07Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"fill-mask",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-10T00:27:42Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bert-base-uncased-scouting
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-scouting
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5443
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| No log | 1.0 | 378 | 1.7727 |
| 2.1016 | 2.0 | 756 | 1.6040 |
| 1.7298 | 3.0 | 1134 | 1.5572 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.10.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
edubz/anne_bradstreet
|
edubz
| 2022-03-09T23:44:03Z | 7 | 2 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T22:03:42Z |
---
license: mit
---
This model was trained on a new dataset composed of available poems by Anne Bradstreet hosted by [Public Domain Poetry.](https://www.public-domain-poetry.com/anne-bradstreet) Specifically I downloaded all 40 poems and fine-tuned a bert-base-uncased text classification model on Amazon SageMaker. For the negative class, I actually generated GPT-2 samples of length 70. That is to say, for each line of Bradstreet I generated a generic GPT-2 reposes. I considered these responses my negative class.
In the classifier, I had a total of 6947 positive lines written by Anne Bradstreet, and 5219 lines generated by GPT-2 in response, totally a dataset of 12,166 labeled lines. I used only the GPT-2 responses in the training set, keeping the actual Bradstreet lines in the positive samples alone.
I split the train and test set in 80/20, leaving a total of 9732 labeled samples in training, and 2435 samples in test.
These I trained on SageMaker, using the Hugging Face deep learning container. I also used SageMaker Training Compiler, which achieved 64 samples per batch on an ml.p3.2xlarge. After 42 minutes of training, on only 5 epochs, I achieved a train loss of 0.0714. Test loss is forthcoming.
In my own tests, the model seems to be always very confident. That is to say, it routinely gives a confidence score of at least 99.8%. All predictions should be single-lines only, as this is how the model was fine-tuned. Multiple lines in a prediction request will always result in a Label0 response, ie not written by Anne Bradstreet, even if pulled directly from her works.
In short, the model seems to know the difference between generic GPT-2 text responding to a Bradstreet prompt, vs the output of a model fine-tuned on Bradstreet text and generating based on Bradstreet responses.
This was developed exclusively for use at an upcoming workshop.
|
megantosh/flair-arabic-dialects-codeswitch-egy-lev
|
megantosh
| 2022-03-09T22:12:57Z | 8 | 0 |
flair
|
[
"flair",
"pytorch",
"token-classification",
"sequence-tagger-model",
"Dialectal Arabic",
"Code-Switching",
"Code-Mixing",
"ar",
"en",
"dataset:4Dialects",
"dataset:MADAR",
"dataset:CSCS",
"license:apache-2.0",
"region:us"
] |
token-classification
| 2022-03-02T23:29:05Z |
---
language:
- ar
- en
license: apache-2.0
datasets:
- 4Dialects
- MADAR
- CSCS
thumbnail: https://www.informatik.hu-berlin.de/en/forschung-en/gebiete/ml-en/resolveuid/a6f82e0d7fa446a59c902cac4cafa9cb/@@images/image/preview
tags:
- flair
- token-classification
- sequence-tagger-model
- Dialectal Arabic
- Code-Switching
- Code-Mixing
metrics:
- f1
widget:
- text: "طلعوا جماعة الممانعة بالسياسة ما بيعرفوا ولا بالصحة بيعرفوا ولا حتى بالدين"
- text: "أعلم أن هذا يبدو غير عادل ، لكن لا يمكن أن يكون هناك ظلم"
- text: "أنا عارف أن الموضوع ده شكله مش عادل ، بس لا يمكن أن يكون فيه ظلم"
---
# Arabic Flair + fastText Part-of-Speech tagging Model (Egyptian and Levant)
Pretrained Part-of-Speech tagging model built on a joint corpus written in Egyptian and Levantine (Jordanian, Lebanese, Palestinian, Syrian) dialects with code-switching of Egyptian Arabic and English. The model is trained using [Flair](https://aclanthology.org/C18-1139/) (forward+backward)and [fastText](https://fasttext.cc) embeddings.
# Pretraining Corpora:
This sequence labeling model was pretrained on three corpora jointly:
1. [4 Dialects](https://huggingface.co/datasets/viewer/?dataset=arabic_pos_dialect)
A Dialectal Arabic Datasets containing four dialects of Arabic, Egyptian (EGY), Levantine (LEV), Gulf (GLF), and Maghrebi (MGR). Each dataset consists of a set of 350 manually segmented and PoS tagged tweets.
2. [UD South Levantine Arabic MADAR](https://universaldependencies.org/treebanks/ajp_madar/index.html)
A Dataset with 100 manually-annotated sentences taken from the [MADAR](https://camel.abudhabi.nyu.edu/madar/) (Multi-Arabic Dialect Applications and Resources) project by [Shorouq Zahra](mailto:[email protected]).
3. Parts of the Cairo Students Code-Switch (CSCS) corpus developed for ["Collection and Analysis of Code-switch Egyptian Arabic-English Speech Corpus"](https://aclanthology.org/L18-1601.pdf) by Hamed et al.
# Usage
```python
from flair.data import Sentence
from flair.models import SequenceTagger
tagger = SequenceTagger.load("megantosh/flair-arabic-dialects-codeswitch-egy-lev")
sentence = Sentence('عمرو عادلي أستاذ للاقتصاد السياسي المساعد في الجامعة الأمريكية بالقاهرة .')
tagger.predict(sentence)
for entity in sentence.get_spans('pos'):
print(entity)
```
Due to the right-to-left in left-to-right context, some formatting errors might occur. and your code might appear like [this](https://ibb.co/ky20Lnq), (link accessed on 2020-10-27)
<!--# Example
# Tagset-->
# Scores & Tagset
<details>
| |precision | recall | f1-score | support|
|--|-----------|------|-------------|--------------|
|INTJ | 0.8182 | 0.9000 |0.8571 | 10|
|OUN | 0.9009 | 0.9402 |0.9201 | 435|
|NUM | 0.9524 | 0.8333 | 0.8889 | 24|
|ADJ |0.8762 | 0.7603 | 0.8142 | 121|
|ADP |0.9903 |0.9623 | 0.9761 |106|
| CCONJ | 0.9600 | 0.9730 | 0.9664 | 74|
|PROPN | 0.9333 | 0.9333 | 0.9333 | 15|
| ADV | 0.9135 | 0.8051 | 0.8559 | 118|
|VERB | 0.8852 | 0.9231 | 0.9038 | 117|
|PRON | 0.9620 | 0.9465 | 0.9542 | 187|
|SCONJ | 0.8571 | 0.9474 | 0.9000 | 19|
|PART | 0.9350 | 0.9791 | 0.9565 | 191|
| DET | 0.9348 | 0.9149 | 0.9247 | 47|
|PUNCT | 1.0000 | 1.0000 | 1.0000 | 35|
| AUX | 0.9286 | 0.9811 | 0.9541 | 53|
|MENTION | 0.9231 | 1.0000 | 0.9600 | 12|
| V | 0.8571 | 0.8780 | 0.8675 | 82|
| FUT-PART+V+PREP+PRON |1.0000 | 0.0000 | 0.0000 | 1|
| PROG-PART+V+PRON+PREP+PRON | 0.0000 | 1.0000 | 0.0000 | 0|
|ADJ+NSUFF | 0.6111 | 0.8462 | 0.7097 | 26|
|NOUN+NSUFF | 0.8182 | 0.8438 | 0.8308 | 64|
|PREP+PRON | 0.9565 | 0.9565 | 0.9565 | 23|
| PUNC | 0.9941 | 1.0000 | 0.9971 | 169|
| EOS |1.0000 | 1.0000 | 1.0000 | 70|
| NOUN+PRON | 0.6986 | 0.8500 | 0.7669 | 60|
| V+PRON | 0.7258 | 0.8036 | 0.7627 | 56|
| PART+PRON | 1.0000 | 0.9474 | 0.9730 | 19|
| PROG-PART+V | 0.8333 | 0.9302 | 0.8791 | 43|
| DET+NOUN | 0.9625 | 1.0000 | 0.9809 | 77|
| NOUN+NSUFF+PRON | 0.9091 | 0.7143 | 0.8000 | 14|
| PROG-PART+V+PRON | 0.7083 | 0.9444 | 0.8095 | 18|
| PREP+NOUN+NSUFF | 0.6667 | 0.4000 | 0.5000 5|
| NOUN+NSUFF+NSUFF | 1.0000 | 0.0000 | 0.0000 | 3|
| CONJ | 0.9722 | 1.0000 | 0.9859 | 35|
| V+PRON+PRON | 0.6364 | 0.5833 | 0.6087 | 12|
| FOREIGN | 0.6667 | 0.6667 | 0.6667 | 3|
| PREP+NOUN | 0.6316 | 0.7500 | 0.6857 | 16|
| DET+NOUN+NSUFF | 0.9000 | 0.9310 | 0.9153 | 29|
| DET+ADJ+NSUFF | 1.0000 | 0.5714 | 0.7273 | 7|
| CONJ+PRON | 1.0000 | 0.8750 | 0.9333 | 8|
| NOUN+CASE | 0.0000 | 0.0000 | 0.0000 | 2|
| DET+ADJ | 1.0000 | 0.6667 | 0.8000 | 6|
| PREP | 1.0000 | 0.9718 | 0.9857 | 71|
| CONJ+FUT-PART+V | 0.0000 | 0.0000 | 0.0000 | 1|
| CONJ+V | 0.6667 | 0.7500 | 0.7059 | 8|
| FUT-PART | 1.0000 | 1.0000 | 1.0000 | 2|
| ADJ+PRON | 1.0000 | 0.0000 | 0.0000 | 8|
| CONJ+PREP+NOUN+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+NOUN+PRON | 0.3750 | 1.0000 | 0.5455 | 3|
| PART+ADJ | 1.0000 | 0.0000 | 0.0000 | 1|
| PART+NOUN | 0.5000 | 1.0000 | 0.6667 | 1|
| CONJ+PREP+NOUN | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+NOUN | 0.7000 | 0.7778 | 0.7368 | 9|
| URL | 1.0000 | 1.0000 | 1.0000 | 3|
| CONJ+FUT-PART | 1.0000 | 0.0000 | 0.0000 | 1|
| FUT-PART+V | 0.8571 | 0.6000 | 0.7059 | 10|
| PREP+NOUN+NSUFF+NSUFF | 1.0000 | 0.0000 | 0.0000 | 1|
| HASH | 1.0000 | 0.9412 | 0.9697 | 17|
| ADJ+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 3|
| PREP+NOUN+PRON | 0.0000 | 0.0000 | 0.0000 | 1|
| EMOT | 1.0000 | 0.8889 | 0.9412 | 18|
| CONJ+PREP | 1.0000 | 0.7500 | 0.8571 | 4|
| PREP+DET+NOUN+NSUFF | 1.0000 | 0.7500 | 0.8571 | 4|
| PRON+DET+NOUN+NSUFF | 0.0000 | 1.0000 | 0.0000 | 0|
| V+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 5|
| V+PRON+PREP+PRON | 0.0000 | 1.0000 | 0.0000 | 0|
| CONJ+NOUN+NSUFF | 0.5000 | 0.5000 | 0.5000 | 2|
| V+NEG-PART | 1.0000 | 0.0000 | 0.0000 | 2|
| PREP+DET+NOUN | 0.9091 | 1.0000 | 0.9524 | 10|
| PREP+V | 1.0000 | 0.0000 | 0.0000 | 2|
| CONJ+PART | 1.0000 | 0.7778 | 0.8750 | 9|
| CONJ+V+PRON | 1.0000 | 1.0000 | 1.0000 | 5|
| PROG-PART+V+PREP+PRON | 1.0000 | 0.5000 | 0.6667 | 2|
| PREP+NOUN+NSUFF+PRON | 1.0000 | 1.0000 | 1.0000 | 1|
| ADJ+CASE | 1.0000 | 0.0000 | 0.0000 | 1|
| PART+NOUN+PRON | 1.0000 | 1.0000 | 1.0000 | 1|
| PART+V | 1.0000 | 0.0000 | 0.0000 | 3|
| PART+V+PRON | 0.0000 | 1.0000 | 0.0000 | 0|
| FUT-PART+V+PRON | 0.0000 | 1.0000 | 0.0000 | 0|
|FUT-PART+V+PRON+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
|CONJ+V+PRON+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+V+PREP+PRON | 0.0000 | 1.0000 | 0.0000 | 0|
|CONJ+DET+NOUN+NSUFF | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+DET+NOUN | 0.6667 | 1.0000 | 0.8000 | 2|
| CONJ+PREP+DET+NOUN | 1.0000 | 1.0000 | 1.0000 | 1|
| PREP+PART | 1.0000 | 0.0000 | 0.0000 | 2|
| PART+V+PRON+NEG-PART | 0.3333 | 0.3333 | 0.3333 | 3|
| PART+V+NEG-PART | 0.3333 | 0.5000 | 0.4000 | 2|
| PART+PREP+NEG-PART | 1.0000 | 1.0000 | 1.0000 | 3|
| PART+PROG-PART+V+NEG-PART | 1.0000 | 0.3333 | 0.5000 | 3|
| PREP+DET+NOUN+NSUFF+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| PREP+PRON+DET+NOUN | 0.0000 | 1.0000 | 0.0000 | 0|
| PART+NSUFF | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+PROG-PART+V+PRON | 1.0000 | 1.0000 | 1.0000 | 1|
| PART+PREP+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+PART+PREP | 1.0000 | 0.0000 | 0.0000 | 1|
| NUM+NSUFF | 0.6667 | 0.6667 | 0.6667 | 3|
| CONJ+PART+V+PRON+NEG-PART | 1.0000 | 1.0000 | 1.0000 | 1|
| PART+NOUN+NEG-PART | 1.0000 | 1.0000 | 1.0000 | 1|
| CONJ+ADJ+NSUFF | 1.0000 | 0.0000 | 0.0000 | 1|
| PREP+ADJ | 1.0000 | 0.0000 | 0.0000 | 1|
| ADJ+NSUFF+PRON | 1.0000 | 0.0000 | 0.0000 | 2|
| CONJ+PROG-PART+V | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+PART+PROG-PART+V+PREP+PRON+NEG-PART | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+PART+PREP+PRON+NEG-PART | 0.0000 | 1.0000 | 0.0000 | 0|
| PREP+PART+PRON | 1.0000 | 0.0000 | 0.0000 | 1|
| CONJ+ADV+NSUFF | 1.0000 | 0.0000 |0.0000 | 1|
| CONJ+ADV | 0.0000 | 1.0000 | 0.0000 | 0|
| PART+NOUN+PRON+NEG-PART | 0.0000 | 1.0000 | 0.0000 | 0|
| CONJ+ADJ | 1.0000 | 1.0000 | 1.0000 | 1|
</details>
- F-score (micro): 0.8974
- F-score (macro): 0.5188
- Accuracy (incl. no class): 0.901
Expand details below to show class scores for each tag. Note that tag compounds (a tag made for multiple agglutinated parts of speech) are considered as separate ones.
# Citation
*if you use this model, please consider citing [this work](https://www.researchgate.net/publication/358956953_Sequence_Labeling_Architectures_in_Diglossia_-_a_case_study_of_Arabic_and_its_dialects):*
```latex
@unpublished{MMHU21
author = "M. Megahed",
title = "Sequence Labeling Architectures in Diglossia",
year = {2021},
doi = "10.13140/RG.2.2.34961.10084"
url = {https://www.researchgate.net/publication/358956953_Sequence_Labeling_Architectures_in_Diglossia_-_a_case_study_of_Arabic_and_its_dialects}
}
```
|
megantosh/flair-arabic-multi-ner
|
megantosh
| 2022-03-09T22:12:22Z | 698 | 5 |
flair
|
[
"flair",
"pytorch",
"Text Classification",
"token-classification",
"sequence-tagger-model",
"ar",
"en",
"dataset:AQMAR",
"dataset:ANERcorp",
"license:apache-2.0",
"region:us"
] |
token-classification
| 2022-03-02T23:29:05Z |
---
language:
- ar
- en
license: apache-2.0
datasets:
- AQMAR
- ANERcorp
thumbnail: https://www.informatik.hu-berlin.de/en/forschung-en/gebiete/ml-en/resolveuid/a6f82e0d7fa446a59c902cac4cafa9cb/@@images/image/preview
tags:
- flair
- Text Classification
- token-classification
- sequence-tagger-model
metrics:
- f1
widget:
- text: أعرف كل شيء عن جيجي
- text: ترتقي شريحة M1 Pro وشريحة M1 Max ببنية شريحة M1 المذهلة إلى مستويات جديدة، إذ تأتيان للمرة الأولى ببنية نظام متكامل في شريحة (SoC) إلى جهاز نوت بوك للمحترفين.
- text: "اختارها خيري بشارة كممثلة، دون سابقة معرفة أو تجربة تمثيلية، لتقف بجانب فاتن حمامة في فيلم «يوم مر ويوم حلو» (1988) وهي ما زالت شابة لم تتخطَ عامها الثاني"
---
# Arabic NER Model using Flair Embeddings
Training was conducted over 94 epochs, using a linear decaying learning rate of 2e-05, starting from 0.225 and a batch size of 32 with GloVe and Flair forward and backward embeddings.
## Original Datasets:
- [AQMAR](http://www.cs.cmu.edu/~ark/ArabicNER/)
- [ANERcorp](http://curtis.ml.cmu.edu/w/courses/index.php/ANERcorp)
## Results:
- F1-score (micro) 0.8666
- F1-score (macro) 0.8488
| | Named Entity Type | True Posititves | False Positives | False Negatives | Precision | Recall | class-F1 |
|------|-|----|----|----|-----------|--------|----------|
| LOC | Location| 539 | 51 | 68 | 0.9136 | 0.8880 | 0.9006 |
| MISC | Miscellaneous|408 | 57 | 89 | 0.8774 | 0.8209 | 0.8482 |
| ORG | Organisation|167 | 43 | 64 | 0.7952 | 0.7229 | 0.7574 |
| PER | Person (no title)|501 | 65 | 60 | 0.8852 | 0.8930 | 0.8891 |
---
# Usage
```python
from flair.data import Sentence
from flair.models import SequenceTagger
import pyarabic.araby as araby
from icecream import ic
tagger = SequenceTagger.load("julien-c/flair-ner")
arTagger = SequenceTagger.load('megantosh/flair-arabic-multi-ner')
sentence = Sentence('George Washington went to Washington .')
arSentence = Sentence('عمرو عادلي أستاذ للاقتصاد السياسي المساعد في الجامعة الأمريكية بالقاهرة .')
# predict NER tags
tagger.predict(sentence)
arTagger.predict(arSentence)
# print sentence with predicted tags
ic(sentence.to_tagged_string)
ic(arSentence.to_tagged_string)
```
# Example
```bash
2021-07-07 14:30:59,649 loading file /Users/mega/.flair/models/flair-ner/f22eb997f66ae2eacad974121069abaefca5fe85fce71b49e527420ff45b9283.941c7c30b38aef8d8a4eb5c1b6dd7fe8583ff723fef457382589ad6a4e859cfc
2021-07-07 14:31:04,654 loading file /Users/mega/.flair/models/flair-arabic-multi-ner/c7af7ddef4fdcc681fcbe1f37719348afd2862b12aa1cfd4f3b93bd2d77282c7.242d030cb106124f7f9f6a88fb9af8e390f581d42eeca013367a86d585ee6dd6
ic| sentence.to_tagged_string: <bound method Sentence.to_tagged_string of Sentence: "George Washington went to Washington ." [− Tokens: 6 − Token-Labels: "George <B-PER> Washington <E-PER> went to Washington <S-LOC> ."]>
ic| arSentence.to_tagged_string: <bound method Sentence.to_tagged_string of Sentence: "عمرو عادلي أستاذ للاقتصاد السياسي المساعد في الجامعة الأمريكية بالقاهرة ." [− Tokens: 11 − Token-Labels: "عمرو <B-PER> عادلي <I-PER> أستاذ للاقتصاد السياسي المساعد في الجامعة <B-ORG> الأمريكية <I-ORG> بالقاهرة <B-LOC> ."]>
ic| entity: <PER-span (1,2): "George Washington">
ic| entity: <LOC-span (5): "Washington">
ic| entity: <PER-span (1,2): "عمرو عادلي">
ic| entity: <ORG-span (8,9): "الجامعة الأمريكية">
ic| entity: <LOC-span (10): "بالقاهرة">
ic| sentence.to_dict(tag_type='ner'):
{"text":"عمرو عادلي أستاذ للاقتصاد السياسي المساعد في الجامعة الأمريكية بالقاهرة .",
"labels":[],
{"entities":[{{{
"text":"عمرو عادلي",
"start_pos":0,
"end_pos":10,
"labels":[PER (0.9826)]},
{"text":"الجامعة الأمريكية",
"start_pos":45,
"end_pos":62,
"labels":[ORG (0.7679)]},
{"text":"بالقاهرة",
"start_pos":64,
"end_pos":72,
"labels":[LOC (0.8079)]}]}
"text":"George Washington went to Washington .",
"labels":[],
"entities":[{
{"text":"George Washington",
"start_pos":0,
"end_pos":17,
"labels":[PER (0.9968)]},
{"text":"Washington""start_pos":26,
"end_pos":36,
"labels":[LOC (0.9994)]}}]}
```
# Model Configuration
```python
SequenceTagger(
(embeddings): StackedEmbeddings(
(list_embedding_0): WordEmbeddings('glove')
(list_embedding_1): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1, inplace=False)
(encoder): Embedding(7125, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=7125, bias=True)
)
)
(list_embedding_2): FlairEmbeddings(
(lm): LanguageModel(
(drop): Dropout(p=0.1, inplace=False)
(encoder): Embedding(7125, 100)
(rnn): LSTM(100, 2048)
(decoder): Linear(in_features=2048, out_features=7125, bias=True)
)
)
)
(word_dropout): WordDropout(p=0.05)
(locked_dropout): LockedDropout(p=0.5)
(embedding2nn): Linear(in_features=4196, out_features=4196, bias=True)
(rnn): LSTM(4196, 256, batch_first=True, bidirectional=True)
(linear): Linear(in_features=512, out_features=15, bias=True)
(beta): 1.0
(weights): None
(weight_tensor) None
```
Due to the right-to-left in left-to-right context, some formatting errors might occur. and your code might appear like [this](https://ibb.co/ky20Lnq), (link accessed on 2020-10-27)
# Citation
*if you use this model, please consider citing [this work](https://www.researchgate.net/publication/358956953_Sequence_Labeling_Architectures_in_Diglossia_-_a_case_study_of_Arabic_and_its_dialects):*
```latex
@unpublished{MMHU21
author = "M. Megahed",
title = "Sequence Labeling Architectures in Diglossia",
year = {2021},
doi = "10.13140/RG.2.2.34961.10084"
url = {https://www.researchgate.net/publication/358956953_Sequence_Labeling_Architectures_in_Diglossia_-_a_case_study_of_Arabic_and_its_dialects}
}
```
|
sudoparsa/wav2vec2-base-finetuned-ks
|
sudoparsa
| 2022-03-09T22:06:19Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"dataset:superb",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2022-03-07T15:38:54Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- superb
metrics:
- accuracy
model-index:
- name: wav2vec2-base-finetuned-ks
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-base-finetuned-ks
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the superb dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0894
- Accuracy: 0.9828
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|:-------------:|:-----:|:----:|:--------:|:---------------:|
| 0.5003 | 1.0 | 399 | 0.9643 | 0.4284 |
| 0.1868 | 2.0 | 798 | 0.9748 | 0.1628 |
| 0.1413 | 3.0 | 1197 | 0.9796 | 0.1128 |
| 0.0965 | 4.0 | 1596 | 0.0950 | 0.9826 |
| 0.0915 | 5.0 | 1995 | 0.0894 | 0.9828 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.10.0+cu111
- Datasets 1.14.0
- Tokenizers 0.10.3
|
antho-data/distilbert-base-uncased-finetuned-emotion
|
antho-data
| 2022-03-09T21:27:17Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T20:30:05Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- emotion
metrics:
- accuracy
- f1
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: emotion
type: emotion
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9235
- name: F1
type: f1
value: 0.9237367861627231
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2294
- Accuracy: 0.9235
- F1: 0.9237
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8637 | 1.0 | 250 | 0.3319 | 0.9075 | 0.9050 |
| 0.2634 | 2.0 | 500 | 0.2294 | 0.9235 | 0.9237 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
saksornr/mt-align-finetuned-LST-en-to-th
|
saksornr
| 2022-03-09T20:41:54Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"marian",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-09T07:45:26Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: mt-align-finetuned-LST-en-to-th
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mt-align-finetuned-LST-en-to-th
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-mul](https://huggingface.co/Helsinki-NLP/opus-mt-en-mul) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 128
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
| No log | 1.0 | 77 | 1.6042 | 13.1732 | 26.144 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.2+cu113
- Datasets 1.18.4
- Tokenizers 0.11.6
|
Kevincp560/bigbird-pegasus-large-arxiv-finetuned-pubmed
|
Kevincp560
| 2022-03-09T19:30:11Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bigbird_pegasus",
"text2text-generation",
"generated_from_trainer",
"dataset:pub_med_summarization_dataset",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-09T17:14:25Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- pub_med_summarization_dataset
metrics:
- rouge
model-index:
- name: bigbird-pegasus-large-arxiv-finetuned-pubmed
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: pub_med_summarization_dataset
type: pub_med_summarization_dataset
args: document
metrics:
- name: Rouge1
type: rouge
value: 45.4807
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bigbird-pegasus-large-arxiv-finetuned-pubmed
This model is a fine-tuned version of [google/bigbird-pegasus-large-arxiv](https://huggingface.co/google/bigbird-pegasus-large-arxiv) on the pub_med_summarization_dataset dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6049
- Rouge1: 45.4807
- Rouge2: 20.0199
- Rougel: 28.3621
- Rougelsum: 41.4618
- Gen Len: 219.144
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 2.594 | 1.0 | 500 | 1.9879 | 33.6364 | 13.5074 | 21.4286 | 29.7158 | 189.014 |
| 1.9146 | 2.0 | 1000 | 1.6494 | 44.0056 | 19.0069 | 27.5142 | 40.0492 | 210.528 |
| 1.7378 | 3.0 | 1500 | 1.6213 | 44.7071 | 19.3559 | 27.6806 | 40.6124 | 213.596 |
| 1.692 | 4.0 | 2000 | 1.6081 | 45.1505 | 19.7355 | 28.06 | 41.0108 | 213.674 |
| 1.6656 | 5.0 | 2500 | 1.6049 | 45.4807 | 20.0199 | 28.3621 | 41.4618 | 219.144 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.9.1
- Datasets 1.18.4
- Tokenizers 0.11.6
|
hyechanjun/reverse-interview-question
|
hyechanjun
| 2022-03-09T18:57:52Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bart",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-09T18:52:33Z |
An AI model that, given a statement, generates a question that would have likely resulted in said statement.
Created for a Senior Project at Calvin University.
|
Noricum/wav2vec2-large-xls-r-300m-de-with-lm
|
Noricum
| 2022-03-09T18:14:21Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2022-03-08T15:45:28Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-300m-de-with-lm
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-large-xls-r-300m-de-with-lm
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.17.0
- Pytorch 1.9.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
ankitkhowal/minutes-of-meeting
|
ankitkhowal
| 2022-03-09T18:08:01Z | 43 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bart",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
Model to summarize the meeting transcripts.
|
petrichorRainbow/mrf-covid-bert
|
petrichorRainbow
| 2022-03-09T17:24:51Z | 2 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2022-03-09T16:59:51Z |
---
license: apache-2.0
---
|
petrichorRainbow/mrf-bert
|
petrichorRainbow
| 2022-03-09T17:12:06Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-09T16:59:17Z |
---
license: apache-2.0
---
|
orzhan/ruroberta-ruatd-binary
|
orzhan
| 2022-03-09T15:36:04Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"text-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T15:28:56Z |
sberbank-ai/ruRoberta-large fine-tuned for Russian Artificial Text Detection shared task
|
Narshion/mWACH_mBERT_System
|
Narshion
| 2022-03-09T13:49:35Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T12:28:12Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# model
This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on mWACH NEO dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6344
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5.0
### Training results
### Framework versions
- Transformers 4.12.4
- Pytorch 1.10.0+cu111
- Datasets 1.15.1
- Tokenizers 0.10.3
|
SGrannemann/marian-finetuned-kde4-en-to-fr
|
SGrannemann
| 2022-03-09T13:39:05Z | 3 | 0 |
transformers
|
[
"transformers",
"tf",
"marian",
"text2text-generation",
"generated_from_keras_callback",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-09T12:27:54Z |
---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: marian-finetuned-kde4-en-to-fr
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# marian-finetuned-kde4-en-to-fr
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-fr](https://huggingface.co/Helsinki-NLP/opus-mt-en-fr) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.6859
- Validation Loss: 0.8062
- Epoch: 2
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 5e-05, 'decay_steps': 17733, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
- training_precision: mixed_float16
### Training results
| Train Loss | Validation Loss | Epoch |
|:----------:|:---------------:|:-----:|
| 1.0582 | 0.8792 | 0 |
| 0.7977 | 0.8250 | 1 |
| 0.6859 | 0.8062 | 2 |
### Framework versions
- Transformers 4.17.0
- TensorFlow 2.8.0
- Datasets 1.18.4
- Tokenizers 0.11.6
|
MarioPenguin/beto_stars
|
MarioPenguin
| 2022-03-09T13:37:45Z | 5 | 0 |
transformers
|
[
"transformers",
"tf",
"bert",
"text-classification",
"generated_from_keras_callback",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T13:30:22Z |
---
tags:
- generated_from_keras_callback
model-index:
- name: beto_stars
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# beto_stars
This model is a fine-tuned version of [dccuchile/bert-base-spanish-wwm-uncased](https://huggingface.co/dccuchile/bert-base-spanish-wwm-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.8954
- Train Accuracy: 0.6248
- Validation Loss: 1.1278
- Validation Accuracy: 0.5148
- Epoch: 14
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'Adam', 'learning_rate': 5e-07, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
- training_precision: float32
### Training results
| Train Loss | Train Accuracy | Validation Loss | Validation Accuracy | Epoch |
|:----------:|:--------------:|:---------------:|:-------------------:|:-----:|
| 1.5935 | 0.2626 | 1.5553 | 0.3482 | 0 |
| 1.5072 | 0.3782 | 1.4289 | 0.4188 | 1 |
| 1.3590 | 0.4312 | 1.2929 | 0.4406 | 2 |
| 1.2463 | 0.4628 | 1.2197 | 0.4682 | 3 |
| 1.1754 | 0.4970 | 1.1785 | 0.4830 | 4 |
| 1.1299 | 0.5114 | 1.1533 | 0.4908 | 5 |
| 1.0847 | 0.5362 | 1.1398 | 0.5006 | 6 |
| 1.0492 | 0.5440 | 1.1273 | 0.5046 | 7 |
| 1.0278 | 0.5592 | 1.1237 | 0.5034 | 8 |
| 1.0031 | 0.5690 | 1.1171 | 0.5118 | 9 |
| 0.9798 | 0.5712 | 1.1163 | 0.5120 | 10 |
| 0.9598 | 0.5894 | 1.1180 | 0.5114 | 11 |
| 0.9406 | 0.5964 | 1.1219 | 0.5122 | 12 |
| 0.9178 | 0.6104 | 1.1269 | 0.5150 | 13 |
| 0.8954 | 0.6248 | 1.1278 | 0.5148 | 14 |
### Framework versions
- Transformers 4.17.0
- TensorFlow 2.8.0
- Datasets 1.18.4
- Tokenizers 0.11.6
|
akshaychaudhary/distilbert-base-uncased-finetuned-combinedmodel1-ner
|
akshaychaudhary
| 2022-03-09T12:59:14Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"token-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2022-03-09T11:01:49Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: distilbert-base-uncased-finetuned-combinedmodel1-ner
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-combinedmodel1-ner
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 2.3126
- Precision: 0.0289
- Recall: 0.1443
- F1: 0.0481
- Accuracy: 0.7058
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log | 1.0 | 312 | 1.5290 | 0.0431 | 0.2278 | 0.0725 | 0.6990 |
| 0.1106 | 2.0 | 624 | 2.0923 | 0.0341 | 0.1722 | 0.0569 | 0.7041 |
| 0.1106 | 3.0 | 936 | 2.3126 | 0.0289 | 0.1443 | 0.0481 | 0.7058 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
nickmuchi/vit-finetuned-chest-xray-pneumonia
|
nickmuchi
| 2022-03-09T12:50:04Z | 385 | 6 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"vit",
"image-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2022-03-09T12:04:10Z |
---
license: apache-2.0
tags:
- image-classification
- generated_from_trainer
metrics:
- accuracy
datasets:
- chest X-rays
widget:
- src: https://drive.google.com/uc?id=1ygVCyEn6mfsNwpT1ZvWxANg5_DvStA7M
example_title: PNEUMONIA
- src: https://drive.google.com/uc?id=1xjcIEDb8kuSd4wF44gCEgsc0PfRvs53m
example_title: NORMAL
model-index:
- name: vit-finetuned-chest-xray-pneumonia
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# vit-finetuned-chest-xray-pneumonia
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the [chest-xray-pneumonia](https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1271
- Accuracy: 0.9551
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 326 | 0.2739 | 0.9167 |
| 0.2238 | 2.0 | 652 | 0.2892 | 0.9071 |
| 0.2238 | 3.0 | 978 | 0.2077 | 0.9407 |
| 0.1385 | 4.0 | 1304 | 0.1349 | 0.9535 |
| 0.1347 | 5.0 | 1630 | 0.1271 | 0.9551 |
| 0.1347 | 6.0 | 1956 | 0.1458 | 0.9535 |
| 0.1112 | 7.0 | 2282 | 0.2040 | 0.9375 |
| 0.1063 | 8.0 | 2608 | 0.1423 | 0.9567 |
| 0.1063 | 9.0 | 2934 | 0.1473 | 0.9535 |
| 0.0944 | 10.0 | 3260 | 0.1385 | 0.9583 |
## Example Images
#### Pneumonia Chest X-Ray

#### Normal Chest X-Ray

### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
ScandinavianMrT/distilbert-SARC
|
ScandinavianMrT
| 2022-03-09T10:29:23Z | 26 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T06:17:28Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: distilbert-SARC
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-SARC
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- eval_loss: 0.4976
- eval_accuracy: 0.7590
- eval_runtime: 268.1875
- eval_samples_per_second: 753.782
- eval_steps_per_second: 47.113
- epoch: 1.0
- step: 50539
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
mrbalazs5/bert-finetuned-squad
|
mrbalazs5
| 2022-03-09T10:01:12Z | 5 | 0 |
transformers
|
[
"transformers",
"tf",
"bert",
"question-answering",
"generated_from_keras_callback",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-07T16:04:24Z |
---
license: apache-2.0
tags:
- generated_from_keras_callback
model-index:
- name: mrbalazs5/bert-finetuned-squad
results: []
---
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# mrbalazs5/bert-finetuned-squad
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.7151
- Epoch: 1
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 66546, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
- training_precision: mixed_float16
### Training results
| Train Loss | Epoch |
|:----------:|:-----:|
| 1.2226 | 0 |
| 0.7151 | 1 |
### Framework versions
- Transformers 4.16.2
- TensorFlow 2.8.0
- Datasets 1.18.3
- Tokenizers 0.11.6
|
kyleinincubated/autonlp-abbb-622117836
|
kyleinincubated
| 2022-03-09T09:30:07Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"autonlp",
"zh",
"dataset:kyleinincubated/autonlp-data-abbb",
"co2_eq_emissions",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T09:27:47Z |
---
tags: autonlp
language: zh
widget:
- text: "I love AutoNLP 🤗"
datasets:
- kyleinincubated/autonlp-data-abbb
co2_eq_emissions: 2.22514962526191
---
# Model Trained Using AutoNLP
- Problem type: Multi-class Classification
- Model ID: 622117836
- CO2 Emissions (in grams): 2.22514962526191
## Validation Metrics
- Loss: 1.2368708848953247
- Accuracy: 0.7973333333333333
- Macro F1: 0.46009076588978487
- Micro F1: 0.7973333333333333
- Weighted F1: 0.7712349116681224
- Macro Precision: 0.4527155928883903
- Micro Precision: 0.7973333333333333
- Weighted Precision: 0.7610710955220162
- Macro Recall: 0.4947868561369568
- Micro Recall: 0.7973333333333333
- Weighted Recall: 0.7973333333333333
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/kyleinincubated/autonlp-abbb-622117836
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("kyleinincubated/autonlp-abbb-622117836", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("kyleinincubated/autonlp-abbb-622117836", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
merve/anime-faces-discriminator
|
merve
| 2022-03-09T09:12:21Z | 0 | 0 |
keras
|
[
"keras",
"tf-keras",
"dcgan",
"dataset:anime-faces",
"license:apache-2.0",
"region:us"
] | null | 2022-03-04T16:47:21Z |
---
license: apache-2.0
library_name: keras
tags:
- dcgan
datasets:
- anime-faces
---
## Model description
Anime face discriminator model using [TensorFlow DCGAN example](https://www.tensorflow.org/tutorials/generative/dcgan).
## Training and evaluation data
Model is trained on [anime faces dataset](https://huggingface.co/datasets/merve/anime-faces).
## Intended use and biases
This model is not intended for production.
|
gsarti/it5-base-headline-generation
|
gsarti
| 2022-03-09T08:07:05Z | 5 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"headline-generation",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- headline-generation
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
model-index:
- name: it5-base-headline-generation
results:
- task:
type: headline-generation
name: "Headline generation"
dataset:
type: headgen_it
name: "HeadGen-IT"
metrics:
- type: rouge1
value: 0.310
name: "Test Rouge1"
- type: rouge2
value: 0.112
name: "Test Rouge2"
- type: rougeL
value: 0.270
name: "Test RougeL"
- type: bertscore
value: 0.433
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
---
# IT5 Base for News Headline Generation 🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Base](https://huggingface.co/gsarti/it5-base) model fine-tuned on news headline generation on the Italian HeadGen-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
hg = pipeline("text2text-generation", model='it5/it5-base-headline-generation')
hg("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-base-headline-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-base-headline-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-base-wiki-summarization
|
gsarti
| 2022-03-09T08:06:40Z | 16 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"wikipedia",
"summarization",
"wits",
"it",
"dataset:wits",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- wits
tags:
- italian
- sequence-to-sequence
- wikipedia
- summarization
- wits
widget:
- text: "La 5ª Commissione ha competenza per i disegni di legge riguardanti le specifiche materie del bilancio, del personale e dei servizi del Ministero dell'economia, nonché per i disegni di legge riguardanti la materia finanziaria. La Commissione è composta da 26 senatori (di cui 2 segretari, 2 vicepresidenti di cui 1 componente esterno, e un presidente) scelti in modo omogeneo tra i componenti di quel ramo del Parlamento, in modo da rispecchiarne le forze politiche presenti. Essi sono scelti dai gruppi parlamentari (e non dal Presidente, come invece accade per l'organismo della Giunta parlamentare): per la nomina dei membri ciascun Gruppo, entro cinque giorni dalla propria costituzione, procede, dandone comunicazione alla Presidenza del Senato, alla designazione dei propri rappresentanti nelle singole Commissioni permanenti. Ogni senatore chiamato a far parte del governo o eletto presidente della Commissione è, per la durata della carica, sostituito dal suo gruppo nella Commissione con un altro senatore, che continuerà ad appartenere anche alla Commissione di provenienza. Tranne in rari casi nessun Senatore può essere assegnato a più di una Commissione permanente. Le Commissioni permanenti sono rinnovate dopo il primo biennio della legislatura ed i loro componenti possono essere confermati."
- text: "Interni della chiesa Si pensa che già ai tempi di Gediminas vi fosse una piccola chiesa, probabilmente in legno. Nel 1408 circa Vitoldo costruì la chiesa dello Spirito Santo che andò in seguito ampliata. Nel 1501 Alessandro Jagellone lo donò al monastero domenicano, il più antico della Lituania, che nel 1679-88 fu ampliato e ricostruito. Di quel periodo sopravvivono le mura della chiesa, mentre l'arredamento interno fu realizzato nel 1749-1770 e la cupola affrontò dei lavori di restauro nel 1752-1760. Nel 1844 le autorità zariste chiusero il monastero e la chiesa divenne parrocchiale. Oggi serve la comunità polacca di Vilnius. Su via Šv. Ignoto fu fondato un monastero domenicano nel 1501. Come molti altri edifici, questo monastero fu convertito in una prigione dalle autorità zariste nel 1807. Costituì un luogo di prigionia per molti patrioti lituani, nello specifico i Filareti, i quali parteciparono alle rivolte del 1831 e del 1863. Organo La chiesa si trova lateralmente rispetto alla strada e non ha una facciata principale ben disegnata. L'altezza, inclusa la cupola, è di 51 m. La parte inferiore della facciata (con piccole torri gemelle) è ricoperta da edifici conventuali e l'esterno presenta caratteristiche architettoniche tipiche del tardo barocco. Celebre per i fantasiosi ornamenti rococò, l'interno della chiesa è tra i più celebri della Lituania per via dei cartigli con vari stemmi e affreschi lungo la navata: vi sono 16 altari nella chiesa. Gli altari e il pulpito sono assai decorati con sculture e ornamenti rotondi e in rilievo. Tra gli affreschi barocchi, si pensi alla composizione multi-figurale intitolata ''Apoteosi dello Spirito Santo'' (neobarocco, XIX secolo) nella cupola, 45 dipinti nella chiesa (tra cui un'immagine di Santa Barbara con un'ambientazione del XVII o XVIII secolo, una di Santa Caterina da Siena in stile rococò di Szymon Czechowicz, un ritratto di Alessandro Jagellone di un artista sconosciuto della seconda metà del XVIII secolo). Un ingresso sotto l'altare conduce alle grandi volte, labirintiche, con molte stanze e cripte: i sotterranei ospitano i resti di centinaia di residenti di Vilnius, alcuni dei quali mummificatisi naturalmente, e sono circondati da leggende metropolitane. Sebbene l'esistenza dei sotterranei fosse nota, i primi sforzi per esplorare e mappare le cripte furono abbandonate nonostante lo sforzo degli studenti dell'Università di Vilnius negli anni '30. Tuttavia, questi ultimi non avevano osservato le corrette procedure archeologiche e causarono infatti molti danni: il modus operandi prevedeva lo smistamento delle ossa ponendo tutti i teschi sugli scaffali e rimuovendoli le tombe. Da allora, i resti sono stati spostati molte volte lasciandoli in uno stato casuale e disorganizzato. Stando alle leggende che aleggiano sul luogo, i resti sarebbero di soldati francesi recatisi in città nel corso della campagna di Russia del 1812 avviata da Napoleone Bonaparte, di vittime dell'Inquisizione o della peste nera. Più romantiche risultano le affermazioni di chi sostiene che i corridoi sotterranei facevano parte di una rete di passaggi più ampia che consentiva agli amanti leggendari Barbara Radziwiłł e Sigismondo II Augusto di incontrarsi in segreto. Nel 2011, gli antropologi dell'Università di Vilnius, guidati da Rimantas Jankauskas, avviarono uno studio sui corpi mummificati, stimando settimane dopo che le volte conservassero i resti di circa 600 persone, tra cui molte donne e bambini dalla metà del XVIII secolo all'inizio del XIX secolo. Il team ha selezionato i cadaveri meglio conservati e ha eseguito la loro tomografia. I risultati mostrano che molte persone erano in sovrappeso e avevano l'alluce valgo, il che ha portato alla conclusione che si trattava di alti borghesi o comunque di cittadini abbienti. "
- text: "Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. "
- text: "Vanni ha la sua prima mostra personale nel 1948, alla Galleria Margherita di Roma. Nel 1949 vince una borsa di studio che lo porterà a studiare ad Amsterdam sotto la guida del pittore neoplastico Friedrich Vordemberge-Gildewart. Nel 1952 vince una Fulbright Scholarship che lo porterà a studiare in America, alla Yale University, sotto la guida di Josef Albers. Dal 1953 al 1960 si stabilisce a Parigi, dove illustra alcuni libri per bambini che in seguito vinceranno il premio del Club des Editeurs. Nel 1954 lavora come consulente del colore per il documentario su Picasso di Luciano Emmer, e nel 1955 comincia la sua lunga collaborazione con la Galleria Schneider, affiancando artisti come Corrado Cagli. Dal 1969 al 1974 lavora su dei bassorilievi in vetro resina sui quali vengono proiettati dei film astratti da lui creati, per creare dei quadri che si trasformino continuamente nel tempo. Nel 1979 lascia Roma per stabilirsi a New York, dove alla carriera di pittore affiancherà quella di professore per la prestigiosa Cooper Union School of Art, dove insegnerà ininterrottamente dal 1984 al 2014. L'opera pittorica di Vanni è segnata da una visione estremamente personale, lontana dalle correnti e dai movimenti che hanno caratterizzato la seconda metà del XX secolo. Memore delle lunghe conversazioni avute da Vanni nella sua primissima gioventù, con il filosofo e pittore futurista Alberto Bragaglia, le sue opere sono contrassegnate da un “eclettismo” formale programmatico, alla base del quale resta costante una conoscenza profonda delle molteplici tecniche artistiche utilizzate (tra cui il mosaico, l’affresco e la tempera ad uovo). Pur esprimendosi per lo più in cicli di opere dove l’astrazione formale è la principale componente figurativa, sono da sottolineare alcune opere dove Vanni ha dato prova di una importante padronanza dell’arte figurativa. Importanti e numerose sono le sue realizzazioni anche nel campo dell’illustrazione. Sue sono le illustrazioni per la novella ''Agostino'' di Alberto Moravia, per il libro ''Love'' di Lowell A. Siff e delle ''Contes de Cristal'' di Alice Coléno. Ha tenuto mostre personali in Italia e all’estero ed esposto in mostre collettive di rappresentanza italiana nei musei e nelle gallerie di ogni parte del mondo. "
metrics:
- rouge
- bertscore
model-index:
- name: it5-base-wiki-summarization
results:
- task:
type: wiki-summarization
name: "Wikipedia Summarization"
dataset:
type: wits
name: "WITS"
metrics:
- type: rouge1
value: 0.369
name: "Test Rouge1"
- type: rouge2
value: 0.217
name: "Test Rouge2"
- type: rougeL
value: 0.333
name: "Test RougeL"
- type: bertscore
value: 0.530
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Base for Wikipedia Summarization 📑 🇮🇹
This repository contains the checkpoint for the [IT5 Base](https://huggingface.co/gsarti/it5-base) model fine-tuned on Wikipedia summarization on the [WITS](https://www.semanticscholar.org/paper/WITS%3A-Wikipedia-for-Italian-Text-Summarization-Casola-Lavelli/ad6c83122e721c7c0db4a40727dac3b4762cd2b1) dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
hg = pipeline("text2text-generation", model='it5/it5-base-wiki-summarization')
hg("Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. "
- text: "Vanni ha la sua prima mostra personale nel 1948, alla Galleria Margherita di Roma. Nel 1949 vince una borsa di studio che lo porterà a studiare ad Amsterdam sotto la guida del pittore neoplastico Friedrich Vordemberge-Gildewart. Nel 1952 vince una Fulbright Scholarship che lo porterà a studiare in America, alla Yale University, sotto la guida di Josef Albers. Dal 1953 al 1960 si stabilisce a Parigi, dove illustra alcuni libri per bambini che in seguito vinceranno il premio del Club des Editeurs. Nel 1954 lavora come consulente del colore per il documentario su Picasso di Luciano Emmer, e nel 1955 comincia la sua lunga collaborazione con la Galleria Schneider, affiancando artisti come Corrado Cagli. Dal 1969 al 1974 lavora su dei bassorilievi in vetro resina sui quali vengono proiettati dei film astratti da lui creati, per creare dei quadri che si trasformino continuamente nel tempo. Nel 1979 lascia Roma per stabilirsi a New York, dove alla carriera di pittore affiancherà quella di professore per la prestigiosa Cooper Union School of Art, dove insegnerà ininterrottamente dal 1984 al 2014. L'opera pittorica di Vanni è segnata da una visione estremamente personale, lontana dalle correnti e dai movimenti che hanno caratterizzato la seconda metà del XX secolo. Memore delle lunghe conversazioni avute da Vanni nella sua primissima gioventù, con il filosofo e pittore futurista Alberto Bragaglia, le sue opere sono contrassegnate da un “eclettismo” formale programmatico, alla base del quale resta costante una conoscenza profonda delle molteplici tecniche artistiche utilizzate (tra cui il mosaico, l’affresco e la tempera ad uovo). Pur esprimendosi per lo più in cicli di opere dove l’astrazione formale è la principale componente figurativa, sono da sottolineare alcune opere dove Vanni ha dato prova di una importante padronanza dell’arte figurativa. Importanti e numerose sono le sue realizzazioni anche nel campo dell’illustrazione. Sue sono le illustrazioni per la novella ''Agostino'' di Alberto Moravia, per il libro ''Love'' di Lowell A. Siff e delle ''Contes de Cristal'' di Alice Coléno. Ha tenuto mostre personali in Italia e all’estero ed esposto in mostre collettive di rappresentanza italiana nei musei e nelle gallerie di ogni parte del mondo.")
>>> [{"generated_text": "L' '''isola di Rabot''' si trova in prossimità dell'isola di Renaud, a sud dell'Argentina."}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-base-wiki-summarization")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-base-wiki-summarization")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-base-question-generation
|
gsarti
| 2022-03-09T08:06:11Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"question-generation",
"squad_it",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- question-generation
- squad_it
- text2text-generation
widget:
- text: "Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una \"grande pestilenza nell' aria\". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola \"peste\" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia"
- text: "Il 14 aprile 2011, ABC ha annullato le lunghe opere di sapone All My Children e One Life to Live dopo 41 e 43 anni in onda, rispettivamente (in seguito al contraccolpo dei tifosi, ABC ha venduto i diritti ad entrambi gli spettacoli a Prospect Park, che alla fine ha rilanciato i saponi su Hulu per un' ulteriore stagione nel 2013 e con entrambe le società che si citano in giudizio per accuse di interferenza con il processo di rilancio degli spettacoli, mancato pagamento delle tasse di licenza. Il talk/lifestyle show che ha sostituito One Life to Live, The Revolution, non è riuscito a generare giudizi soddisfacenti ed è stato a sua volta annullato dopo soli sette mesi. La stagione 2011-12 ha visto l' ABC cadere al quarto posto nel 18-49 demografico nonostante rinnovando una manciata di nuovi spettacoli (compresi i drammi matricole Scandal, Revenge e Once Upon a Time) per la seconda stagione. Risposta: Hulu"
- text: "L' American Broadcasting Company (ABC) (stlized nel suo logo come abc dal 1957) è una rete televisiva commerciale americana trasmissione televisiva che è di proprietà del Disney-ABC Television Group, una controllata della divisione Disney Media Networks di The Walt Disney Company. La rete fa parte delle grandi reti televisive Big Three. La rete ha sede a Columbus Avenue e West 66th Street a Manhattan, con ulteriori uffici e stabilimenti di produzione a New York City, Los Angeles e Burbank, California. Risposta: Manhattan"
- text: "La disobbedienza civile non rivoluzionaria è una semplice disobbedienza delle leggi sulla base del fatto che sono giudicate \"sbagliate\" da una coscienza individuale, o come parte di uno sforzo per rendere alcune leggi inefficaci, per causarne l' abrogazione, o per esercitare pressioni per ottenere i propri desideri politici su qualche altra questione. La disobbedienza civile rivoluzionaria è più che altro un tentativo attivo di rovesciare un governo (o di cambiare le tradizioni culturali, i costumi sociali, le credenze religiose, ecc. La rivoluzione non deve necessariamente essere politica, cioè \"rivoluzione culturale\", implica semplicemente un cambiamento radicale e diffuso in una sezione del tessuto sociale). Gli atti di Gandhi sono stati descritti come disobbedienza civile rivoluzionaria. È stato affermato che gli ungheresi sotto Ferenc Deák hanno diretto una disobbedienza civile rivoluzionaria contro il governo austriaco. Thoreau ha anche scritto di disobbedienza civile realizzando \"rivoluzione pacifica\". Howard Zinn, Harvey Wheeler e altri hanno identificato il diritto sposato nella Dichiarazione d' Indipendenza di \"alterare o abolire\" un governo ingiusto come principio di disobbedienza civile. Risposta: Ferenc Deák"
metrics:
- rouge
- bertscore
model-index:
- name: it5-base-question-generation
results:
- task:
type: question-generation
name: "Question generation"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: rouge1
value: 0.382
name: "Test Rouge1"
- type: rouge2
value: 0.199
name: "Test Rouge2"
- type: rougeL
value: 0.354
name: "Test RougeL"
- type: bertscore
value: 0.516
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Base for Question Generation 💭 🇮🇹
This repository contains the checkpoint for the [IT5 Base](https://huggingface.co/gsarti/it5-base) model fine-tuned on question generation on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qg = pipeline("text2text-generation", model='it5/it5-base-question-generation')
qg("Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una "grande pestilenza nell\' aria". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola "peste" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia")
>>> [{"generated_text": "Per chi è stato redatto il referto medico?"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-base-question-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-base-question-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-base-question-answering
|
gsarti
| 2022-03-09T08:05:47Z | 25 | 2 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"squad_it",
"text2text-question-answering",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- squad_it
- text2text-question-answering
- text2text-generation
widget:
- text: "In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?"
- text: "L' embargo non era uniforme in tutta Europa. Dei nove membri della Comunità Economica Europea (CEE), i Paesi Bassi hanno dovuto affrontare un embargo totale, il Regno Unito e la Francia hanno ricevuto forniture quasi ininterrotte (poichè si sono rifiutati di consentire all' America di utilizzare i loro aerodromi e le armi e forniture embargo sia agli arabi che agli israeliani), mentre gli altri sei hanno dovuto affrontare tagli parziali. Il Regno Unito era tradizionalmente un alleato di Israele, e il governo di Harold Wilson ha sostenuto gli israeliani durante la guerra dei sei giorni. Il suo successore, Ted Heath, ribaltò questa politica nel 1970, chiedendo a Israele di ritirarsi ai suoi confini prima del 1967. Domanda: Il Regno Unito e la Francia non hanno avuto interruzioni dell' approvvigionamento petrolifero in quanto non hanno consentito a quale paese di utilizzare il loro aeroporto?"
- text: "Nel 1962, il grafico Paul Rand ridisegna il logo ABC nella sua forma più conosciuta (e attuale) con le lettere minuscole \"abc\" racchiuse in un unico cerchio nero. Il nuovo logo esordisce in onda per le promozioni di ABC all' inizio della stagione 1963-64. Le lettere ricordano fortemente il carattere tipografico Bauhaus disegnato da Herbert Bayer negli anni Venti, ma condividono anche similitudini con diversi altri caratteri, come ITC Avant Garde e Horatio, e lo Chalet più simile. La semplicità del logo ha reso più facile la riprogettazione e la duplicazione, il che ha conferito un beneficio per ABC (soprattutto prima dell' avvento della computer grafica). Domanda: Di quale carattere tipografico ricordano le lettere dell' iconico logo ABC?"
- text: "La fotorespirazione può verificarsi quando la concentrazione di ossigeno è troppo elevata. Rubisco non è in grado di distinguere molto bene tra ossigeno e anidride carbonica, quindi può accidentalmente aggiungere O2 invece di CO2 a RuBP. Questo processo riduce l' efficienza della fotosintesi: consuma ATP e ossigeno, rilascia CO2 e non produce zucchero. Può sprecare fino alla metà del carbonio fissato dal ciclo di Calvin. Diversi meccanismi si sono evoluti in diversi lignaggi che aumentano la concentrazione di anidride carbonica rispetto all' ossigeno all' interno del cloroplasto, aumentando l' efficienza della fotosintesi. Questi meccanismi sono chiamati meccanismi di concentrazione dell' anidride carbonica, o CCM. Tra questi figurano il metabolismo degli acidi crassulaceanici, la fissazione del carbonio C4 e i pirenoidi. I cloroplasti negli impianti C4 sono notevoli in quanto presentano un chiaro dimorfismo cloroplastico. Domanda: Che cosa può fare rubisco per errore?"
metrics:
- f1
- exact-match
model-index:
- name: it5-base-question-answering
results:
- task:
type: question-answering
name: "Question Answering"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: f1
value: 0.761
name: "Test F1"
- type: exact-match
value: 0.663
name: "Test Exact Match"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Base for Question Answering ⁉️ 🇮🇹
This repository contains the checkpoint for the [IT5 Base](https://huggingface.co/gsarti/it5-base) model fine-tuned on extractive question answering on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qa = pipeline("text2text-generation", model='it5/it5-base-question-answering')
qa("In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?")
>>> [{"generated_text": "ultimo massimo glaciale"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-base-question-answering")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-base-question-answering")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-base-repubblica-to-ilgiornale
|
gsarti
| 2022-03-09T08:05:15Z | 7 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: it5-base-repubblica-to-ilgiornale
results:
- task:
type: headline-style-transfer-repubblica-to-ilgiornale
name: "Headline style transfer (Repubblica to Il Giornale)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.272
name: "Test Rouge1"
- type: rouge2
value: 0.089
name: "Test Rouge2"
- type: rougeL
value: 0.235
name: "Test RougeL"
- type: bertscore
value: 0.396
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.883
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.880
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Base for News Headline Style Transfer (Repubblica to Il Giornale) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Base](https://huggingface.co/gsarti/it5-base) model fine-tuned on news headline style transfer in the Repubblica to Il Giornale direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Il Giornale from the full body of an article written in the style of Repubblica. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
r2g = pipeline("text2text-generation", model='it5/it5-base-repubblica-to-ilgiornale')
r2g("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-base-repubblica-to-ilgiornale")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-base-repubblica-to-ilgiornale")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-small-ilgiornale-to-repubblica
|
gsarti
| 2022-03-09T08:03:52Z | 7 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: it5-small-ilgiornale-to-repubblica
results:
- task:
type: headline-style-transfer-ilgiornale-to-repubblica
name: "Headline style transfer (Il Giornale to Repubblica)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.270
name: "Test Rouge1"
- type: rouge2
value: 0.092
name: "Test Rouge2"
- type: rougeL
value: 0.239
name: "Test RougeL"
- type: bertscore
value: 0.404
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.909
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.869
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "8g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Small for News Headline Style Transfer (Il Giornale to Repubblica) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Small](https://huggingface.co/gsarti/it5-small) model fine-tuned on news headline style transfer in the Il Giornale to Repubblica direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Repubblica from the full body of an article written in the style of Il Giornale. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
g2r = pipeline("text2text-generation", model='it5/it5-small-ilgiornale-to-repubblica')
g2r("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-small-ilgiornale-to-repubblica")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-small-ilgiornale-to-repubblica")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-ilgiornale-to-repubblica
|
gsarti
| 2022-03-09T08:02:59Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: mt5-base-ilgiornale-to-repubblica
results:
- task:
type: headline-style-transfer-ilgiornale-to-repubblica
name: "Headline style transfer (Il Giornale to Repubblica)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.282
name: "Test Rouge1"
- type: rouge2
value: 0.101
name: "Test Rouge2"
- type: rougeL
value: 0.248
name: "Test RougeL"
- type: bertscore
value: 0.411
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.815
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.773
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Base for News Headline Style Transfer (Il Giornale to Repubblica) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [mT5 Base](https://huggingface.co/google/mt5-base) model fine-tuned on news headline style transfer in the Il Giornale to Repubblica direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Repubblica from the full body of an article written in the style of Il Giornale. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
g2r = pipeline("text2text-generation", model='it5/mt5-base-ilgiornale-to-repubblica')
g2r("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-base-ilgiornale-to-repubblica")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-base-ilgiornale-to-repubblica")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-small-repubblica-to-ilgiornale
|
gsarti
| 2022-03-09T08:02:27Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: it5-small-repubblica-to-ilgiornale
results:
- task:
type: headline-style-transfer-repubblica-to-ilgiornale
name: "Headline style transfer (Repubblica to Il Giornale)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.255
name: "Test Rouge1"
- type: rouge2
value: 0.080
name: "Test Rouge2"
- type: rougeL
value: 0.223
name: "Test RougeL"
- type: bertscore
value: 0.380
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.887
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.894
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "8g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Small for News Headline Style Transfer (Repubblica to Il Giornale) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Small](https://huggingface.co/gsarti/it5-small) model fine-tuned on news headline style transfer in the Repubblica to Il Giornale direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Il Giornale from the full body of an article written in the style of Repubblica. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
r2g = pipeline("text2text-generation", model='it5/it5-small-repubblica-to-ilgiornale')
r2g("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-small-repubblica-to-ilgiornale")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-small-repubblica-to-ilgiornale")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-large-repubblica-to-ilgiornale
|
gsarti
| 2022-03-09T08:01:50Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: it5-large-repubblica-to-ilgiornale
results:
- task:
type: headline-style-transfer-repubblica-to-ilgiornale
name: "Headline style transfer (Repubblica to Il Giornale)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.270
name: "Test Rouge1"
- type: rouge2
value: 0.089
name: "Test Rouge2"
- type: rougeL
value: 0.237
name: "Test RougeL"
- type: bertscore
value: 0.400
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.883
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.880
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "51g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Large for News Headline Style Transfer (Repubblica to Il Giornale) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Large](https://huggingface.co/gsarti/it5-large) model fine-tuned on news headline style transfer in the Repubblica to Il Giornale direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Il Giornale from the full body of an article written in the style of Repubblica. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
r2g = pipeline("text2text-generation", model='it5/it5-large-repubblica-to-ilgiornale')
r2g("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-large-repubblica-to-ilgiornale")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-large-repubblica-to-ilgiornale")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-small-repubblica-to-ilgiornale
|
gsarti
| 2022-03-09T08:01:24Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: mt5-small-repubblica-to-ilgiornale
results:
- task:
type: headline-style-transfer-repubblica-to-ilgiornale
name: "Headline style transfer (Repubblica to Il Giornale)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.231
name: "Test Rouge1"
- type: rouge2
value: 0.069
name: "Test Rouge2"
- type: rougeL
value: 0.204
name: "Test RougeL"
- type: bertscore
value: 0.363
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.769
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.842
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Small for News Headline Style Transfer (Repubblica to Il Giornale) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [mT5 Small](https://huggingface.co/google/mt5-small) model fine-tuned on news headline style transfer in the Repubblica to Il Giornale direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Il Giornale from the full body of an article written in the style of Repubblica. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
r2g = pipeline("text2text-generation", model='it5/mt5-small-repubblica-to-ilgiornale')
r2g("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-small-repubblica-to-ilgiornale")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-small-repubblica-to-ilgiornale")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-repubblica-to-ilgiornale
|
gsarti
| 2022-03-09T08:00:57Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"style-transfer",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- style-transfer
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
- headline-headline-consistency-classifier
- headline-article-consistency-classifier
model-index:
- name: mt5-base-repubblica-to-ilgiornale
results:
- task:
type: headline-style-transfer-repubblica-to-ilgiornale
name: "Headline style transfer (Repubblica to Il Giornale)"
dataset:
type: gsarti/change_it
name: "CHANGE-IT"
metrics:
- type: rouge1
value: 0.256
name: "Test Rouge1"
- type: rouge2
value: 0.080
name: "Test Rouge2"
- type: rougeL
value: 0.223
name: "Test RougeL"
- type: bertscore
value: 0.384
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: headline-headline-consistency-classifier
value: 0.775
name: "Test Headline-Headline Consistency Accuracy"
- type: headline-article-consistency-classifier
value: 0.825
name: "Test Headline-Article Consistency Accuracy"
co2_eq_emissions:
emissions: "40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Small for News Headline Style Transfer (Repubblica to Il Giornale) 🗞️➡️🗞️ 🇮🇹
This repository contains the checkpoint for the [mT5 Small](https://huggingface.co/google/mt5-small) model fine-tuned on news headline style transfer in the Repubblica to Il Giornale direction on the Italian CHANGE-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
The model is trained to generate an headline in the style of Il Giornale from the full body of an article written in the style of Repubblica. Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
r2g = pipeline("text2text-generation", model='it5/mt5-small-repubblica-to-ilgiornale')
r2g("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-small-repubblica-to-ilgiornale")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-small-repubblica-to-ilgiornale")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-small-headline-generation
|
gsarti
| 2022-03-09T08:00:22Z | 17 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"headline-generation",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- headline-generation
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
model-index:
- name: it5-small-headline-generation
results:
- task:
type: headline-generation
name: "Headline generation"
dataset:
type: headgen_it
name: "HeadGen-IT"
metrics:
- type: rouge1
value: 0.287
name: "Test Rouge1"
- type: rouge2
value: 0.100
name: "Test Rouge2"
- type: rougeL
value: 0.253
name: "Test RougeL"
- type: bertscore
value: 0.414
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "8g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Small for News Headline Generation 📣 🇮🇹
This repository contains the checkpoint for the [IT5 Small](https://huggingface.co/gsarti/it5-small) model fine-tuned on news headline generation on the Italian HeadGen-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
hg = pipeline("text2text-generation", model='it5/it5-small-headline-generation')
hg("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-small-headline-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-small-headline-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-large-headline-generation
|
gsarti
| 2022-03-09T07:59:47Z | 7 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"newspaper",
"ilgiornale",
"repubblica",
"headline-generation",
"it",
"dataset:gsarti/change_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- gsarti/change_it
tags:
- italian
- sequence-to-sequence
- newspaper
- ilgiornale
- repubblica
- headline-generation
widget:
- text: "WASHINGTON - La Corea del Nord torna dopo nove anni nella blacklist Usa degli Stati considerati sponsor del terrorismo. Come Iran, Siria e Sudan. Lo ha deciso Donald Trump , che ha preferito dare l'annuncio non durante il suo recente viaggio in Asia ma ieri, in una riunione del governo alla Casa Bianca. 'Oggi gli Stati Uniti designeranno la Corea del nord come uno stato sponsor del terrorismo', ha tuonato il tycoon, anticipando che sarà formalizzata oggi dal dipartimento di stato e sarà accompagnata da nuove e più severe sanzioni. 'Il livello più alto' mai imposto a Pyongyang, ha promesso. 'Avrebbe dovuto succedere molto tempo fa', ha aggiunto, scaricando per l'ennesima volta la responsabilità dell'attuale crisi sull'amministrazione Obama. Poi si è scagliato contro un 'regime assassino' che 'deve mettere fine allo sviluppo del suo programma illegale nucleare e balistico'. Per giustificare la svolta, Trump ha accusato Pyongyang non solo di 'minacciare il mondo con una devastazione nucleare' ma anche di aver 'ripetutamente sostenuto atti di terrorismo internazionale', compreso omicidi in suolo straniero. Il riferimento è all' uccisione all'aeroporto della capitale malese di Kim Jong Nam , il fratellastro del leader nordcoreano Kim Jong Un , ma non ci sono altri episodi noti. Tanto che alcuni esperti, come pure dirigenti Usa coperti dall'anonimato, dubitano che Pyongyang risponda ai criteri per una tale designazione. La mossa appare altamente simbolica, dato che la Corea del Nord è già pesantemente sanzionata a livello internazionale. Per il segretario di stato Rex Tillerson è solo l'ultima di una serie di passi per rafforzare la pressione su Pyongyang e costringerla a sedersi ad un tavolo perché gli Usa hanno sempre 'speranza nella diplomazia'. Ma nello stesso tempo è un monito per 'fermare e dissuadere' altri Paesi dal sostenere la Corea del Nord, finita nella blacklist 'anche per l'uso di armi chimiche'. Ma la mossa potrebbe anche essere controproducente, provocando una risposta di Kim o minando gli sforzi per sollecitare Pechino ad una maggiore pressione su Pyongyang. In ogni caso non aiuta il dialogo diretto tra Usa e Corea del Nord, che sembrava essere stato avviato in modo riservato. Come non aiutano gli scambi di insulti fra Trump e Kim. Nord Corea, Trump: 'Cerco di essere amico di Kim, sarebbe una bella cosa per il mondo'. Pyongyang era stata messa nella lista Usa degli Stati sponsor del terrorismo per aver fatto esplodere nel 1987 un volo della Korean Air uccidendo tutti i 115 passeggeri a bordo. Ma l'amministrazione di George W. Bush l'aveva rimossa sperando di far avanzare i negoziati sulla denuclearizzazione della penisola coreana. Il governo giapponese sostiene la decisione degli Stati Uniti di inserire la Corea del Nord nella lista degli stati che sponsorizzano il terrorismo, pur riconoscendo che l'annuncio potrebbe provocare una reazione immediata del regime di Pyongyang. Il premier Shinzo Abe ha accolto con consenso il comunicato Usa e ha detto alla stampa che servirà a incrementare la pressione sulla Corea del Nord. Il ministro della Difesa Itsunori Onodera , pur valutando positivamente la notifica, ha spiegato che si attendono azioni provocatorie dallo stato eremita, ribadendo che è vitale rimanere vigili. Secondo la stampa nipponica Abe aveva richiesto al dipartimento di Stato Usa di mettere la Corea del Nord sulla lista durante l'incontro col presidente Usa Donald Trump a Tokyo a inizio mese. L'ultimo lancio di missile balistico condotto da Pyongyang nell'oceano Pacifico, sorvolando il mare del Giappone, risale allo scorso settembre."
- text: "ROMA - Una nuova droga killer è stata sequestrata per la prima volta in Europa dagli investigatori del Nas. Si tratta di una nuova \"miscela psicoattiva altamente tossica\" per la prima volta individuata da forze di polizia, simile all'eroina sintetica, ma molto più economica e letale. Tanto che i 20 grammi scoperti sarebbero stati sufficienti per fabbricare ben 20.000 dosi e lo stesso contatto attraverso la pelle può provocare intossicazione. Individuata per la prima volta, la nuova droga presenta una struttura simile al farmaco sedativo Fentanyl ma con effetti molto più devastanti per l'organismo. Proveniva dell'estero ed era contenuta in un plico postale indirizzato in una città del centro Italia: è stata intercettata tramite accertamenti sul web grazie a un'operazione di intelligence che ha visto come protagonisti i militari della Sezione operativa centrale del Comando carabinieri per la Tutela della salute (Nas). Economica e letale, secondo gli investigatori \"in confronto l'eroina è quasi 'acqua fresca', anzi, proprio per la sua economicità, in alcuni casi viene venduta dai pusher a giovani conviti di comprare eroina\". La diffusione di nuove droghe sintetiche che continuamente appaiono sui mercati necessita di un'attività investigativa costante e complessa. Si tratta infatti di sostanze dalla struttura molecolare molto simile a quella del Fentanyl ma ogni volta leggermente diversa. Di qui la difficoltà di individuarle e l'importanza del nuovo sequestro. \"La chiamano impropriamente 'eroina sintetica' - spiega il comandante dei Nas, generale Adelmo Lusi - per il tipo di effetto psicotropo simile, ma dal punto di vista della tossicità è molto peggio: con 25 milligrammi di eroina ci si sballa, con 25mg di simil-fentanyl, come quello appena sequestrato, si muore\". Le indagini sono partite da ricoveri per overdose in ospedale, in cui arrivavano ragazzi che non rispondevano al trattamento disintossicante per l'eroina. La nuova sostanza verrà ora segnalata per l'inserimento tra le tabelle ministeriali degli stupefacenti prevista dal Dpr 309/1990."
- text: "Fragile come il burro. Il nostro territorio è precario. Ne sanno qualcosa i comuni che sono stati investititi dal maltempo . Il dissesto idrogeologico imperversa su tutto il territorio. Infatti, oltre 6.600 comuni , pari all’82% del totale, sono in aree ad elevato rischio idrogeologico, pari al 10% della sua superficie. La popolazione potenzialmente esposta è stimata in 5,8 milioni di persone. I dati emergono dalle recenti analisi fatte da Legambiente e Protezione civile, che mettono in evidenza come in 10 anni in Italia sia raddoppiata l’area dei territori colpiti da alluvioni e frane , passando da una media di quattro regioni all’anno a otto regioni. Nella classifica delle regioni a maggior rischio idrogeologico prima è la Calabria con il 100% dei comuni esposti; al 100% ci sono anche la provincia di Trento, il Molise, la Basilicata, l’Umbria, la Valle d’Aosta. Poi Marche, Liguria al 99%; Lazio, Toscana al 98%; Abruzzo (96%), Emilia-Romagna (95%), Campania e Friuli Venezia Giulia al 92%, Piemonte (87%), Sardegna (81%), Puglia (78%), Sicilia (71%), Lombardia (60%), provincia di Bolzano (59%), Veneto (56%). Tra le cause che condizionano ed amplificano il rischio idrogeologico c’è l’azione dell’uomo (abbandono e degrado, cementificazione, consumo di suolo, abusivismo, disboscamento e incendi). Ma anche e soprattutto la mancanza di una seria manutenzione ordinaria e non ad una organica politica di prevenzione."
- text: "Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\"."
metrics:
- rouge
- bertscore
model-index:
- name: it5-large-headline-generation
results:
- task:
type: headline-generation
name: "Headline generation"
dataset:
type: headgen_it
name: "HeadGen-IT"
metrics:
- type: rouge1
value: 0.308
name: "Test Rouge1"
- type: rouge2
value: 0.113
name: "Test Rouge2"
- type: rougeL
value: 0.270
name: "Test RougeL"
- type: bertscore
value: 0.430
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "51g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Large for News Headline Generation 📣 🇮🇹
This repository contains the checkpoint for the [IT5 Large](https://huggingface.co/gsarti/it5-large) model fine-tuned on news headline generation on the Italian HeadGen-IT dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
hg = pipeline("text2text-generation", model='it5/it5-large-headline-generation')
hg("Arriva dal Partito nazionalista basco (Pnv) la conferma che i cinque deputati che siedono in parlamento voteranno la sfiducia al governo guidato da Mariano Rajoy. Pochi voti, ma significativi quelli della formazione politica di Aitor Esteban, che interverrà nel pomeriggio. Pur con dimensioni molto ridotte, il partito basco si è trovato a fare da ago della bilancia in aula. E il sostegno alla mozione presentata dai Socialisti potrebbe significare per il primo ministro non trovare quei 176 voti che gli servono per continuare a governare. \" Perché dovrei dimettermi io che per il momento ho la fiducia della Camera e quella che mi è stato data alle urne \", ha detto oggi Rajoy nel suo intervento in aula, mentre procedeva la discussione sulla mozione di sfiducia. Il voto dei baschi ora cambia le carte in tavola e fa crescere ulteriormente la pressione sul premier perché rassegni le sue dimissioni. La sfiducia al premier, o un'eventuale scelta di dimettersi, porterebbe alle estreme conseguenze lo scandalo per corruzione che ha investito il Partito popolare. Ma per ora sembra pensare a tutt'altro. \"Non ha intenzione di dimettersi - ha detto il segretario generale del Partito popolare , María Dolores de Cospedal - Non gioverebbe all'interesse generale o agli interessi del Pp\".")
>>> [{"generated_text": "il nazionalista rajoy: 'voteremo la sfiducia'"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-large-headline-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-large-headline-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-large-question-answering
|
gsarti
| 2022-03-09T07:57:53Z | 18 | 5 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"squad_it",
"text2text-question-answering",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- squad_it
- text2text-question-answering
- text2text-generation
widget:
- text: "In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?"
- text: "L' embargo non era uniforme in tutta Europa. Dei nove membri della Comunità Economica Europea (CEE), i Paesi Bassi hanno dovuto affrontare un embargo totale, il Regno Unito e la Francia hanno ricevuto forniture quasi ininterrotte (poichè si sono rifiutati di consentire all' America di utilizzare i loro aerodromi e le armi e forniture embargo sia agli arabi che agli israeliani), mentre gli altri sei hanno dovuto affrontare tagli parziali. Il Regno Unito era tradizionalmente un alleato di Israele, e il governo di Harold Wilson ha sostenuto gli israeliani durante la guerra dei sei giorni. Il suo successore, Ted Heath, ribaltò questa politica nel 1970, chiedendo a Israele di ritirarsi ai suoi confini prima del 1967. Domanda: Il Regno Unito e la Francia non hanno avuto interruzioni dell' approvvigionamento petrolifero in quanto non hanno consentito a quale paese di utilizzare il loro aeroporto?"
- text: "Nel 1962, il grafico Paul Rand ridisegna il logo ABC nella sua forma più conosciuta (e attuale) con le lettere minuscole \"abc\" racchiuse in un unico cerchio nero. Il nuovo logo esordisce in onda per le promozioni di ABC all' inizio della stagione 1963-64. Le lettere ricordano fortemente il carattere tipografico Bauhaus disegnato da Herbert Bayer negli anni Venti, ma condividono anche similitudini con diversi altri caratteri, come ITC Avant Garde e Horatio, e lo Chalet più simile. La semplicità del logo ha reso più facile la riprogettazione e la duplicazione, il che ha conferito un beneficio per ABC (soprattutto prima dell' avvento della computer grafica). Domanda: Di quale carattere tipografico ricordano le lettere dell' iconico logo ABC?"
- text: "La fotorespirazione può verificarsi quando la concentrazione di ossigeno è troppo elevata. Rubisco non è in grado di distinguere molto bene tra ossigeno e anidride carbonica, quindi può accidentalmente aggiungere O2 invece di CO2 a RuBP. Questo processo riduce l' efficienza della fotosintesi: consuma ATP e ossigeno, rilascia CO2 e non produce zucchero. Può sprecare fino alla metà del carbonio fissato dal ciclo di Calvin. Diversi meccanismi si sono evoluti in diversi lignaggi che aumentano la concentrazione di anidride carbonica rispetto all' ossigeno all' interno del cloroplasto, aumentando l' efficienza della fotosintesi. Questi meccanismi sono chiamati meccanismi di concentrazione dell' anidride carbonica, o CCM. Tra questi figurano il metabolismo degli acidi crassulaceanici, la fissazione del carbonio C4 e i pirenoidi. I cloroplasti negli impianti C4 sono notevoli in quanto presentano un chiaro dimorfismo cloroplastico. Domanda: Che cosa può fare rubisco per errore?"
metrics:
- f1
- exact-match
model-index:
- name: it5-large-question-answering
results:
- task:
type: question-answering
name: "Question Answering"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: f1
value: 0.780
name: "Test F1"
- type: exact-match
value: 0.691
name: "Test Exact Match"
co2_eq_emissions:
emissions: 51g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Large for Question Answering ⁉️ 🇮🇹
This repository contains the checkpoint for the [IT5 Large](https://huggingface.co/gsarti/it5-large) model fine-tuned on extractive question answering on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qa = pipeline("text2text-generation", model='it5/it5-large-question-answering')
qa("In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?")
>>> [{"generated_text": "ultimo massimo glaciale"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-large-question-answering")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-large-question-answering")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-question-answering
|
gsarti
| 2022-03-09T07:57:29Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"squad_it",
"text2text-question-answering",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- squad_it
- text2text-question-answering
- text2text-generation
widget:
- text: "In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?"
- text: "L' embargo non era uniforme in tutta Europa. Dei nove membri della Comunità Economica Europea (CEE), i Paesi Bassi hanno dovuto affrontare un embargo totale, il Regno Unito e la Francia hanno ricevuto forniture quasi ininterrotte (poichè si sono rifiutati di consentire all' America di utilizzare i loro aerodromi e le armi e forniture embargo sia agli arabi che agli israeliani), mentre gli altri sei hanno dovuto affrontare tagli parziali. Il Regno Unito era tradizionalmente un alleato di Israele, e il governo di Harold Wilson ha sostenuto gli israeliani durante la guerra dei sei giorni. Il suo successore, Ted Heath, ribaltò questa politica nel 1970, chiedendo a Israele di ritirarsi ai suoi confini prima del 1967. Domanda: Il Regno Unito e la Francia non hanno avuto interruzioni dell' approvvigionamento petrolifero in quanto non hanno consentito a quale paese di utilizzare il loro aeroporto?"
- text: "Nel 1962, il grafico Paul Rand ridisegna il logo ABC nella sua forma più conosciuta (e attuale) con le lettere minuscole \"abc\" racchiuse in un unico cerchio nero. Il nuovo logo esordisce in onda per le promozioni di ABC all' inizio della stagione 1963-64. Le lettere ricordano fortemente il carattere tipografico Bauhaus disegnato da Herbert Bayer negli anni Venti, ma condividono anche similitudini con diversi altri caratteri, come ITC Avant Garde e Horatio, e lo Chalet più simile. La semplicità del logo ha reso più facile la riprogettazione e la duplicazione, il che ha conferito un beneficio per ABC (soprattutto prima dell' avvento della computer grafica). Domanda: Di quale carattere tipografico ricordano le lettere dell' iconico logo ABC?"
- text: "La fotorespirazione può verificarsi quando la concentrazione di ossigeno è troppo elevata. Rubisco non è in grado di distinguere molto bene tra ossigeno e anidride carbonica, quindi può accidentalmente aggiungere O2 invece di CO2 a RuBP. Questo processo riduce l' efficienza della fotosintesi: consuma ATP e ossigeno, rilascia CO2 e non produce zucchero. Può sprecare fino alla metà del carbonio fissato dal ciclo di Calvin. Diversi meccanismi si sono evoluti in diversi lignaggi che aumentano la concentrazione di anidride carbonica rispetto all' ossigeno all' interno del cloroplasto, aumentando l' efficienza della fotosintesi. Questi meccanismi sono chiamati meccanismi di concentrazione dell' anidride carbonica, o CCM. Tra questi figurano il metabolismo degli acidi crassulaceanici, la fissazione del carbonio C4 e i pirenoidi. I cloroplasti negli impianti C4 sono notevoli in quanto presentano un chiaro dimorfismo cloroplastico. Domanda: Che cosa può fare rubisco per errore?"
metrics:
- f1
- exact-match
model-index:
- name: mt5-base-question-answering
results:
- task:
type: question-answering
name: "Question Answering"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: f1
value: 0.757
name: "Test F1"
- type: exact-match
value: 0.663
name: "Test Exact Match"
co2_eq_emissions:
emissions: 40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Base for Question Answering ⁉️ 🇮🇹
This repository contains the checkpoint for the [mT5 Base](https://huggingface.co/google/mt5-base) model fine-tuned on extractive question answering on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qa = pipeline("text2text-generation", model='it5/mt5-base-question-answering')
qa("In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?")
>>> [{"generated_text": "ultimo massimo glaciale"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-base-question-answering")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-base-question-answering")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-small-question-answering
|
gsarti
| 2022-03-09T07:57:03Z | 15 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"squad_it",
"text2text-question-answering",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- squad_it
- text2text-question-answering
- text2text-generation
widget:
- text: "In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?"
- text: "L' embargo non era uniforme in tutta Europa. Dei nove membri della Comunità Economica Europea (CEE), i Paesi Bassi hanno dovuto affrontare un embargo totale, il Regno Unito e la Francia hanno ricevuto forniture quasi ininterrotte (poichè si sono rifiutati di consentire all' America di utilizzare i loro aerodromi e le armi e forniture embargo sia agli arabi che agli israeliani), mentre gli altri sei hanno dovuto affrontare tagli parziali. Il Regno Unito era tradizionalmente un alleato di Israele, e il governo di Harold Wilson ha sostenuto gli israeliani durante la guerra dei sei giorni. Il suo successore, Ted Heath, ribaltò questa politica nel 1970, chiedendo a Israele di ritirarsi ai suoi confini prima del 1967. Domanda: Il Regno Unito e la Francia non hanno avuto interruzioni dell' approvvigionamento petrolifero in quanto non hanno consentito a quale paese di utilizzare il loro aeroporto?"
- text: "Nel 1962, il grafico Paul Rand ridisegna il logo ABC nella sua forma più conosciuta (e attuale) con le lettere minuscole \"abc\" racchiuse in un unico cerchio nero. Il nuovo logo esordisce in onda per le promozioni di ABC all' inizio della stagione 1963-64. Le lettere ricordano fortemente il carattere tipografico Bauhaus disegnato da Herbert Bayer negli anni Venti, ma condividono anche similitudini con diversi altri caratteri, come ITC Avant Garde e Horatio, e lo Chalet più simile. La semplicità del logo ha reso più facile la riprogettazione e la duplicazione, il che ha conferito un beneficio per ABC (soprattutto prima dell' avvento della computer grafica). Domanda: Di quale carattere tipografico ricordano le lettere dell' iconico logo ABC?"
- text: "La fotorespirazione può verificarsi quando la concentrazione di ossigeno è troppo elevata. Rubisco non è in grado di distinguere molto bene tra ossigeno e anidride carbonica, quindi può accidentalmente aggiungere O2 invece di CO2 a RuBP. Questo processo riduce l' efficienza della fotosintesi: consuma ATP e ossigeno, rilascia CO2 e non produce zucchero. Può sprecare fino alla metà del carbonio fissato dal ciclo di Calvin. Diversi meccanismi si sono evoluti in diversi lignaggi che aumentano la concentrazione di anidride carbonica rispetto all' ossigeno all' interno del cloroplasto, aumentando l' efficienza della fotosintesi. Questi meccanismi sono chiamati meccanismi di concentrazione dell' anidride carbonica, o CCM. Tra questi figurano il metabolismo degli acidi crassulaceanici, la fissazione del carbonio C4 e i pirenoidi. I cloroplasti negli impianti C4 sono notevoli in quanto presentano un chiaro dimorfismo cloroplastico. Domanda: Che cosa può fare rubisco per errore?"
metrics:
- f1
- exact-match
model-index:
- name: mt5-small-question-answering
results:
- task:
type: question-answering
name: "Question Answering"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: f1
value: 0.660
name: "Test F1"
- type: exact-match
value: 0.560
name: "Test Exact Match"
co2_eq_emissions:
emissions: 17g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Small for Question Answering ⁉️ 🇮🇹
This repository contains the checkpoint for the [mT5 Small](https://huggingface.co/google/mt5-small) model fine-tuned on extractive question answering on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qa = pipeline("text2text-generation", model='it5/mt5-small-question-answering')
qa("In seguito all' evento di estinzione del Cretaceo-Paleogene, l' estinzione dei dinosauri e il clima umido possono aver permesso alla foresta pluviale tropicale di diffondersi in tutto il continente. Dal 66-34 Mya, la foresta pluviale si estendeva fino a sud fino a 45°. Le fluttuazioni climatiche degli ultimi 34 milioni di anni hanno permesso alle regioni della savana di espandersi fino ai tropici. Durante l' Oligocene, ad esempio, la foresta pluviale ha attraversato una banda relativamente stretta. Si espandeva di nuovo durante il Miocene medio, poi si ritrasse ad una formazione prevalentemente interna all' ultimo massimo glaciale. Tuttavia, la foresta pluviale è riuscita ancora a prosperare durante questi periodi glaciali, consentendo la sopravvivenza e l' evoluzione di un' ampia varietà di specie. Domanda: La foresta pluviale amazzonica è diventata per lo più una foresta interna intorno a quale evento globale?")
>>> [{"generated_text": "ultimo massimo glaciale"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-small-question-answering")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-small-question-answering")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-small-question-generation
|
gsarti
| 2022-03-09T07:55:38Z | 14 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"question-generation",
"squad_it",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- question-generation
- squad_it
- text2text-generation
widget:
- text: "Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una \"grande pestilenza nell' aria\". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola \"peste\" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia"
- text: "Il 14 aprile 2011, ABC ha annullato le lunghe opere di sapone All My Children e One Life to Live dopo 41 e 43 anni in onda, rispettivamente (in seguito al contraccolpo dei tifosi, ABC ha venduto i diritti ad entrambi gli spettacoli a Prospect Park, che alla fine ha rilanciato i saponi su Hulu per un' ulteriore stagione nel 2013 e con entrambe le società che si citano in giudizio per accuse di interferenza con il processo di rilancio degli spettacoli, mancato pagamento delle tasse di licenza. Il talk/lifestyle show che ha sostituito One Life to Live, The Revolution, non è riuscito a generare giudizi soddisfacenti ed è stato a sua volta annullato dopo soli sette mesi. La stagione 2011-12 ha visto l' ABC cadere al quarto posto nel 18-49 demografico nonostante rinnovando una manciata di nuovi spettacoli (compresi i drammi matricole Scandal, Revenge e Once Upon a Time) per la seconda stagione. Risposta: Hulu"
- text: "L' American Broadcasting Company (ABC) (stlized nel suo logo come abc dal 1957) è una rete televisiva commerciale americana trasmissione televisiva che è di proprietà del Disney-ABC Television Group, una controllata della divisione Disney Media Networks di The Walt Disney Company. La rete fa parte delle grandi reti televisive Big Three. La rete ha sede a Columbus Avenue e West 66th Street a Manhattan, con ulteriori uffici e stabilimenti di produzione a New York City, Los Angeles e Burbank, California. Risposta: Manhattan"
- text: "La disobbedienza civile non rivoluzionaria è una semplice disobbedienza delle leggi sulla base del fatto che sono giudicate \"sbagliate\" da una coscienza individuale, o come parte di uno sforzo per rendere alcune leggi inefficaci, per causarne l' abrogazione, o per esercitare pressioni per ottenere i propri desideri politici su qualche altra questione. La disobbedienza civile rivoluzionaria è più che altro un tentativo attivo di rovesciare un governo (o di cambiare le tradizioni culturali, i costumi sociali, le credenze religiose, ecc. La rivoluzione non deve necessariamente essere politica, cioè \"rivoluzione culturale\", implica semplicemente un cambiamento radicale e diffuso in una sezione del tessuto sociale). Gli atti di Gandhi sono stati descritti come disobbedienza civile rivoluzionaria. È stato affermato che gli ungheresi sotto Ferenc Deák hanno diretto una disobbedienza civile rivoluzionaria contro il governo austriaco. Thoreau ha anche scritto di disobbedienza civile realizzando \"rivoluzione pacifica\". Howard Zinn, Harvey Wheeler e altri hanno identificato il diritto sposato nella Dichiarazione d' Indipendenza di \"alterare o abolire\" un governo ingiusto come principio di disobbedienza civile. Risposta: Ferenc Deák"
metrics:
- rouge
- bertscore
model-index:
- name: it5-small-question-generation
results:
- task:
type: question-generation
name: "Question generation"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: rouge1
value: 0.367
name: "Test Rouge1"
- type: rouge2
value: 0.189
name: "Test Rouge2"
- type: rougeL
value: 0.344
name: "Test RougeL"
- type: bertscore
value: 0.505
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "8g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Small for Question Generation 💭 🇮🇹
This repository contains the checkpoint for the [IT5 Small](https://huggingface.co/gsarti/it5-small) model fine-tuned on question generation on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qg = pipeline("text2text-generation", model='it5/it5-small-question-generation')
qg("Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una "grande pestilenza nell\' aria". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola "peste" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia")
>>> [{"generated_text": "Per chi è stato redatto il referto medico?"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-small-question-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-small-question-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-question-generation
|
gsarti
| 2022-03-09T07:54:16Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"question-generation",
"squad_it",
"it",
"dataset:squad_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- squad_it
tags:
- italian
- sequence-to-sequence
- question-generation
- squad_it
- text2text-generation
widget:
- text: "Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una \"grande pestilenza nell' aria\". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola \"peste\" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia"
- text: "Il 14 aprile 2011, ABC ha annullato le lunghe opere di sapone All My Children e One Life to Live dopo 41 e 43 anni in onda, rispettivamente (in seguito al contraccolpo dei tifosi, ABC ha venduto i diritti ad entrambi gli spettacoli a Prospect Park, che alla fine ha rilanciato i saponi su Hulu per un' ulteriore stagione nel 2013 e con entrambe le società che si citano in giudizio per accuse di interferenza con il processo di rilancio degli spettacoli, mancato pagamento delle tasse di licenza. Il talk/lifestyle show che ha sostituito One Life to Live, The Revolution, non è riuscito a generare giudizi soddisfacenti ed è stato a sua volta annullato dopo soli sette mesi. La stagione 2011-12 ha visto l' ABC cadere al quarto posto nel 18-49 demografico nonostante rinnovando una manciata di nuovi spettacoli (compresi i drammi matricole Scandal, Revenge e Once Upon a Time) per la seconda stagione. Risposta: Hulu"
- text: "L' American Broadcasting Company (ABC) (stlized nel suo logo come abc dal 1957) è una rete televisiva commerciale americana trasmissione televisiva che è di proprietà del Disney-ABC Television Group, una controllata della divisione Disney Media Networks di The Walt Disney Company. La rete fa parte delle grandi reti televisive Big Three. La rete ha sede a Columbus Avenue e West 66th Street a Manhattan, con ulteriori uffici e stabilimenti di produzione a New York City, Los Angeles e Burbank, California. Risposta: Manhattan"
- text: "La disobbedienza civile non rivoluzionaria è una semplice disobbedienza delle leggi sulla base del fatto che sono giudicate \"sbagliate\" da una coscienza individuale, o come parte di uno sforzo per rendere alcune leggi inefficaci, per causarne l' abrogazione, o per esercitare pressioni per ottenere i propri desideri politici su qualche altra questione. La disobbedienza civile rivoluzionaria è più che altro un tentativo attivo di rovesciare un governo (o di cambiare le tradizioni culturali, i costumi sociali, le credenze religiose, ecc. La rivoluzione non deve necessariamente essere politica, cioè \"rivoluzione culturale\", implica semplicemente un cambiamento radicale e diffuso in una sezione del tessuto sociale). Gli atti di Gandhi sono stati descritti come disobbedienza civile rivoluzionaria. È stato affermato che gli ungheresi sotto Ferenc Deák hanno diretto una disobbedienza civile rivoluzionaria contro il governo austriaco. Thoreau ha anche scritto di disobbedienza civile realizzando \"rivoluzione pacifica\". Howard Zinn, Harvey Wheeler e altri hanno identificato il diritto sposato nella Dichiarazione d' Indipendenza di \"alterare o abolire\" un governo ingiusto come principio di disobbedienza civile. Risposta: Ferenc Deák"
metrics:
- rouge
- bertscore
model-index:
- name: mt5-base-question-generation
results:
- task:
type: question-generation
name: "Question generation"
dataset:
type: squad_it
name: "SQuAD-IT"
metrics:
- type: rouge1
value: 0.346
name: "Test Rouge1"
- type: rouge2
value: 0.174
name: "Test Rouge2"
- type: rougeL
value: 0.324
name: "Test RougeL"
- type: bertscore
value: 0.495
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Base for Question Generation 💭 🇮🇹
This repository contains the checkpoint for the [mT5 Base](https://huggingface.co/google/mt5-base) model fine-tuned on question generation on the [SQuAD-IT corpus](https://huggingface.co/datasets/squad_it) as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
qg = pipeline("text2text-generation", model='it5/mt5-base-question-generation')
qg("Le conoscenze mediche erano stagnanti durante il Medioevo. Il resoconto più autorevole di allora è venuto dalla facoltà di medicina di Parigi in un rapporto al re di Francia che ha incolpato i cieli, sotto forma di una congiunzione di tre pianeti nel 1345 che causò una "grande pestilenza nell\' aria". Questa relazione è diventata la prima e più diffusa di una serie di casi di peste che cercava di dare consigli ai malati. Che la peste fosse causata dalla cattiva aria divenne la teoria più accettata. Oggi, questo è conosciuto come la teoria di Miasma. La parola "peste" non aveva un significato particolare in questo momento, e solo la ricorrenza dei focolai durante il Medioevo gli diede il nome che è diventato il termine medico. Risposta: re di Francia")
>>> [{"generated_text": "Per chi è stato redatto il referto medico?"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-base-question-generation")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-base-question-generation")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-small-news-summarization
|
gsarti
| 2022-03-09T07:52:53Z | 96 | 3 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"fanpage",
"ilpost",
"summarization",
"it",
"dataset:ARTeLab/fanpage",
"dataset:ARTeLab/ilpost",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- ARTeLab/fanpage
- ARTeLab/ilpost
tags:
- italian
- sequence-to-sequence
- fanpage
- ilpost
- summarization
widget:
- text: "Non lo vuole sposare. E’ quanto emerge all’interno dell’ultima intervista di Raffaella Fico che, ringraziando Mancini per i buoni consigli elargiti al suo fidanzato, rimanda l’idea del matrimonio per qualche anno ancora. La soubrette, che è stata recentemente protagonista di una dedica di Supermario, non ha ancora intenzione di accasarsi perché è sicura che per mettersi la fede al dito ci sia ancora tempo. Nonostante il suo Mario sia uno degli sportivi più desiderati al mondo, l’ex protagonista del Grande Fratello non ha alcuna intenzione di cedere seriamente alla sua corte. Solo qualche giorno fa, infatti, dopo l’ultima bravata di Balotelli, Mancini gli aveva consigliato di sposare la sua Raffaella e di mettere la testa a posto. Chi pensava che sarebbe stato Mario a rispondere, però, si è sbagliato. A mettere le cose bene in chiaro è la Fico che, intervistata dall’emittente radiofonica Rtl 102.5, dice: È presto per sposarsi, siamo ancora molto giovani. È giusto che prima uno si realizzi nel proprio lavoro. E poi successivamente perché no, ci si può anche pensare. Quando si è giovani capita di fare qualche pazzia, quindi ci sta. Comunque i tabloid inglesi sono totalmente accaniti sulla sua vita privata quando poi dovrebbero interessarsi di più di quello che fa sul campo. Lui non fa le cose con cattiveria, ma quando si è giovani si fanno determinate cose senza stare a pensare se sono giuste o sbagliate. Mario ha gli obiettivi puntati addosso: più per la sua vita privata che come giocatore. Per me può anche andare in uno strip club, se non fa niente di male, con gli amici, però devo dire che alla fine torna sempre da me, sono la sua preferita."
- text: "Valerio è giovanissimo ma già una star. Fuori dall’Ariston ragazzine e meno ragazzine passano ore anche sotto la pioggia per vederlo. Lui è forte del suo talento e sicuro. Partecipa in gara tra i “big” di diritto, per essere arrivato in finalissima nel programma Amici di Maria De Filippi e presenta il brano Per tutte le volte che scritta per lui da Pierdavide Carone. Valerio Scanu è stato eliminato. Ma non è detta l'ultima parola: il duetto di questa sera con Alessandra Amoroso potrebbe risollevarlo e farlo rientrare in gara. Che cosa è successo alla giuria visto che sei stato eliminato anche se l’esibizione era perfetta? Nn lo so. Sono andate bene le esibizioni, ero emozionato ma tranquillo. Ero contento ma ho cantato bene. Non sono passato e stasera ci sarà il ballottaggio… Quali sono le differenze tra Amici e Sanremo? Sono due cose diverse. Amici ti prepara a salire sul palco di amici. A Sanremo ci devi arrivare… ho fatto più di sessanta serate nel tour estivo, poi promozione del secondo disco. Una bella palestra. Sono cresciuto anche umanamente. Sono riuscito a percepire quello che il pubblico trasmette. L’umiltà? Prima di tutto. Sennò non sarei qui."
- text: "L’azienda statunitense Broadcom, uno dei più grandi produttori di semiconduttori al mondo, ha presentato un’offerta per acquisire Qualcomm, altra grande società degli Stati Uniti conosciuta soprattutto per la sua produzione di microprocessori Snapdragon (ARM), utilizzati in centinaia di milioni di smartphone in giro per il mondo. Broadcom ha proposto di acquistare ogni azione di Qualcomm al prezzo di 70 dollari, per un valore complessivo di circa 105 miliardi di dollari (130 miliardi se si comprendono 25 miliardi di debiti netti) . Se l’operazione dovesse essere approvata, sarebbe una delle più grandi acquisizioni di sempre nella storia del settore tecnologico degli Stati Uniti. Broadcom ha perfezionato per mesi la sua proposta di acquisto e, secondo i media statunitensi, avrebbe già preso contatti con Qualcomm per trovare un accordo. Secondo gli analisti, Qualcomm potrebbe comunque opporsi all’acquisizione perché il prezzo offerto è di poco superiore a quello dell’attuale valore delle azioni dell’azienda. Ci potrebbero essere inoltre complicazioni sul piano dell’antitrust da valutare, prima di un’eventuale acquisizione."
- text: "Dal 31 maggio è infine partita la piattaforma ITsART, a più di un anno da quando – durante il primo lockdown – il ministro della Cultura Dario Franceschini ne aveva parlato come di «una sorta di Netflix della cultura», pensata per «offrire a tutto il mondo la cultura italiana a pagamento». È presto per dare giudizi definitivi sulla piattaforma, e di certo sarà difficile farlo anche più avanti senza numeri precisi. Al momento, l’unica cosa che si può fare è guardare com’è fatto il sito, contare quanti contenuti ci sono (circa 700 “titoli”, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro varietà. Intanto, una cosa notata da più parti è che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente."
metrics:
- rouge
model-index:
- name: it5-small-news-summarization
results:
- task:
type: news-summarization
name: "News Summarization"
dataset:
type: newssum-it
name: "NewsSum-IT"
metrics:
- type: rouge1
value: 0.333
name: "Test Rouge1 IlPost"
- type: rouge2
value: 0.162
name: "Test Rouge2 IlPost"
- type: rougeL
value: 0.273
name: "Test RougeL IlPost"
- type: bertscore
value: 0.395
name: "Test BERTScore IlPost"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
- type: rouge1
value: 0.328
name: "Test Rouge1 Fanpage"
- type: rouge2
value: 0.148
name: "Test Rouge2 Fanpage"
- type: rougeL
value: 0.242
name: "Test RougeL Fanpage"
- type: bertscore
value: 0.377
name: "Test BERTScore Fanpage"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "8g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# IT5 Small for News Summarization ✂️🗞️ 🇮🇹
This repository contains the checkpoint for the [IT5 Small](https://huggingface.co/gsarti/it5-small) model fine-tuned on news summarization on the [Fanpage](https://huggingface.co/datasets/ARTeLab/fanpage) and [Il Post](https://huggingface.co/datasets/ARTeLab/ilpost) corpora as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
newsum = pipeline("summarization", model='it5/it5-small-news-summarization')
newsum("Dal 31 maggio è infine partita la piattaforma ITsART, a più di un anno da quando – durante il primo lockdown – il ministro della Cultura Dario Franceschini ne aveva parlato come di «una sorta di Netflix della cultura», pensata per «offrire a tutto il mondo la cultura italiana a pagamento». È presto per dare giudizi definitivi sulla piattaforma, e di certo sarà difficile farlo anche più avanti senza numeri precisi. Al momento, l’unica cosa che si può fare è guardare com’è fatto il sito, contare quanti contenuti ci sono (circa 700 “titoli”, tra film, documentari, spettacoli teatrali e musicali e altri eventi) e provare a dare un giudizio sul loro valore e sulla loro varietà. Intanto, una cosa notata da più parti è che diversi contenuti di ITsART sono a pagamento sulla piattaforma sebbene altrove, per esempio su RaiPlay, siano invece disponibili gratuitamente.")
>>> [{"generated_text": "ITsART, la Netflix della cultura italiana, parte da maggio. Film, documentari, spettacoli teatrali e musicali disponibili sul nuovo sito a pagamento."}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-small-news-summarization")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-small-news-summarization")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-wiki-summarization
|
gsarti
| 2022-03-09T07:51:31Z | 13 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"wikipedia",
"summarization",
"wits",
"it",
"dataset:wits",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- wits
tags:
- italian
- sequence-to-sequence
- wikipedia
- summarization
- wits
widget:
- text: "La 5ª Commissione ha competenza per i disegni di legge riguardanti le specifiche materie del bilancio, del personale e dei servizi del Ministero dell'economia, nonché per i disegni di legge riguardanti la materia finanziaria. La Commissione è composta da 26 senatori (di cui 2 segretari, 2 vicepresidenti di cui 1 componente esterno, e un presidente) scelti in modo omogeneo tra i componenti di quel ramo del Parlamento, in modo da rispecchiarne le forze politiche presenti. Essi sono scelti dai gruppi parlamentari (e non dal Presidente, come invece accade per l'organismo della Giunta parlamentare): per la nomina dei membri ciascun Gruppo, entro cinque giorni dalla propria costituzione, procede, dandone comunicazione alla Presidenza del Senato, alla designazione dei propri rappresentanti nelle singole Commissioni permanenti. Ogni senatore chiamato a far parte del governo o eletto presidente della Commissione è, per la durata della carica, sostituito dal suo gruppo nella Commissione con un altro senatore, che continuerà ad appartenere anche alla Commissione di provenienza. Tranne in rari casi nessun Senatore può essere assegnato a più di una Commissione permanente. Le Commissioni permanenti sono rinnovate dopo il primo biennio della legislatura ed i loro componenti possono essere confermati."
- text: "Interni della chiesa Si pensa che già ai tempi di Gediminas vi fosse una piccola chiesa, probabilmente in legno. Nel 1408 circa Vitoldo costruì la chiesa dello Spirito Santo che andò in seguito ampliata. Nel 1501 Alessandro Jagellone lo donò al monastero domenicano, il più antico della Lituania, che nel 1679-88 fu ampliato e ricostruito. Di quel periodo sopravvivono le mura della chiesa, mentre l'arredamento interno fu realizzato nel 1749-1770 e la cupola affrontò dei lavori di restauro nel 1752-1760. Nel 1844 le autorità zariste chiusero il monastero e la chiesa divenne parrocchiale. Oggi serve la comunità polacca di Vilnius. Su via Šv. Ignoto fu fondato un monastero domenicano nel 1501. Come molti altri edifici, questo monastero fu convertito in una prigione dalle autorità zariste nel 1807. Costituì un luogo di prigionia per molti patrioti lituani, nello specifico i Filareti, i quali parteciparono alle rivolte del 1831 e del 1863. Organo La chiesa si trova lateralmente rispetto alla strada e non ha una facciata principale ben disegnata. L'altezza, inclusa la cupola, è di 51 m. La parte inferiore della facciata (con piccole torri gemelle) è ricoperta da edifici conventuali e l'esterno presenta caratteristiche architettoniche tipiche del tardo barocco. Celebre per i fantasiosi ornamenti rococò, l'interno della chiesa è tra i più celebri della Lituania per via dei cartigli con vari stemmi e affreschi lungo la navata: vi sono 16 altari nella chiesa. Gli altari e il pulpito sono assai decorati con sculture e ornamenti rotondi e in rilievo. Tra gli affreschi barocchi, si pensi alla composizione multi-figurale intitolata ''Apoteosi dello Spirito Santo'' (neobarocco, XIX secolo) nella cupola, 45 dipinti nella chiesa (tra cui un'immagine di Santa Barbara con un'ambientazione del XVII o XVIII secolo, una di Santa Caterina da Siena in stile rococò di Szymon Czechowicz, un ritratto di Alessandro Jagellone di un artista sconosciuto della seconda metà del XVIII secolo). Un ingresso sotto l'altare conduce alle grandi volte, labirintiche, con molte stanze e cripte: i sotterranei ospitano i resti di centinaia di residenti di Vilnius, alcuni dei quali mummificatisi naturalmente, e sono circondati da leggende metropolitane. Sebbene l'esistenza dei sotterranei fosse nota, i primi sforzi per esplorare e mappare le cripte furono abbandonate nonostante lo sforzo degli studenti dell'Università di Vilnius negli anni '30. Tuttavia, questi ultimi non avevano osservato le corrette procedure archeologiche e causarono infatti molti danni: il modus operandi prevedeva lo smistamento delle ossa ponendo tutti i teschi sugli scaffali e rimuovendoli le tombe. Da allora, i resti sono stati spostati molte volte lasciandoli in uno stato casuale e disorganizzato. Stando alle leggende che aleggiano sul luogo, i resti sarebbero di soldati francesi recatisi in città nel corso della campagna di Russia del 1812 avviata da Napoleone Bonaparte, di vittime dell'Inquisizione o della peste nera. Più romantiche risultano le affermazioni di chi sostiene che i corridoi sotterranei facevano parte di una rete di passaggi più ampia che consentiva agli amanti leggendari Barbara Radziwiłł e Sigismondo II Augusto di incontrarsi in segreto. Nel 2011, gli antropologi dell'Università di Vilnius, guidati da Rimantas Jankauskas, avviarono uno studio sui corpi mummificati, stimando settimane dopo che le volte conservassero i resti di circa 600 persone, tra cui molte donne e bambini dalla metà del XVIII secolo all'inizio del XIX secolo. Il team ha selezionato i cadaveri meglio conservati e ha eseguito la loro tomografia. I risultati mostrano che molte persone erano in sovrappeso e avevano l'alluce valgo, il che ha portato alla conclusione che si trattava di alti borghesi o comunque di cittadini abbienti. "
- text: "Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. "
- text: "Vanni ha la sua prima mostra personale nel 1948, alla Galleria Margherita di Roma. Nel 1949 vince una borsa di studio che lo porterà a studiare ad Amsterdam sotto la guida del pittore neoplastico Friedrich Vordemberge-Gildewart. Nel 1952 vince una Fulbright Scholarship che lo porterà a studiare in America, alla Yale University, sotto la guida di Josef Albers. Dal 1953 al 1960 si stabilisce a Parigi, dove illustra alcuni libri per bambini che in seguito vinceranno il premio del Club des Editeurs. Nel 1954 lavora come consulente del colore per il documentario su Picasso di Luciano Emmer, e nel 1955 comincia la sua lunga collaborazione con la Galleria Schneider, affiancando artisti come Corrado Cagli. Dal 1969 al 1974 lavora su dei bassorilievi in vetro resina sui quali vengono proiettati dei film astratti da lui creati, per creare dei quadri che si trasformino continuamente nel tempo. Nel 1979 lascia Roma per stabilirsi a New York, dove alla carriera di pittore affiancherà quella di professore per la prestigiosa Cooper Union School of Art, dove insegnerà ininterrottamente dal 1984 al 2014. L'opera pittorica di Vanni è segnata da una visione estremamente personale, lontana dalle correnti e dai movimenti che hanno caratterizzato la seconda metà del XX secolo. Memore delle lunghe conversazioni avute da Vanni nella sua primissima gioventù, con il filosofo e pittore futurista Alberto Bragaglia, le sue opere sono contrassegnate da un “eclettismo” formale programmatico, alla base del quale resta costante una conoscenza profonda delle molteplici tecniche artistiche utilizzate (tra cui il mosaico, l’affresco e la tempera ad uovo). Pur esprimendosi per lo più in cicli di opere dove l’astrazione formale è la principale componente figurativa, sono da sottolineare alcune opere dove Vanni ha dato prova di una importante padronanza dell’arte figurativa. Importanti e numerose sono le sue realizzazioni anche nel campo dell’illustrazione. Sue sono le illustrazioni per la novella ''Agostino'' di Alberto Moravia, per il libro ''Love'' di Lowell A. Siff e delle ''Contes de Cristal'' di Alice Coléno. Ha tenuto mostre personali in Italia e all’estero ed esposto in mostre collettive di rappresentanza italiana nei musei e nelle gallerie di ogni parte del mondo. "
metrics:
- rouge
- bertscore
model-index:
- name: mt5-base-wiki-summarization
results:
- task:
type: wiki-summarization
name: "Wikipedia Summarization"
dataset:
type: wits
name: "WITS"
metrics:
- type: rouge1
value: 0.348
name: "Test Rouge1"
- type: rouge2
value: 0.200
name: "Test Rouge2"
- type: rougeL
value: 0.315
name: "Test RougeL"
- type: bertscore
value: 0.520
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Base for Wikipedia Summarization ✂️📑 🇮🇹
This repository contains the checkpoint for the [mT5 Base](https://huggingface.co/google/mt5-base) model fine-tuned on Wikipedia summarization on the [WITS](https://www.semanticscholar.org/paper/WITS%3A-Wikipedia-for-Italian-Text-Summarization-Casola-Lavelli/ad6c83122e721c7c0db4a40727dac3b4762cd2b1) dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
wikisum = pipeline("summarization", model='it5/mt5-base-wiki-summarization')
wikisum("Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. ")
>>> [{"generated_text": "L' '''isola di Rabot''' si trova in prossimità dell'isola di Renaud, a sud dell'Argentina."}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-base-wiki-summarization")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-base-wiki-summarization")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-small-wiki-summarization
|
gsarti
| 2022-03-09T07:51:07Z | 9 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"mt5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"wikipedia",
"summarization",
"wits",
"it",
"dataset:wits",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
datasets:
- wits
tags:
- italian
- sequence-to-sequence
- wikipedia
- summarization
- wits
widget:
- text: "La 5ª Commissione ha competenza per i disegni di legge riguardanti le specifiche materie del bilancio, del personale e dei servizi del Ministero dell'economia, nonché per i disegni di legge riguardanti la materia finanziaria. La Commissione è composta da 26 senatori (di cui 2 segretari, 2 vicepresidenti di cui 1 componente esterno, e un presidente) scelti in modo omogeneo tra i componenti di quel ramo del Parlamento, in modo da rispecchiarne le forze politiche presenti. Essi sono scelti dai gruppi parlamentari (e non dal Presidente, come invece accade per l'organismo della Giunta parlamentare): per la nomina dei membri ciascun Gruppo, entro cinque giorni dalla propria costituzione, procede, dandone comunicazione alla Presidenza del Senato, alla designazione dei propri rappresentanti nelle singole Commissioni permanenti. Ogni senatore chiamato a far parte del governo o eletto presidente della Commissione è, per la durata della carica, sostituito dal suo gruppo nella Commissione con un altro senatore, che continuerà ad appartenere anche alla Commissione di provenienza. Tranne in rari casi nessun Senatore può essere assegnato a più di una Commissione permanente. Le Commissioni permanenti sono rinnovate dopo il primo biennio della legislatura ed i loro componenti possono essere confermati."
- text: "Interni della chiesa Si pensa che già ai tempi di Gediminas vi fosse una piccola chiesa, probabilmente in legno. Nel 1408 circa Vitoldo costruì la chiesa dello Spirito Santo che andò in seguito ampliata. Nel 1501 Alessandro Jagellone lo donò al monastero domenicano, il più antico della Lituania, che nel 1679-88 fu ampliato e ricostruito. Di quel periodo sopravvivono le mura della chiesa, mentre l'arredamento interno fu realizzato nel 1749-1770 e la cupola affrontò dei lavori di restauro nel 1752-1760. Nel 1844 le autorità zariste chiusero il monastero e la chiesa divenne parrocchiale. Oggi serve la comunità polacca di Vilnius. Su via Šv. Ignoto fu fondato un monastero domenicano nel 1501. Come molti altri edifici, questo monastero fu convertito in una prigione dalle autorità zariste nel 1807. Costituì un luogo di prigionia per molti patrioti lituani, nello specifico i Filareti, i quali parteciparono alle rivolte del 1831 e del 1863. Organo La chiesa si trova lateralmente rispetto alla strada e non ha una facciata principale ben disegnata. L'altezza, inclusa la cupola, è di 51 m. La parte inferiore della facciata (con piccole torri gemelle) è ricoperta da edifici conventuali e l'esterno presenta caratteristiche architettoniche tipiche del tardo barocco. Celebre per i fantasiosi ornamenti rococò, l'interno della chiesa è tra i più celebri della Lituania per via dei cartigli con vari stemmi e affreschi lungo la navata: vi sono 16 altari nella chiesa. Gli altari e il pulpito sono assai decorati con sculture e ornamenti rotondi e in rilievo. Tra gli affreschi barocchi, si pensi alla composizione multi-figurale intitolata ''Apoteosi dello Spirito Santo'' (neobarocco, XIX secolo) nella cupola, 45 dipinti nella chiesa (tra cui un'immagine di Santa Barbara con un'ambientazione del XVII o XVIII secolo, una di Santa Caterina da Siena in stile rococò di Szymon Czechowicz, un ritratto di Alessandro Jagellone di un artista sconosciuto della seconda metà del XVIII secolo). Un ingresso sotto l'altare conduce alle grandi volte, labirintiche, con molte stanze e cripte: i sotterranei ospitano i resti di centinaia di residenti di Vilnius, alcuni dei quali mummificatisi naturalmente, e sono circondati da leggende metropolitane. Sebbene l'esistenza dei sotterranei fosse nota, i primi sforzi per esplorare e mappare le cripte furono abbandonate nonostante lo sforzo degli studenti dell'Università di Vilnius negli anni '30. Tuttavia, questi ultimi non avevano osservato le corrette procedure archeologiche e causarono infatti molti danni: il modus operandi prevedeva lo smistamento delle ossa ponendo tutti i teschi sugli scaffali e rimuovendoli le tombe. Da allora, i resti sono stati spostati molte volte lasciandoli in uno stato casuale e disorganizzato. Stando alle leggende che aleggiano sul luogo, i resti sarebbero di soldati francesi recatisi in città nel corso della campagna di Russia del 1812 avviata da Napoleone Bonaparte, di vittime dell'Inquisizione o della peste nera. Più romantiche risultano le affermazioni di chi sostiene che i corridoi sotterranei facevano parte di una rete di passaggi più ampia che consentiva agli amanti leggendari Barbara Radziwiłł e Sigismondo II Augusto di incontrarsi in segreto. Nel 2011, gli antropologi dell'Università di Vilnius, guidati da Rimantas Jankauskas, avviarono uno studio sui corpi mummificati, stimando settimane dopo che le volte conservassero i resti di circa 600 persone, tra cui molte donne e bambini dalla metà del XVIII secolo all'inizio del XIX secolo. Il team ha selezionato i cadaveri meglio conservati e ha eseguito la loro tomografia. I risultati mostrano che molte persone erano in sovrappeso e avevano l'alluce valgo, il che ha portato alla conclusione che si trattava di alti borghesi o comunque di cittadini abbienti. "
- text: "Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. "
- text: "Vanni ha la sua prima mostra personale nel 1948, alla Galleria Margherita di Roma. Nel 1949 vince una borsa di studio che lo porterà a studiare ad Amsterdam sotto la guida del pittore neoplastico Friedrich Vordemberge-Gildewart. Nel 1952 vince una Fulbright Scholarship che lo porterà a studiare in America, alla Yale University, sotto la guida di Josef Albers. Dal 1953 al 1960 si stabilisce a Parigi, dove illustra alcuni libri per bambini che in seguito vinceranno il premio del Club des Editeurs. Nel 1954 lavora come consulente del colore per il documentario su Picasso di Luciano Emmer, e nel 1955 comincia la sua lunga collaborazione con la Galleria Schneider, affiancando artisti come Corrado Cagli. Dal 1969 al 1974 lavora su dei bassorilievi in vetro resina sui quali vengono proiettati dei film astratti da lui creati, per creare dei quadri che si trasformino continuamente nel tempo. Nel 1979 lascia Roma per stabilirsi a New York, dove alla carriera di pittore affiancherà quella di professore per la prestigiosa Cooper Union School of Art, dove insegnerà ininterrottamente dal 1984 al 2014. L'opera pittorica di Vanni è segnata da una visione estremamente personale, lontana dalle correnti e dai movimenti che hanno caratterizzato la seconda metà del XX secolo. Memore delle lunghe conversazioni avute da Vanni nella sua primissima gioventù, con il filosofo e pittore futurista Alberto Bragaglia, le sue opere sono contrassegnate da un “eclettismo” formale programmatico, alla base del quale resta costante una conoscenza profonda delle molteplici tecniche artistiche utilizzate (tra cui il mosaico, l’affresco e la tempera ad uovo). Pur esprimendosi per lo più in cicli di opere dove l’astrazione formale è la principale componente figurativa, sono da sottolineare alcune opere dove Vanni ha dato prova di una importante padronanza dell’arte figurativa. Importanti e numerose sono le sue realizzazioni anche nel campo dell’illustrazione. Sue sono le illustrazioni per la novella ''Agostino'' di Alberto Moravia, per il libro ''Love'' di Lowell A. Siff e delle ''Contes de Cristal'' di Alice Coléno. Ha tenuto mostre personali in Italia e all’estero ed esposto in mostre collettive di rappresentanza italiana nei musei e nelle gallerie di ogni parte del mondo. "
metrics:
- rouge
- bertscore
model-index:
- name: mt5-small-wiki-summarization
results:
- task:
type: wiki-summarization
name: "Wikipedia Summarization"
dataset:
type: wits
name: "WITS"
metrics:
- type: rouge1
value: 0.347
name: "Test Rouge1"
- type: rouge2
value: 0.200
name: "Test Rouge2"
- type: rougeL
value: 0.316
name: "Test RougeL"
- type: bertscore
value: 0.517
name: "Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "14g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
thumbnail: https://gsarti.com/publication/it5/featured.png
---
# mT5 Small for Wikipedia Summarization ✂️📑 🇮🇹
This repository contains the checkpoint for the [mT5 Small](https://huggingface.co/google/mt5-small) model fine-tuned on Wikipedia summarization on the [WITS](https://www.semanticscholar.org/paper/WITS%3A-Wikipedia-for-Italian-Text-Summarization-Casola-Lavelli/ad6c83122e721c7c0db4a40727dac3b4762cd2b1) dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
wikisum = pipeline("summarization", model='it5/mt5-small-wiki-summarization')
wikisum("Le dimensioni dell'isola sono di 8 km di lunghezza e di 3,2 km di larghezza. Si trova a 1,6 km a sud-est dell'isola di Renaud, dalla quale è separata dal passaggio Rodman. La sua altezza è di 100 m. Fu scoperta dall'esploratore e baleniere britannico John Biscoe nel 1832 e venne mappata durante una spedizione antartica francese realizzata nel primo decennio del XX secolo. Al comando della spedizione era Jean-Baptiste Charcot e il nome fu scelto per onorare l'esploratore e geografo francese Charles Rabot. === Rivendicazioni territoriali === * Secondo l'Argentina appartiene al dipartimento dell'Antartide Argentina nella provincia della Terra del Fuoco. * Secondo il Cile appartiene al comune antartico della provincia cilena antartica nella regione di Magallanes e dell'Antartico cileno. * Secondo il Regno Unito fa parte del territorio antartico britannico. Per il Trattato Antartico tali rivendicazioni sono sospese. Sull'isola è presente il rifugio Guillochon, sito storico antartico. ")
>>> [{"generated_text": "L' '''isola di Rabot''' si trova in prossimità dell'isola di Renaud, a sud dell'Argentina."}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-small-wiki-summarization")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-small-wiki-summarization")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/it5-large-informal-to-formal
|
gsarti
| 2022-03-09T07:48:09Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"style-transfer",
"formality-style-transfer",
"it",
"dataset:yahoo/xformal_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
tags:
- italian
- sequence-to-sequence
- style-transfer
- formality-style-transfer
datasets:
- yahoo/xformal_it
widget:
- text: "maronn qualcuno mi spieg' CHECCOSA SUCCEDE?!?!"
- text: "wellaaaaaaa, ma fraté sei proprio troppo simpatiko, grazieeee!!"
- text: "nn capisco xke tt i ragazzi lo fanno"
- text: "IT5 è SUPERMEGA BRAVISSIMO a capire tt il vernacolo italiano!!!"
metrics:
- rouge
- bertscore
model-index:
- name: it5-large-informal-to-formal
results:
- task:
type: formality-style-transfer
name: "Informal-to-formal Style Transfer"
dataset:
type: xformal_it
name: "XFORMAL (Italian Subset)"
metrics:
- type: rouge1
value: 0.663
name: "Avg. Test Rouge1"
- type: rouge2
value: 0.477
name: "Avg. Test Rouge2"
- type: rougeL
value: 0.645
name: "Avg. Test RougeL"
- type: bertscore
value: 0.714
name: "Avg. Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "51g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
---
# IT5 Base for Informal-to-formal Style Transfer 🧐
This repository contains the checkpoint for the [IT5 Large](https://huggingface.co/gsarti/it5-large) model fine-tuned on Informal-to-formal style transfer on the Italian subset of the XFORMAL dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
i2f = pipeline("text2text-generation", model='it5/it5-large-informal-to-formal')
i2f("nn capisco xke tt i ragazzi lo fanno")
>>> [{"generated_text": "non comprendo perché tutti i ragazzi agiscono così"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/it5-large-informal-to-formal")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/it5-large-informal-to-formal")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={{IT5}: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint 2203.03759},
url={https://arxiv.org/abs/2203.03759},
year={2022},
month={mar}
}
```
|
gsarti/mt5-base-formal-to-informal
|
gsarti
| 2022-03-09T07:44:08Z | 6 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"jax",
"tensorboard",
"t5",
"text2text-generation",
"italian",
"sequence-to-sequence",
"style-transfer",
"formality-style-transfer",
"it",
"dataset:yahoo/xformal_it",
"arxiv:2203.03759",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-02T23:29:05Z |
---
language:
- it
license: apache-2.0
tags:
- italian
- sequence-to-sequence
- style-transfer
- formality-style-transfer
datasets:
- yahoo/xformal_it
widget:
- text: "Questa performance è a dir poco spiacevole."
- text: "In attesa di un Suo cortese riscontro, Le auguriamo un piacevole proseguimento di giornata."
- text: "Questa visione mi procura una goduria indescrivibile."
- text: "qualora ciò possa interessarti, ti pregherei di contattarmi."
metrics:
- rouge
- bertscore
model-index:
- name: mt5-base-formal-to-informal
results:
- task:
type: formality-style-transfer
name: "Formal-to-informal Style Transfer"
dataset:
type: xformal_it
name: "XFORMAL (Italian Subset)"
metrics:
- type: rouge1
value: 0.653
name: "Avg. Test Rouge1"
- type: rouge2
value: 0.449
name: "Avg. Test Rouge2"
- type: rougeL
value: 0.632
name: "Avg. Test RougeL"
- type: bertscore
value: 0.667
name: "Avg. Test BERTScore"
args:
- model_type: "dbmdz/bert-base-italian-xxl-uncased"
- lang: "it"
- num_layers: 10
- rescale_with_baseline: True
- baseline_path: "bertscore_baseline_ita.tsv"
co2_eq_emissions:
emissions: "40g"
source: "Google Cloud Platform Carbon Footprint"
training_type: "fine-tuning"
geographical_location: "Eemshaven, Netherlands, Europe"
hardware_used: "1 TPU v3-8 VM"
---
# mT5 Base for Formal-to-informal Style Transfer 🤗
This repository contains the checkpoint for the [mT5 Base](https://huggingface.co/google/mt5-base) model fine-tuned on Formal-to-informal style transfer on the Italian subset of the XFORMAL dataset as part of the experiments of the paper [IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation](https://arxiv.org/abs/2203.03759) by [Gabriele Sarti](https://gsarti.com) and [Malvina Nissim](https://malvinanissim.github.io).
A comprehensive overview of other released materials is provided in the [gsarti/it5](https://github.com/gsarti/it5) repository. Refer to the paper for additional details concerning the reported scores and the evaluation approach.
## Using the model
Model checkpoints are available for usage in Tensorflow, Pytorch and JAX. They can be used directly with pipelines as:
```python
from transformers import pipelines
f2i = pipeline("text2text-generation", model='it5/mt5-base-formal-to-informal')
f2i("Vi ringrazio infinitamente per vostra disponibilità")
>>> [{"generated_text": "e grazie per la vostra disponibilità!"}]
```
or loaded using autoclasses:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("it5/mt5-base-formal-to-informal")
model = AutoModelForSeq2SeqLM.from_pretrained("it5/mt5-base-formal-to-informal")
```
If you use this model in your research, please cite our work as:
```bibtex
@article{sarti-nissim-2022-it5,
title={IT5: Large-scale Text-to-text Pretraining for Italian Language Understanding and Generation},
author={Sarti, Gabriele and Nissim, Malvina},
journal={ArXiv preprint TBD},
url={TBD},
year={2022}
}
```
|
mrm8488/spanish-TinyBERT-betito-finetuned-xnli-es
|
mrm8488
| 2022-03-09T07:29:03Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:xnli",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-08T20:55:51Z |
---
tags:
- generated_from_trainer
datasets:
- xnli
metrics:
- accuracy
model-index:
- name: spanish-TinyBERT-betito-finetuned-xnli-es
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: xnli
type: xnli
args: es
metrics:
- name: Accuracy
type: accuracy
value: 0.7475049900199601
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# spanish-TinyBERT-betito-finetuned-xnli-es
This model is a fine-tuned version of [mrm8488/spanish-TinyBERT-betito](https://huggingface.co/mrm8488/spanish-TinyBERT-betito) on the xnli dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7104
- Accuracy: 0.7475
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2.50838112218154e-05
- train_batch_size: 8
- eval_batch_size: 64
- seed: 13
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:------:|:---------------:|:--------:|
| 0.7191 | 1.0 | 49399 | 0.6829 | 0.7112 |
| 0.6323 | 2.0 | 98798 | 0.6527 | 0.7305 |
| 0.5727 | 3.0 | 148197 | 0.6531 | 0.7465 |
| 0.4964 | 4.0 | 197596 | 0.7079 | 0.7427 |
| 0.4929 | 5.0 | 246995 | 0.7104 | 0.7475 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
M-Quan/wav2vec2-demo
|
M-Quan
| 2022-03-09T06:20:54Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2022-03-09T01:26:39Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-demo
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-demo
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4239
- Wer: 0.3508
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 3.4093 | 4.0 | 500 | 1.2405 | 0.8685 |
| 0.5597 | 8.0 | 1000 | 0.4538 | 0.4437 |
| 0.2113 | 12.0 | 1500 | 0.4106 | 0.3749 |
| 0.1188 | 16.0 | 2000 | 0.4609 | 0.3775 |
| 0.0776 | 20.0 | 2500 | 0.4239 | 0.3508 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.10.3
|
nickmuchi/vit-base-xray-pneumonia
|
nickmuchi
| 2022-03-09T05:43:35Z | 153 | 4 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"vit",
"image-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2022-03-09T02:04:50Z |
---
license: apache-2.0
tags:
- image-classification
- generated_from_trainer
datasets:
- chest xrays
widget:
- src: https://drive.google.com/uc?id=1yqnhD4Wjt4Y_NGLtijTGGaaw9GL497kQ
example_title: PNEUMONIA
- src: https://drive.google.com/uc?id=1xjcIEDb8kuSd4wF44gCEgsc0PfRvs53m
example_title: NORMAL
metrics:
- accuracy
model-index:
- name: vit-base-xray-pneumonia
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# vit-base-xray-pneumonia
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the [chest-xray-pneumonia](https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3387
- Accuracy: 0.9006
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.1233 | 0.31 | 100 | 1.1662 | 0.6651 |
| 0.0868 | 0.61 | 200 | 0.3387 | 0.9006 |
| 0.1387 | 0.92 | 300 | 0.5297 | 0.8237 |
| 0.1264 | 1.23 | 400 | 0.4566 | 0.8590 |
| 0.0829 | 1.53 | 500 | 0.6832 | 0.8285 |
| 0.0734 | 1.84 | 600 | 0.4886 | 0.8157 |
| 0.0132 | 2.15 | 700 | 1.3639 | 0.7292 |
| 0.0877 | 2.45 | 800 | 0.5258 | 0.8846 |
| 0.0516 | 2.76 | 900 | 0.8772 | 0.8013 |
| 0.0637 | 3.07 | 1000 | 0.4947 | 0.8558 |
| 0.0022 | 3.37 | 1100 | 1.0062 | 0.8045 |
| 0.0555 | 3.68 | 1200 | 0.7822 | 0.8285 |
| 0.0405 | 3.99 | 1300 | 1.9288 | 0.6779 |
| 0.0012 | 4.29 | 1400 | 1.2153 | 0.7981 |
| 0.0034 | 4.6 | 1500 | 1.8931 | 0.7308 |
| 0.0339 | 4.91 | 1600 | 0.9071 | 0.8590 |
| 0.0013 | 5.21 | 1700 | 1.6266 | 0.7580 |
| 0.0373 | 5.52 | 1800 | 1.5252 | 0.7676 |
| 0.001 | 5.83 | 1900 | 1.2748 | 0.7869 |
| 0.0005 | 6.13 | 2000 | 1.2103 | 0.8061 |
| 0.0004 | 6.44 | 2100 | 1.3133 | 0.7981 |
| 0.0004 | 6.75 | 2200 | 1.2200 | 0.8045 |
| 0.0004 | 7.06 | 2300 | 1.2834 | 0.7933 |
| 0.0004 | 7.36 | 2400 | 1.3080 | 0.7949 |
| 0.0003 | 7.67 | 2500 | 1.3814 | 0.7917 |
| 0.0004 | 7.98 | 2600 | 1.2853 | 0.7965 |
| 0.0003 | 8.28 | 2700 | 1.3644 | 0.7933 |
| 0.0003 | 8.59 | 2800 | 1.3137 | 0.8013 |
| 0.0003 | 8.9 | 2900 | 1.3507 | 0.7997 |
| 0.0003 | 9.2 | 3000 | 1.3751 | 0.7997 |
| 0.0003 | 9.51 | 3100 | 1.3884 | 0.7981 |
| 0.0003 | 9.82 | 3200 | 1.3831 | 0.7997 |
## Example Images
#### Pneumonia Chest X-Ray

#### Normal Chest X-Ray

### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
megagonlabs/cocosum-cont-few
|
megagonlabs
| 2022-03-09T05:18:18Z | 0 | 0 | null |
[
"license:bsd-3-clause",
"region:us"
] | null | 2022-03-07T23:29:46Z |
---
license: bsd-3-clause
---
See original GitHub repo for more details [here](https://github.com/megagonlabs/cocosum)
|
josu/albert-pt-br
|
josu
| 2022-03-09T03:41:04Z | 6 | 1 |
transformers
|
[
"transformers",
"pytorch",
"albert",
"fill-mask",
"portuguese",
"brazil",
"pt_BR",
"pt",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-09T02:32:11Z |
---
language: pt
tags:
- portuguese
- brazil
- pt_BR
widget:
- text: Marte está no [MASK] solar.
---
``` python
from transformers import pipeline, AlbertTokenizer, AlbertForMaskedLM
model = AlbertForMaskedLM.from_pretrained('josu/albert-pt-br')
tokenizer = AlbertTokenizer.from_pretrained('josu/albert-pt-br')
unmasker = pipeline('fill-mask', model=model, tokenizer=tokenizer ,device=0)
text = 'Marte está no [MASK] solar.'
unmasker(text)
[{'score': 0.7004144191741943,
'token': 244,
'token_str': 'sistema',
'sequence': 'marte esta no sistema solar.'},
{'score': 0.02539917267858982,
'token': 4077,
'token_str': 'solar',
'sequence': 'marte esta no solar solar.'},
{'score': 0.020301498472690582,
'token': 49,
'token_str': 'seu',
'sequence': 'marte esta no seu solar.'},
{'score': 0.01753508299589157,
'token': 482,
'token_str': 'centro',
'sequence': 'marte esta no centro solar.'},
{'score': 0.013344300910830498,
'token': 1401,
'token_str': 'plano',
'sequence': 'marte esta no plano solar.'}]
```
|
jcai1/ss_ver1
|
jcai1
| 2022-03-09T03:03:20Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T01:28:26Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: ss_ver1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# ss_ver1
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on an unknown dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---:|
| No log | 1.0 | 436 | 0.0001 | 1.0 | 0.0 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
daisyxie21/bert-base-uncased-8-50-0.01
|
daisyxie21
| 2022-03-09T02:13:19Z | 7 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:glue",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-08T16:10:03Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- matthews_correlation
model-index:
- name: bert-base-uncased-8-50-0.01
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: glue
type: glue
args: cola
metrics:
- name: Matthews Correlation
type: matthews_correlation
value: 0.0
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-base-uncased-8-50-0.01
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9219
- Matthews Correlation: 0.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.01
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|:-------------:|:-----:|:-----:|:---------------:|:--------------------:|
| No log | 1.0 | 400 | 0.9219 | 0.0 |
| 1.2047 | 2.0 | 800 | 1.8168 | 0.0 |
| 1.0707 | 3.0 | 1200 | 1.4474 | 0.0 |
| 1.0538 | 4.0 | 1600 | 1.5223 | 0.0 |
| 1.316 | 5.0 | 2000 | 0.8467 | 0.0 |
| 1.316 | 6.0 | 2400 | 1.0906 | 0.0 |
| 1.2739 | 7.0 | 2800 | 0.6851 | 0.0 |
| 1.1342 | 8.0 | 3200 | 1.3170 | 0.0 |
| 1.2572 | 9.0 | 3600 | 0.8870 | 0.0 |
| 1.0237 | 10.0 | 4000 | 1.3236 | 0.0 |
| 1.0237 | 11.0 | 4400 | 0.9025 | 0.0 |
| 0.9597 | 12.0 | 4800 | 0.7757 | 0.0 |
| 1.0946 | 13.0 | 5200 | 1.2551 | 0.0 |
| 1.0011 | 14.0 | 5600 | 1.1606 | 0.0 |
| 1.1111 | 15.0 | 6000 | 0.6040 | 0.0 |
| 1.1111 | 16.0 | 6400 | 1.4347 | 0.0 |
| 1.0098 | 17.0 | 6800 | 0.6218 | 0.0 |
| 1.0829 | 18.0 | 7200 | 0.4979 | 0.0 |
| 0.9131 | 19.0 | 7600 | 1.3040 | 0.0 |
| 0.879 | 20.0 | 8000 | 2.0309 | 0.0 |
| 0.879 | 21.0 | 8400 | 0.5150 | 0.0 |
| 0.9646 | 22.0 | 8800 | 0.4850 | 0.0 |
| 0.9625 | 23.0 | 9200 | 0.5076 | 0.0 |
| 0.9129 | 24.0 | 9600 | 1.1277 | 0.0 |
| 0.8839 | 25.0 | 10000 | 0.9403 | 0.0 |
| 0.8839 | 26.0 | 10400 | 1.6226 | 0.0 |
| 0.9264 | 27.0 | 10800 | 0.6049 | 0.0 |
| 0.7999 | 28.0 | 11200 | 0.9549 | 0.0 |
| 0.752 | 29.0 | 11600 | 0.6757 | 0.0 |
| 0.7675 | 30.0 | 12000 | 0.7320 | 0.0 |
| 0.7675 | 31.0 | 12400 | 0.8393 | 0.0 |
| 0.6887 | 32.0 | 12800 | 0.5977 | 0.0 |
| 0.7563 | 33.0 | 13200 | 0.4815 | 0.0 |
| 0.7671 | 34.0 | 13600 | 0.5457 | 0.0 |
| 0.7227 | 35.0 | 14000 | 0.7384 | 0.0 |
| 0.7227 | 36.0 | 14400 | 0.7749 | 0.0 |
| 0.7308 | 37.0 | 14800 | 0.4726 | 0.0 |
| 0.7191 | 38.0 | 15200 | 0.5069 | 0.0 |
| 0.6846 | 39.0 | 15600 | 0.4762 | 0.0 |
| 0.6151 | 40.0 | 16000 | 0.4738 | 0.0 |
| 0.6151 | 41.0 | 16400 | 0.5114 | 0.0 |
| 0.5982 | 42.0 | 16800 | 0.4866 | 0.0 |
| 0.6199 | 43.0 | 17200 | 0.4717 | 0.0 |
| 0.5737 | 44.0 | 17600 | 0.7651 | 0.0 |
| 0.5703 | 45.0 | 18000 | 0.8008 | 0.0 |
| 0.5703 | 46.0 | 18400 | 0.5391 | 0.0 |
| 0.5748 | 47.0 | 18800 | 0.5097 | 0.0 |
| 0.5297 | 48.0 | 19200 | 0.4731 | 0.0 |
| 0.4902 | 49.0 | 19600 | 0.4720 | 0.0 |
| 0.4955 | 50.0 | 20000 | 0.4748 | 0.0 |
### Framework versions
- Transformers 4.16.2
- Pytorch 1.9.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
aaraki/distilbert-base-uncased-finetuned-cola
|
aaraki
| 2022-03-09T02:08:47Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:glue",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-09T01:56:17Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- matthews_correlation
model-index:
- name: distilbert-base-uncased-finetuned-cola
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: glue
type: glue
args: cola
metrics:
- name: Matthews Correlation
type: matthews_correlation
value: 0.40967417350821667
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-cola
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5026
- Matthews Correlation: 0.4097
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|:-------------:|:-----:|:----:|:---------------:|:--------------------:|
| 0.5335 | 1.0 | 535 | 0.5026 | 0.4097 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
KoichiYasuoka/roberta-base-ukrainian
|
KoichiYasuoka
| 2022-03-08T23:33:19Z | 22 | 2 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"ukrainian",
"masked-lm",
"ubertext",
"uk",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2022-03-08T23:25:41Z |
---
language:
- "uk"
tags:
- "ukrainian"
- "masked-lm"
- "ubertext"
license: "cc-by-sa-4.0"
pipeline_tag: "fill-mask"
mask_token: "[MASK]"
---
# roberta-base-ukrainian
## Model Description
This is a RoBERTa model pre-trained on [Корпус UberText](https://lang.org.ua/uk/corpora/#anchor4). You can fine-tune `roberta-base-ukrainian` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-base-ukrainian-upos), dependency-parsing, and so on.
## How to Use
```py
from transformers import AutoTokenizer,AutoModelForMaskedLM
tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-ukrainian")
model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-base-ukrainian")
```
|
OrfeasTsk/bert-base-uncased-finetuned-nq
|
OrfeasTsk
| 2022-03-08T21:44:45Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"question-answering",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-08T21:35:09Z |
{ 'max_seq_length': 384,
'batch_size': 8,
'learning_rate': {'val': 5e-5, 'schelduler': 'Linear'},
'max_clip_norm': None,
'epochs': 2
}
|
huggingartists/lady-gaga
|
huggingartists
| 2022-03-08T20:28:19Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"jax",
"gpt2",
"text-generation",
"huggingartists",
"lyrics",
"lm-head",
"causal-lm",
"en",
"dataset:huggingartists/lady-gaga",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-02T23:29:05Z |
---
language: en
datasets:
- huggingartists/lady-gaga
tags:
- huggingartists
- lyrics
- lm-head
- causal-lm
widget:
- text: "I am"
---
<div class="inline-flex flex-col" style="line-height: 1.5;">
<div class="flex">
<div
style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://images.genius.com/e7e76c378cb43b4b1ff03947d5c0481a.400x400x1.jpg')">
</div>
</div>
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
<div style="text-align: center; font-size: 16px; font-weight: 800">Lady Gaga</div>
<a href="https://genius.com/artists/lady-gaga">
<div style="text-align: center; font-size: 14px;">@lady-gaga</div>
</a>
</div>
I was made with [huggingartists](https://github.com/AlekseyKorshuk/huggingartists).
Create your own bot based on your favorite artist with [the demo](https://colab.research.google.com/github/AlekseyKorshuk/huggingartists/blob/master/huggingartists-demo.ipynb)!
## How does it work?
To understand how the model was developed, check the [W&B report](https://wandb.ai/huggingartists/huggingartists/reportlist).
## Training data
The model was trained on lyrics from Lady Gaga.
Dataset is available [here](https://huggingface.co/datasets/huggingartists/lady-gaga).
And can be used with:
```python
from datasets import load_dataset
dataset = load_dataset("huggingartists/lady-gaga")
```
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/17c0d4ej/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
## Training procedure
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on Lady Gaga's lyrics.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/2j7yp9qd) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/2j7yp9qd/artifacts) is logged and versioned.
## How to use
You can use this model directly with a pipeline for text generation:
```python
from transformers import pipeline
generator = pipeline('text-generation',
model='huggingartists/lady-gaga')
generator("I am", num_return_sequences=5)
```
Or with Transformers library:
```python
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("huggingartists/lady-gaga")
model = AutoModelWithLMHead.from_pretrained("huggingartists/lady-gaga")
```
## Limitations and bias
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
In addition, the data present in the user's tweets further affects the text generated by the model.
## About
*Built by Aleksey Korshuk*
[](https://github.com/AlekseyKorshuk)
[](https://twitter.com/intent/follow?screen_name=alekseykorshuk)
[](https://t.me/joinchat/_CQ04KjcJ-4yZTky)
For more details, visit the project repository.
[](https://github.com/AlekseyKorshuk/huggingartists)
|
akozlo/lib_bal
|
akozlo
| 2022-03-08T20:19:15Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-08T20:14:41Z |
---
license: mit
tags:
- generated_from_trainer
model-index:
- name: lib_balanced_gpt_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# lib_balanced_gpt_model
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.15.0
- Pytorch 1.10.1+cu102
- Datasets 1.17.0
- Tokenizers 0.10.3
hello
|
OrfeasTsk/bert-base-uncased-finetuned-triviaqa
|
OrfeasTsk
| 2022-03-08T18:48:47Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"question-answering",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-08T18:38:16Z |
{ 'max_seq_length': 384,
'batch_size': 8,
'learning_rate': {'val': 5e-5, 'schelduler': 'Linear'},
'max_clip_norm': None,
'epochs': 2
}
|
OrfeasTsk/bert-base-uncased-finetuned-squadv2-large-batch
|
OrfeasTsk
| 2022-03-08T18:34:54Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"question-answering",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-08T18:23:33Z |
{ 'max_seq_length': 384,
'batch_size': 24,
'learning_rate': {'val': 3e-5, 'schelduler': 'Linear'},
'max_clip_norm': None,
'epochs': 2
}
|
z-uo/bert-qasper
|
z-uo
| 2022-03-08T18:31:21Z | 68 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"question-answering",
"en",
"dataset:z-uo/qasper-squad",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2022-03-08T09:28:02Z |
---
language: en
datasets:
- z-uo/qasper-squad
---
# bert-base for QA with qasper
Train from bert-base-uncased.
How to use by python code:
```python
from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
# Load model with pipeline
model_name = "z-uo/bert-qasper"
nlp = pipeline('question-answering', model=model_name, tokenizer=model_name)
# Get predictions
QA_input = {
'question': 'what they propose?',
'context': "In this paper, we provide an innovative contribution in the research domain dedicated to crop mapping by exploiting the of Sentinel-2 satellite images time series, with the specific aim to extract information on 'where and when' crops are grown. The final goal is to set up a workflow able to reliably identify (classify) the different crops that are grown in a given area by exploiting an end-to-end (3+2)D convolutional neural network (CNN) for semantic segmentation. The method also has the ambition to provide information, at pixel level, regarding the period in which a given crop is cultivated during the season. To this end, we propose a solution called Class Activation Interval (CAI) which allows us to interpret, for each pixel, the reasoning made by CNN in the classification determining in which time interval, of the input time series, the class is likely to be present or not. Our experiments, using a public domain dataset, show that the approach is able to accurately detect crop classes with an overall accuracy of about 93% and that the network can detect discriminatory time intervals in which crop is cultivated. These results have twofold importance: (i) demonstrate the ability of the network to correctly interpret the investigated physical process (i.e., bare soil condition, plant growth, senescence and harvesting according to specific cultivated variety) and (ii) provide further information to the end-user (e.g., the presence of crops and its temporal dynamics)."
}
res = nlp(QA_input)
# Load model & tokenizer without pipeline
model = AutoModelForQuestionAnswering.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
```
|
Ameer05/bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch-tweak-lr-8-100-1
|
Ameer05
| 2022-03-08T16:43:01Z | 9 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bart",
"text2text-generation",
"summarization",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-08T08:57:44Z |
---
tags:
- summarization
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch-tweak-lr-8-100-1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch-tweak-lr-8-100-1
This model is a fine-tuned version of [Ameer05/model-token-repo](https://huggingface.co/Ameer05/model-token-repo) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.6315
- Rouge1: 61.441
- Rouge2: 52.9403
- Rougel: 58.3426
- Rougelsum: 60.8249
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 100
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|
| No log | 0.91 | 5 | 2.0139 | 53.4301 | 46.6698 | 50.644 | 53.3985 |
| No log | 1.91 | 10 | 1.6309 | 61.4629 | 53.8884 | 59.0867 | 60.8823 |
| No log | 2.91 | 15 | 1.5379 | 61.2938 | 53.7208 | 59.0644 | 60.7381 |
| No log | 3.91 | 20 | 1.4470 | 63.2667 | 55.9273 | 60.5112 | 62.7538 |
| 1.5454 | 4.91 | 25 | 1.4353 | 62.7166 | 54.8328 | 60.0101 | 62.1378 |
| 1.5454 | 5.91 | 30 | 1.4411 | 59.7469 | 51.9068 | 57.036 | 58.9474 |
| 1.5454 | 6.91 | 35 | 1.5195 | 64.152 | 57.1447 | 61.362 | 63.5951 |
| 1.5454 | 7.91 | 40 | 1.6174 | 60.1464 | 51.5654 | 57.1676 | 59.4405 |
| 0.5429 | 8.91 | 45 | 1.7451 | 61.9696 | 53.6421 | 58.5884 | 61.3286 |
| 0.5429 | 9.91 | 50 | 1.9081 | 60.3296 | 52.3052 | 57.6518 | 59.7854 |
| 0.5429 | 10.91 | 55 | 1.9721 | 61.5597 | 51.9027 | 57.1184 | 60.6717 |
| 0.5429 | 11.91 | 60 | 2.0471 | 61.2222 | 53.9475 | 58.725 | 60.6668 |
| 0.5429 | 12.91 | 65 | 2.1422 | 60.1915 | 52.0627 | 56.9955 | 59.438 |
| 0.1506 | 13.91 | 70 | 2.1542 | 61.6915 | 53.045 | 58.1727 | 60.8765 |
| 0.1506 | 14.91 | 75 | 2.1885 | 59.8069 | 51.6543 | 56.8112 | 59.2055 |
| 0.1506 | 15.91 | 80 | 2.3146 | 61.695 | 53.2666 | 57.9003 | 61.1108 |
| 0.1506 | 16.91 | 85 | 2.3147 | 60.4482 | 52.1694 | 57.0649 | 59.7882 |
| 0.0452 | 17.91 | 90 | 2.1731 | 60.0259 | 51.5046 | 56.7399 | 59.2955 |
| 0.0452 | 18.91 | 95 | 2.2690 | 60.0534 | 52.4819 | 57.1631 | 59.5056 |
| 0.0452 | 19.91 | 100 | 2.2990 | 58.0737 | 48.8098 | 54.5684 | 57.3187 |
| 0.0452 | 20.91 | 105 | 2.2704 | 61.8982 | 53.9077 | 58.6909 | 61.4252 |
| 0.0267 | 21.91 | 110 | 2.3012 | 62.0174 | 53.5427 | 58.5278 | 61.1921 |
| 0.0267 | 22.91 | 115 | 2.3569 | 61.6327 | 53.7387 | 58.8908 | 61.1623 |
| 0.0267 | 23.91 | 120 | 2.3579 | 60.228 | 52.3747 | 58.1448 | 59.7322 |
| 0.0267 | 24.91 | 125 | 2.3389 | 60.4902 | 51.7935 | 57.0689 | 59.7132 |
| 0.0267 | 25.91 | 130 | 2.3168 | 58.8469 | 50.3181 | 55.7386 | 58.3598 |
| 0.0211 | 26.91 | 135 | 2.4147 | 59.4225 | 50.8405 | 56.503 | 58.7221 |
| 0.0211 | 27.91 | 140 | 2.3631 | 59.7489 | 51.2137 | 57.3204 | 59.3348 |
| 0.0211 | 28.91 | 145 | 2.3850 | 60.1718 | 51.4176 | 57.2152 | 59.5157 |
| 0.0211 | 29.91 | 150 | 2.4610 | 60.1433 | 51.433 | 56.6256 | 59.3265 |
| 0.0175 | 30.91 | 155 | 2.4400 | 58.8345 | 49.7031 | 55.3079 | 57.9236 |
| 0.0175 | 31.91 | 160 | 2.4506 | 59.209 | 50.1626 | 55.6451 | 58.5791 |
| 0.0175 | 32.91 | 165 | 2.4316 | 59.7713 | 50.8999 | 56.4235 | 58.9845 |
| 0.0175 | 33.91 | 170 | 2.2781 | 60.1822 | 51.9435 | 57.4586 | 59.6766 |
| 0.0175 | 34.91 | 175 | 2.3849 | 58.2328 | 49.2106 | 55.1516 | 57.5072 |
| 0.0141 | 35.91 | 180 | 2.4872 | 58.4916 | 50.3345 | 55.5991 | 58.1131 |
| 0.0141 | 36.91 | 185 | 2.4883 | 59.0957 | 49.76 | 55.3567 | 58.076 |
| 0.0141 | 37.91 | 190 | 2.4327 | 58.091 | 48.8628 | 54.8678 | 57.5406 |
| 0.0141 | 38.91 | 195 | 2.4998 | 57.7428 | 48.7366 | 54.2166 | 56.7643 |
| 0.0089 | 39.91 | 200 | 2.4107 | 60.1662 | 51.9832 | 57.1372 | 59.6989 |
| 0.0089 | 40.91 | 205 | 2.4700 | 58.2159 | 49.3934 | 54.9265 | 57.4126 |
| 0.0089 | 41.91 | 210 | 2.4833 | 58.7434 | 49.6619 | 55.5239 | 57.9562 |
| 0.0089 | 42.91 | 215 | 2.4703 | 60.2984 | 51.3168 | 56.9082 | 59.3958 |
| 0.0062 | 43.91 | 220 | 2.5306 | 60.5455 | 52.1189 | 57.3213 | 60.0232 |
| 0.0062 | 44.91 | 225 | 2.5181 | 60.2149 | 51.2187 | 56.1935 | 59.3471 |
| 0.0062 | 45.91 | 230 | 2.4871 | 59.8013 | 51.6114 | 56.0911 | 59.0902 |
| 0.0062 | 46.91 | 235 | 2.4811 | 58.0271 | 48.9441 | 54.3108 | 57.3647 |
| 0.0062 | 47.91 | 240 | 2.5290 | 62.5087 | 54.6149 | 59.638 | 62.0455 |
| 0.0072 | 48.91 | 245 | 2.5194 | 58.7193 | 49.9679 | 55.6517 | 58.1569 |
| 0.0072 | 49.91 | 250 | 2.5708 | 58.4626 | 49.5257 | 54.5032 | 58.1413 |
| 0.0072 | 50.91 | 255 | 2.6449 | 58.446 | 49.4625 | 55.1092 | 58.03 |
| 0.0072 | 51.91 | 260 | 2.5592 | 58.859 | 49.4398 | 55.1503 | 57.9663 |
| 0.0056 | 52.91 | 265 | 2.5086 | 59.7322 | 51.3051 | 56.5401 | 59.2726 |
| 0.0056 | 53.91 | 270 | 2.4846 | 57.8603 | 48.2408 | 54.3847 | 57.115 |
| 0.0056 | 54.91 | 275 | 2.5509 | 58.9506 | 50.045 | 55.6658 | 58.3618 |
| 0.0056 | 55.91 | 280 | 2.5032 | 60.2524 | 51.8167 | 56.98 | 59.7506 |
| 0.0056 | 56.91 | 285 | 2.5012 | 60.0596 | 51.4924 | 56.7181 | 59.5037 |
| 0.0054 | 57.91 | 290 | 2.5176 | 61.0622 | 52.6235 | 57.9317 | 60.5036 |
| 0.0054 | 58.91 | 295 | 2.5024 | 62.9246 | 54.8544 | 59.9824 | 62.5584 |
| 0.0054 | 59.91 | 300 | 2.5687 | 62.2602 | 53.9673 | 58.9862 | 61.5837 |
| 0.0054 | 60.91 | 305 | 2.5890 | 62.5706 | 54.227 | 59.2032 | 62.125 |
| 0.0036 | 61.91 | 310 | 2.5454 | 62.1565 | 53.2585 | 58.7169 | 61.3943 |
| 0.0036 | 62.91 | 315 | 2.5629 | 62.8292 | 54.6781 | 59.9889 | 62.254 |
| 0.0036 | 63.91 | 320 | 2.5581 | 58.8394 | 50.4421 | 56.0742 | 58.1945 |
| 0.0036 | 64.91 | 325 | 2.5532 | 59.5814 | 51.1335 | 56.5841 | 59.196 |
| 0.0031 | 65.91 | 330 | 2.5826 | 59.0485 | 50.3992 | 55.5283 | 58.3757 |
| 0.0031 | 66.91 | 335 | 2.5815 | 61.4832 | 52.7977 | 57.7351 | 60.9888 |
| 0.0031 | 67.91 | 340 | 2.5865 | 61.7836 | 53.6797 | 58.6743 | 61.3765 |
| 0.0031 | 68.91 | 345 | 2.6007 | 61.2253 | 52.8781 | 57.7006 | 60.717 |
| 0.0031 | 69.91 | 350 | 2.6210 | 60.717 | 52.4933 | 57.5089 | 60.4196 |
| 0.0035 | 70.91 | 355 | 2.6169 | 61.3491 | 53.3932 | 58.2288 | 60.8793 |
| 0.0035 | 71.91 | 360 | 2.6025 | 62.0101 | 54.0289 | 59.0822 | 61.7202 |
| 0.0035 | 72.91 | 365 | 2.5705 | 61.2227 | 52.9937 | 58.2493 | 60.6631 |
| 0.0035 | 73.91 | 370 | 2.5623 | 59.1718 | 50.7827 | 56.1851 | 58.7118 |
| 0.002 | 74.91 | 375 | 2.5536 | 58.4201 | 49.6923 | 55.0398 | 57.7707 |
| 0.002 | 75.91 | 380 | 2.5478 | 60.2307 | 51.7503 | 57.3173 | 59.692 |
| 0.002 | 76.91 | 385 | 2.6039 | 58.7637 | 49.741 | 55.5341 | 58.0784 |
| 0.002 | 77.91 | 390 | 2.6371 | 59.3929 | 50.6444 | 55.9887 | 58.813 |
| 0.002 | 78.91 | 395 | 2.6238 | 59.0572 | 50.605 | 55.6631 | 58.4366 |
| 0.0019 | 79.91 | 400 | 2.5783 | 57.9852 | 49.2588 | 54.822 | 57.4643 |
| 0.0019 | 80.91 | 405 | 2.5982 | 58.0218 | 49.1651 | 54.9876 | 57.4066 |
| 0.0019 | 81.91 | 410 | 2.6141 | 60.3133 | 51.5723 | 56.9476 | 59.715 |
| 0.0019 | 82.91 | 415 | 2.5904 | 60.8199 | 51.8956 | 58.406 | 60.323 |
| 0.0017 | 83.91 | 420 | 2.5718 | 60.3449 | 51.1433 | 57.6984 | 59.7513 |
| 0.0017 | 84.91 | 425 | 2.5737 | 60.151 | 51.1986 | 57.3376 | 59.378 |
| 0.0017 | 85.91 | 430 | 2.5807 | 60.9273 | 52.2469 | 58.2038 | 60.1642 |
| 0.0017 | 86.91 | 435 | 2.5900 | 60.1846 | 51.6144 | 57.5407 | 59.5109 |
| 0.0011 | 87.91 | 440 | 2.6066 | 62.0776 | 53.6022 | 59.157 | 61.6201 |
| 0.0011 | 88.91 | 445 | 2.6231 | 61.8822 | 53.5232 | 58.965 | 61.401 |
| 0.0011 | 89.91 | 450 | 2.6273 | 60.3358 | 51.9941 | 57.3823 | 59.7729 |
| 0.0011 | 90.91 | 455 | 2.6194 | 60.0196 | 51.6134 | 57.1357 | 59.4594 |
| 0.0011 | 91.91 | 460 | 2.6118 | 60.6898 | 52.1328 | 57.3076 | 60.0351 |
| 0.0015 | 92.91 | 465 | 2.6032 | 61.2119 | 52.5034 | 57.8098 | 60.6634 |
| 0.0015 | 93.91 | 470 | 2.6040 | 61.4812 | 52.8197 | 57.9668 | 60.8767 |
| 0.0015 | 94.91 | 475 | 2.6158 | 61.4046 | 52.8905 | 57.8958 | 60.804 |
| 0.0015 | 95.91 | 480 | 2.6280 | 62.1764 | 53.8521 | 58.8608 | 61.6138 |
| 0.0012 | 96.91 | 485 | 2.6304 | 62.2028 | 53.8967 | 58.8976 | 61.6409 |
| 0.0012 | 97.91 | 490 | 2.6328 | 61.7371 | 53.3908 | 58.4107 | 61.1382 |
| 0.0012 | 98.91 | 495 | 2.6331 | 61.441 | 52.9403 | 58.3426 | 60.8249 |
| 0.0012 | 99.91 | 500 | 2.6315 | 61.441 | 52.9403 | 58.3426 | 60.8249 |
### Framework versions
- Transformers 4.15.0
- Pytorch 1.9.1
- Datasets 1.18.4
- Tokenizers 0.10.3
|
gayanin/bart-med-term-mlm
|
gayanin
| 2022-03-08T15:46:48Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bart",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-08T12:09:15Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: bart-med-term-mlm
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-med-term-mlm
This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2506
- Rouge2 Precision: 0.8338
- Rouge2 Recall: 0.6005
- Rouge2 Fmeasure: 0.6775
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge2 Precision | Rouge2 Recall | Rouge2 Fmeasure |
|:-------------:|:-----:|:-----:|:---------------:|:----------------:|:-------------:|:---------------:|
| 0.3426 | 1.0 | 15827 | 0.3029 | 0.8184 | 0.5913 | 0.6664 |
| 0.2911 | 2.0 | 31654 | 0.2694 | 0.8278 | 0.5963 | 0.6727 |
| 0.2571 | 3.0 | 47481 | 0.2549 | 0.8318 | 0.5985 | 0.6753 |
| 0.2303 | 4.0 | 63308 | 0.2506 | 0.8338 | 0.6005 | 0.6775 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
mmgyorke/vit-world-landmarks
|
mmgyorke
| 2022-03-08T14:58:47Z | 72 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"vit",
"image-classification",
"huggingpics",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
image-classification
| 2022-03-08T14:40:15Z |
---
tags:
- image-classification
- pytorch
- huggingpics
metrics:
- accuracy
model-index:
- name: vit-world-landmarks
results:
- task:
name: Image Classification
type: image-classification
metrics:
- name: Accuracy
type: accuracy
value: 1.0
---
# vit-world-landmarks
Autogenerated by HuggingPics🤗🖼️
Create your own image classifier for **anything** by running [the demo on Google Colab](https://colab.research.google.com/github/nateraw/huggingpics/blob/main/HuggingPics.ipynb).
Report any issues with the demo at the [github repo](https://github.com/nateraw/huggingpics).
## Example Images
#### arc de triomphe

#### big ben

#### la sagrada familia

#### leaning tower of pisa

#### taj mahal

|
AlekseyKorshuk/bert-finetuned-ner
|
AlekseyKorshuk
| 2022-03-08T14:27:56Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bert",
"token-classification",
"generated_from_trainer",
"dataset:wnut_17",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2022-03-08T12:40:05Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- wnut_17
model-index:
- name: bert-finetuned-ner
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-finetuned-ner
This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the wnut_17 dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log | 1.0 | 425 | 0.3961 | 0.5707 | 0.2847 | 0.3799 | 0.9058 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
anniezx/111
|
anniezx
| 2022-03-08T08:38:09Z | 0 | 0 | null |
[
"license:artistic-2.0",
"region:us"
] | null | 2022-03-08T08:38:09Z |
---
license: artistic-2.0
---
|
Narsil/totallysafe
|
Narsil
| 2022-03-08T08:23:38Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"gpt2",
"text-generation",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-07T22:13:40Z |
---
pipeline_tag: text-generation
---
This is a totally safe and groundbreaking model.
GPT3 performance with under 10Mo model.
|
huggingtweets/betonkoepfin-littlehorney-plusbibi1
|
huggingtweets
| 2022-03-08T07:46:04Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"huggingtweets",
"en",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-08T07:44:50Z |
---
language: en
thumbnail: http://www.huggingtweets.com/betonkoepfin-littlehorney-plusbibi1/1646725560421/predictions.png
tags:
- huggingtweets
widget:
- text: "My dream is"
---
<div class="inline-flex flex-col" style="line-height: 1.5;">
<div class="flex">
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1386970823681052680/oA_4HBKl_400x400.jpg')">
</div>
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1425205160578588673/LBMG1HOO_400x400.jpg')">
</div>
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1500892464772751365/6uhqt-Jx_400x400.jpg')">
</div>
</div>
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 AI CYBORG 🤖</div>
<div style="text-align: center; font-size: 16px; font-weight: 800">Bibi und Anna & Betty S. & Vanny_Bunny™</div>
<div style="text-align: center; font-size: 14px;">@betonkoepfin-littlehorney-plusbibi1</div>
</div>
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
Create your own bot based on your favorite user with [the demo](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb)!
## How does it work?
The model uses the following pipeline.

To understand how the model was developed, check the [W&B report](https://wandb.ai/wandb/huggingtweets/reports/HuggingTweets-Train-a-Model-to-Generate-Tweets--VmlldzoxMTY5MjI).
## Training data
The model was trained on tweets from Bibi und Anna & Betty S. & Vanny_Bunny™.
| Data | Bibi und Anna | Betty S. | Vanny_Bunny™ |
| --- | --- | --- | --- |
| Tweets downloaded | 1818 | 3243 | 3185 |
| Retweets | 9 | 213 | 494 |
| Short tweets | 341 | 552 | 339 |
| Tweets kept | 1468 | 2478 | 2352 |
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/3nxb6yoh/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
## Training procedure
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @betonkoepfin-littlehorney-plusbibi1's tweets.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/365gy60z) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/365gy60z/artifacts) is logged and versioned.
## How to use
You can use this model directly with a pipeline for text generation:
```python
from transformers import pipeline
generator = pipeline('text-generation',
model='huggingtweets/betonkoepfin-littlehorney-plusbibi1')
generator("My dream is", num_return_sequences=5)
```
## Limitations and bias
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
In addition, the data present in the user's tweets further affects the text generated by the model.
## About
*Built by Boris Dayma*
[](https://twitter.com/intent/follow?screen_name=borisdayma)
For more details, visit the project repository.
[](https://github.com/borisdayma/huggingtweets)
|
zhiweitong/dpr-ctx_encoder-single-nq-base
|
zhiweitong
| 2022-03-08T07:28:29Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"dpr",
"en",
"dataset:wiki_dpr",
"dataset:natural_questions",
"endpoints_compatible",
"region:us"
] | null | 2022-03-07T12:19:50Z |
---
language: en
datasets:
- wiki_dpr
- natural_questions
---
# dpr-ctx_encoder-single-nq-base
This encoder is used with [zhiweitong/dpr-answer_encoder-single-nq-base](https://huggingface.co/zhiweitong/dpr-answer_encoder-single-nq-base)
|
zhiweitong/dpr-answer_encoder-single-nq-base
|
zhiweitong
| 2022-03-08T07:25:05Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"dpr",
"feature-extraction",
"en",
"dataset:natural_questions",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2022-03-07T07:49:38Z |
---
language: en
datasets:
- natural_questions
---
# dpr-answer_encoder-single-nq-base
This encoder is used with [zhiweitong/dpr-ctx_encoder-single-nq-base](https://huggingface.co/zhiweitong/dpr-ctx_encoder-single-nq-base)
|
MikhailGalperin/distilbert-base-uncased-finetuned-ner
|
MikhailGalperin
| 2022-03-08T06:49:43Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"token-classification",
"generated_from_trainer",
"dataset:conll2003",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2022-03-07T20:29:52Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- conll2003
model-index:
- name: distilbert-base-uncased-finetuned-ner
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-ner
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the conll2003 dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
Ameer05/bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch
|
Ameer05
| 2022-03-08T05:53:14Z | 9 | 1 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"bart",
"text2text-generation",
"summarization",
"generated_from_trainer",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
summarization
| 2022-03-08T05:33:06Z |
---
tags:
- summarization
- generated_from_trainer
metrics:
- rouge
model-index:
- name: bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-large-cnn-samsum-rescom-finetuned-resume-summarizer-10-epoch
This model is a fine-tuned version of [Ameer05/model-token-repo](https://huggingface.co/Ameer05/model-token-repo) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5216
- Rouge1: 59.5791
- Rouge2: 51.3273
- Rougel: 56.9984
- Rougelsum: 59.1424
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5.6e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|
| No log | 0.91 | 5 | 2.0124 | 53.776 | 46.7427 | 50.7565 | 53.5502 |
| No log | 1.91 | 10 | 1.6353 | 61.8019 | 53.8614 | 58.9744 | 61.339 |
| No log | 2.91 | 15 | 1.5321 | 59.7045 | 51.5968 | 57.0823 | 59.2417 |
| No log | 3.91 | 20 | 1.4569 | 62.4379 | 54.5464 | 59.9202 | 61.9242 |
| 1.5608 | 4.91 | 25 | 1.4613 | 63.3808 | 55.8818 | 61.432 | 63.0208 |
| 1.5608 | 5.91 | 30 | 1.4321 | 59.6761 | 50.9812 | 56.7977 | 59.1214 |
| 1.5608 | 6.91 | 35 | 1.4753 | 62.6439 | 54.7158 | 60.3831 | 62.1046 |
| 1.5608 | 7.91 | 40 | 1.4783 | 60.2735 | 52.7462 | 57.77 | 59.9725 |
| 0.6428 | 8.91 | 45 | 1.4974 | 62.8691 | 54.9062 | 60.3496 | 62.5132 |
| 0.6428 | 9.91 | 50 | 1.5216 | 59.5791 | 51.3273 | 56.9984 | 59.1424 |
### Framework versions
- Transformers 4.15.0
- Pytorch 1.9.1
- Datasets 1.18.4
- Tokenizers 0.10.3
|
jiobiala24/wav2vec2-base-cv
|
jiobiala24
| 2022-03-08T05:42:48Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"dataset:common_voice",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2022-03-08T00:03:37Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-base-cv
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-base-cv
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the common_voice dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1562
- Wer: 0.3804
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 30
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 4.563 | 3.18 | 500 | 2.9826 | 1.0 |
| 2.0012 | 6.37 | 1000 | 0.9528 | 0.5354 |
| 0.4841 | 9.55 | 1500 | 0.8838 | 0.4325 |
| 0.2748 | 12.74 | 2000 | 0.9437 | 0.4130 |
| 0.1881 | 15.92 | 2500 | 0.9603 | 0.4005 |
| 0.1426 | 19.11 | 3000 | 1.0605 | 0.3955 |
| 0.1134 | 22.29 | 3500 | 1.0733 | 0.3897 |
| 0.0963 | 25.48 | 4000 | 1.1387 | 0.3835 |
| 0.0829 | 28.66 | 4500 | 1.1562 | 0.3804 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.10.0+cu111
- Datasets 1.13.3
- Tokenizers 0.10.3
|
huggingtweets/fitdollar
|
huggingtweets
| 2022-03-08T05:18:01Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"huggingtweets",
"en",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-08T05:17:20Z |
---
language: en
thumbnail: http://www.huggingtweets.com/fitdollar/1646716677087/predictions.png
tags:
- huggingtweets
widget:
- text: "My dream is"
---
<div class="inline-flex flex-col" style="line-height: 1.5;">
<div class="flex">
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1421952831796350976/rFuw5k2v_400x400.jpg')">
</div>
<div
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
</div>
<div
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
</div>
</div>
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 AI BOT 🤖</div>
<div style="text-align: center; font-size: 16px; font-weight: 800">Fit$</div>
<div style="text-align: center; font-size: 14px;">@fitdollar</div>
</div>
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
Create your own bot based on your favorite user with [the demo](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb)!
## How does it work?
The model uses the following pipeline.

To understand how the model was developed, check the [W&B report](https://wandb.ai/wandb/huggingtweets/reports/HuggingTweets-Train-a-Model-to-Generate-Tweets--VmlldzoxMTY5MjI).
## Training data
The model was trained on tweets from Fit$.
| Data | Fit$ |
| --- | --- |
| Tweets downloaded | 1235 |
| Retweets | 139 |
| Short tweets | 219 |
| Tweets kept | 877 |
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/1nxpnpfh/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
## Training procedure
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @fitdollar's tweets.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/3f78vjfv) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/3f78vjfv/artifacts) is logged and versioned.
## How to use
You can use this model directly with a pipeline for text generation:
```python
from transformers import pipeline
generator = pipeline('text-generation',
model='huggingtweets/fitdollar')
generator("My dream is", num_return_sequences=5)
```
## Limitations and bias
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
In addition, the data present in the user's tweets further affects the text generated by the model.
## About
*Built by Boris Dayma*
[](https://twitter.com/intent/follow?screen_name=borisdayma)
For more details, visit the project repository.
[](https://github.com/borisdayma/huggingtweets)
|
oskrmiguel/t5-small-finetuned-es-to-pt
|
oskrmiguel
| 2022-03-08T03:15:16Z | 7 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"dataset:tatoeba",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-08T02:54:55Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- tatoeba
metrics:
- bleu
model-index:
- name: t5-small-finetuned-es-to-pt
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: tatoeba
type: tatoeba
args: es-pt
metrics:
- name: Bleu
type: bleu
value: 15.0473
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# t5-small-finetuned-es-to-pt
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the tatoeba dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5557
- Bleu: 15.0473
- Gen Len: 15.8693
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
| 2.2027 | 1.0 | 1907 | 1.7884 | 11.6192 | 15.8829 |
| 1.9296 | 2.0 | 3814 | 1.6034 | 14.201 | 15.8935 |
| 1.8364 | 3.0 | 5721 | 1.5557 | 15.0473 | 15.8693 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
gayanin/t5-small-paraphrasing-mlm
|
gayanin
| 2022-03-08T01:54:54Z | 10 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-07T21:54:14Z |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: t5-small-paraphrasing-mlm
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# t5-small-paraphrasing-mlm
This model is a fine-tuned version of [gayanin/t5-small-paraphrase-pubmed](https://huggingface.co/gayanin/t5-small-paraphrase-pubmed) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7030
- Rouge2 Precision: 0.6576
- Rouge2 Recall: 0.4712
- Rouge2 Fmeasure: 0.532
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge2 Precision | Rouge2 Recall | Rouge2 Fmeasure |
|:-------------:|:-----:|:------:|:---------------:|:----------------:|:-------------:|:---------------:|
| 0.9215 | 1.0 | 13833 | 0.8050 | 0.6352 | 0.454 | 0.5131 |
| 0.855 | 2.0 | 27666 | 0.7679 | 0.6411 | 0.4589 | 0.5184 |
| 0.8387 | 3.0 | 41499 | 0.7464 | 0.6464 | 0.4626 | 0.5226 |
| 0.8267 | 4.0 | 55332 | 0.7315 | 0.6513 | 0.4671 | 0.5273 |
| 0.7879 | 5.0 | 69165 | 0.7217 | 0.6534 | 0.4687 | 0.529 |
| 0.7738 | 6.0 | 82998 | 0.7142 | 0.6548 | 0.4688 | 0.5295 |
| 0.7793 | 7.0 | 96831 | 0.7094 | 0.6553 | 0.4694 | 0.53 |
| 0.7654 | 8.0 | 110664 | 0.7056 | 0.6573 | 0.4704 | 0.5313 |
| 0.7675 | 9.0 | 124497 | 0.7036 | 0.6577 | 0.4712 | 0.532 |
| 0.7662 | 10.0 | 138330 | 0.7030 | 0.6576 | 0.4712 | 0.532 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6
|
huggingtweets/lilbratmia-littlehorney-plusbibi1
|
huggingtweets
| 2022-03-07T21:45:31Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"huggingtweets",
"en",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2022-03-07T21:35:06Z |
---
language: en
thumbnail: http://www.huggingtweets.com/lilbratmia-littlehorney-plusbibi1/1646689525715/predictions.png
tags:
- huggingtweets
widget:
- text: "My dream is"
---
<div class="inline-flex flex-col" style="line-height: 1.5;">
<div class="flex">
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1386970823681052680/oA_4HBKl_400x400.jpg')">
</div>
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1500892464772751365/6uhqt-Jx_400x400.jpg')">
</div>
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1483439308166123530/vKFDbs48_400x400.jpg')">
</div>
</div>
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 AI CYBORG 🤖</div>
<div style="text-align: center; font-size: 16px; font-weight: 800">Bibi und Anna & Vanny_Bunny™ & 💞 Mia 💞</div>
<div style="text-align: center; font-size: 14px;">@lilbratmia-littlehorney-plusbibi1</div>
</div>
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
Create your own bot based on your favorite user with [the demo](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb)!
## How does it work?
The model uses the following pipeline.

To understand how the model was developed, check the [W&B report](https://wandb.ai/wandb/huggingtweets/reports/HuggingTweets-Train-a-Model-to-Generate-Tweets--VmlldzoxMTY5MjI).
## Training data
The model was trained on tweets from Bibi und Anna & Vanny_Bunny™ & 💞 Mia 💞.
| Data | Bibi und Anna | Vanny_Bunny™ | 💞 Mia 💞 |
| --- | --- | --- | --- |
| Tweets downloaded | 1818 | 3230 | 3247 |
| Retweets | 9 | 503 | 134 |
| Short tweets | 341 | 343 | 1189 |
| Tweets kept | 1468 | 2384 | 1924 |
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/hm55g9hx/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
## Training procedure
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @lilbratmia-littlehorney-plusbibi1's tweets.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/3dezdv7k) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/3dezdv7k/artifacts) is logged and versioned.
## How to use
You can use this model directly with a pipeline for text generation:
```python
from transformers import pipeline
generator = pipeline('text-generation',
model='huggingtweets/lilbratmia-littlehorney-plusbibi1')
generator("My dream is", num_return_sequences=5)
```
## Limitations and bias
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
In addition, the data present in the user's tweets further affects the text generated by the model.
## About
*Built by Boris Dayma*
[](https://twitter.com/intent/follow?screen_name=borisdayma)
For more details, visit the project repository.
[](https://github.com/borisdayma/huggingtweets)
|
Manauu17/roberta_sentiments_es
|
Manauu17
| 2022-03-07T20:10:33Z | 4 | 2 |
transformers
|
[
"transformers",
"pytorch",
"tf",
"roberta",
"text-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2022-03-03T13:50:56Z |
# roberta_sentiments_es , a Sentiment Analysis model for Spanish sentences
This is a roBERTa-base model trained on ~58M tweets and finetuned for sentiment analysis. This model currently supports Spanish sentences
## Example of classification
```python
from transformers import AutoModelForSequenceClassification
from transformers import TFAutoModelForSequenceClassification
from transformers import AutoTokenizer
import numpy as np
import pandas as pd
from scipy.special import softmax
MODEL = 'Manauu17/roberta_sentiments_es_en'
tokenizer = AutoTokenizer.from_pretrained(MODEL)
# PyTorch
model = AutoModelForSequenceClassification.from_pretrained(MODEL)
text = ['@usuario siempre es bueno la opinión de un playo',
'Bendito año el que me espera']
encoded_input = tokenizer(text, return_tensors='pt', padding=True, truncation=True)
output = model(**encoded_input)
scores = output[0].detach().numpy()
# TensorFlow
model = TFAutoModelForSequenceClassification.from_pretrained(MODEL)
text = ['La guerra no es buena para nadie.','Espero que mi jefe me de mañana libre']
encoded_input = tokenizer(text, return_tensors='tf', padding=True, truncation=True)
output = model(encoded_input)
scores = output[0].numpy()
# Results
def get_scores(model_output, labels_dict):
scores = softmax(model_output)
frame = pd.DataFrame(scores, columns=labels.values())
frame.style.highlight_max(axis=1,color="green")
return frame
```
Output:
```
# PyTorch
get_scores(scores, labels_dict).style.highlight_max(axis=1, color="green")
Negative Neutral Positive
0 0.000607 0.004851 0.906596
1 0.079812 0.006650 0.001484
# TensorFlow
get_scores(scores, labels_dict).style.highlight_max(axis=1, color="green")
Negative Neutral Positive
0 0.017030 0.008920 0.000667
1 0.000260 0.001695 0.971429
```
|
espnet/Karthik_DSTC2_asr_train_asr_wav2vec_transformer
|
espnet
| 2022-03-07T19:38:16Z | 1 | 0 |
espnet
|
[
"espnet",
"tensorboard",
"audio",
"automatic-speech-recognition",
"en",
"dataset:sinhala",
"arxiv:1804.00015",
"license:cc-by-4.0",
"region:us"
] |
automatic-speech-recognition
| 2022-03-07T16:09:26Z |
---
tags:
- espnet
- audio
- automatic-speech-recognition
language: en
datasets:
- sinhala
license: cc-by-4.0
---
## ESPnet2 ASR pretrained model
### `espnet/Karthik_DSTC2_asr_train_asr_wav2vec_transformer`
This model was trained by Karthik using DSTC2/asr1 recipe in [espnet](https://github.com/espnet/espnet/).
### Demo: How to use in ESPnet2
```python
# coming soon
```
### Citing ESPnet
```BibTex
@inproceedings{watanabe2018espnet,
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson {Enrique Yalta Soplin} and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
title={{ESPnet}: End-to-End Speech Processing Toolkit},
year={2018},
booktitle={Proceedings of Interspeech},
pages={2207--2211},
doi={10.21437/Interspeech.2018-1456},
url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
}
```
or arXiv:
```bibtex
@misc{watanabe2018espnet,
title={ESPnet: End-to-End Speech Processing Toolkit},
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Enrique Yalta Soplin and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
year={2018},
eprint={1804.00015},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
pyf98/librispeech_conformer
|
pyf98
| 2022-03-07T18:33:17Z | 4 | 0 |
espnet
|
[
"espnet",
"audio",
"automatic-speech-recognition",
"en",
"dataset:librispeech",
"arxiv:1804.00015",
"license:cc-by-4.0",
"region:us"
] |
automatic-speech-recognition
| 2022-03-07T18:16:05Z |
---
tags:
- espnet
- audio
- automatic-speech-recognition
language: en
datasets:
- librispeech
license: cc-by-4.0
---
## ESPnet2 ASR model
### `pyf98/librispeech_conformer`
This model was trained by Yifan Peng using librispeech recipe in [espnet](https://github.com/espnet/espnet/).
### Demo: How to use in ESPnet2
```bash
cd espnet
git checkout c3569453a408fd4ff4173d9c1d2062c88d1fc060
pip install -e .
cd egs2/librispeech/asr1
./run.sh --skip_data_prep false --skip_train true --download_model pyf98/librispeech_conformer
```
<!-- Generated by scripts/utils/show_asr_result.sh -->
# RESULTS
## Environments
- date: `Mon Mar 7 12:26:10 EST 2022`
- python version: `3.9.7 (default, Sep 16 2021, 13:09:58) [GCC 7.5.0]`
- espnet version: `espnet 0.10.7a1`
- pytorch version: `pytorch 1.10.1`
- Git hash: `c3569453a408fd4ff4173d9c1d2062c88d1fc060`
- Commit date: `Sun Mar 6 23:58:36 2022 -0500`
## asr_train_asr_conformer8_raw_en_bpe5000_sp
### WER
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
|beam60_ctc0.2/dev_clean|2703|54402|98.0|1.8|0.2|0.2|2.2|27.2|
|beam60_ctc0.2/dev_other|2864|50948|95.1|4.4|0.5|0.5|5.4|43.3|
|beam60_ctc0.2/test_clean|2620|52576|97.9|1.9|0.2|0.3|2.4|28.8|
|beam60_ctc0.2/test_other|2939|52343|95.2|4.3|0.5|0.6|5.4|45.5|
|beam60_ctc0.2_lm0.6/dev_clean|2703|54402|98.3|1.4|0.3|0.2|1.9|23.7|
|beam60_ctc0.2_lm0.6/dev_other|2864|50948|96.2|3.3|0.4|0.4|4.2|37.2|
|beam60_ctc0.2_lm0.6/test_clean|2620|52576|98.2|1.5|0.3|0.2|2.0|24.3|
|beam60_ctc0.2_lm0.6/test_other|2939|52343|96.1|3.3|0.6|0.4|4.4|39.9|
|beam60_ctc0.3/dev_clean|2703|54402|98.1|1.8|0.2|0.2|2.1|27.3|
|beam60_ctc0.3/dev_other|2864|50948|95.2|4.4|0.4|0.5|5.4|43.7|
|beam60_ctc0.3/test_clean|2620|52576|97.9|1.9|0.2|0.3|2.3|29.0|
|beam60_ctc0.3/test_other|2939|52343|95.2|4.3|0.4|0.6|5.4|45.7|
|beam60_ctc0.3_lm0.6/dev_clean|2703|54402|98.4|1.4|0.2|0.2|1.8|23.5|
|beam60_ctc0.3_lm0.6/dev_other|2864|50948|96.2|3.4|0.4|0.4|4.1|37.4|
|beam60_ctc0.3_lm0.6/test_clean|2620|52576|98.3|1.5|0.2|0.2|1.9|24.1|
|beam60_ctc0.3_lm0.6/test_other|2939|52343|96.2|3.3|0.5|0.5|4.3|39.9|
### CER
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
|beam60_ctc0.2/dev_clean|2703|288456|99.4|0.3|0.3|0.2|0.8|27.2|
|beam60_ctc0.2/dev_other|2864|265951|98.1|1.1|0.8|0.6|2.5|43.3|
|beam60_ctc0.2/test_clean|2620|281530|99.4|0.3|0.3|0.2|0.8|28.8|
|beam60_ctc0.2/test_other|2939|272758|98.3|1.0|0.7|0.6|2.3|45.5|
|beam60_ctc0.2_lm0.6/dev_clean|2703|288456|99.4|0.3|0.3|0.2|0.8|23.7|
|beam60_ctc0.2_lm0.6/dev_other|2864|265951|98.4|0.9|0.7|0.5|2.1|37.2|
|beam60_ctc0.2_lm0.6/test_clean|2620|281530|99.4|0.2|0.4|0.2|0.8|24.3|
|beam60_ctc0.2_lm0.6/test_other|2939|272758|98.5|0.8|0.8|0.5|2.0|39.9|
|beam60_ctc0.3/dev_clean|2703|288456|99.5|0.3|0.2|0.2|0.7|27.3|
|beam60_ctc0.3/dev_other|2864|265951|98.2|1.1|0.7|0.6|2.4|43.7|
|beam60_ctc0.3/test_clean|2620|281530|99.4|0.3|0.3|0.2|0.8|29.0|
|beam60_ctc0.3/test_other|2939|272758|98.4|0.9|0.7|0.6|2.2|45.7|
|beam60_ctc0.3_lm0.6/dev_clean|2703|288456|99.5|0.2|0.2|0.2|0.7|23.5|
|beam60_ctc0.3_lm0.6/dev_other|2864|265951|98.5|0.9|0.7|0.5|2.0|37.4|
|beam60_ctc0.3_lm0.6/test_clean|2620|281530|99.5|0.2|0.3|0.2|0.7|24.1|
|beam60_ctc0.3_lm0.6/test_other|2939|272758|98.6|0.7|0.7|0.5|1.9|39.9|
### TER
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
|beam60_ctc0.2/dev_clean|2703|68010|97.5|1.8|0.7|0.3|2.9|27.2|
|beam60_ctc0.2/dev_other|2864|63110|94.1|4.4|1.6|0.9|6.8|43.3|
|beam60_ctc0.2/test_clean|2620|65818|97.4|1.8|0.8|0.3|2.9|28.8|
|beam60_ctc0.2/test_other|2939|65101|94.1|4.1|1.8|0.8|6.7|45.5|
|beam60_ctc0.2_lm0.6/dev_clean|2703|68010|97.8|1.4|0.8|0.3|2.5|23.7|
|beam60_ctc0.2_lm0.6/dev_other|2864|63110|95.1|3.5|1.5|0.7|5.6|37.2|
|beam60_ctc0.2_lm0.6/test_clean|2620|65818|97.6|1.5|0.9|0.3|2.7|24.3|
|beam60_ctc0.2_lm0.6/test_other|2939|65101|95.0|3.2|1.8|0.6|5.6|39.9|
|beam60_ctc0.3/dev_clean|2703|68010|97.6|1.8|0.7|0.3|2.8|27.3|
|beam60_ctc0.3/dev_other|2864|63110|94.1|4.4|1.5|0.9|6.8|43.7|
|beam60_ctc0.3/test_clean|2620|65818|97.4|1.8|0.7|0.3|2.9|29.0|
|beam60_ctc0.3/test_other|2939|65101|94.2|4.1|1.7|0.8|6.6|45.7|
|beam60_ctc0.3_lm0.6/dev_clean|2703|68010|97.9|1.5|0.7|0.3|2.4|23.5|
|beam60_ctc0.3_lm0.6/dev_other|2864|63110|95.1|3.5|1.4|0.6|5.6|37.4|
|beam60_ctc0.3_lm0.6/test_clean|2620|65818|97.7|1.5|0.8|0.3|2.5|24.1|
|beam60_ctc0.3_lm0.6/test_other|2939|65101|95.1|3.2|1.7|0.6|5.5|39.9|
## ASR config
<details><summary>expand</summary>
```
config: conf/tuning/train_asr_conformer8.yaml
print_config: false
log_level: INFO
dry_run: false
iterator_type: sequence
output_dir: exp/asr_train_asr_conformer8_raw_en_bpe5000_sp
ngpu: 1
seed: 0
num_workers: 4
num_att_plot: 3
dist_backend: nccl
dist_init_method: env://
dist_world_size: 3
dist_rank: 0
local_rank: 0
dist_master_addr: localhost
dist_master_port: 59673
dist_launcher: null
multiprocessing_distributed: true
unused_parameters: false
sharded_ddp: false
cudnn_enabled: true
cudnn_benchmark: false
cudnn_deterministic: true
collect_stats: false
write_collected_feats: false
max_epoch: 50
patience: null
val_scheduler_criterion:
- valid
- loss
early_stopping_criterion:
- valid
- loss
- min
best_model_criterion:
- - valid
- acc
- max
keep_nbest_models: 10
nbest_averaging_interval: 0
grad_clip: 5.0
grad_clip_type: 2.0
grad_noise: false
accum_grad: 4
no_forward_run: false
resume: true
train_dtype: float32
use_amp: true
log_interval: null
use_matplotlib: true
use_tensorboard: true
use_wandb: false
wandb_project: null
wandb_id: null
wandb_entity: null
wandb_name: null
wandb_model_log_interval: -1
detect_anomaly: false
pretrain_path: null
init_param: []
ignore_init_mismatch: false
freeze_param: []
num_iters_per_epoch: null
batch_size: 20
valid_batch_size: null
batch_bins: 35000000
valid_batch_bins: null
train_shape_file:
- exp/asr_stats_raw_en_bpe5000_sp/train/speech_shape
- exp/asr_stats_raw_en_bpe5000_sp/train/text_shape.bpe
valid_shape_file:
- exp/asr_stats_raw_en_bpe5000_sp/valid/speech_shape
- exp/asr_stats_raw_en_bpe5000_sp/valid/text_shape.bpe
batch_type: numel
valid_batch_type: null
fold_length:
- 80000
- 150
sort_in_batch: descending
sort_batch: descending
multiple_iterator: false
chunk_length: 500
chunk_shift_ratio: 0.5
num_cache_chunks: 1024
train_data_path_and_name_and_type:
- - dump/raw/train_960_sp/wav.scp
- speech
- sound
- - dump/raw/train_960_sp/text
- text
- text
valid_data_path_and_name_and_type:
- - dump/raw/dev/wav.scp
- speech
- sound
- - dump/raw/dev/text
- text
- text
allow_variable_data_keys: false
max_cache_size: 0.0
max_cache_fd: 32
valid_max_cache_size: null
optim: adam
optim_conf:
lr: 0.0025
weight_decay: 1.0e-06
scheduler: warmuplr
scheduler_conf:
warmup_steps: 40000
token_list:
- <blank>
- <unk>
- ▁THE
- S
- ▁AND
- ▁OF
- ▁TO
- ▁A
- ▁IN
- ▁I
- ▁HE
- ▁THAT
- ▁WAS
- ED
- ▁IT
- ''''
- ▁HIS
- ING
- ▁YOU
- ▁WITH
- ▁FOR
- ▁HAD
- T
- ▁AS
- ▁HER
- ▁IS
- ▁BE
- ▁BUT
- ▁NOT
- ▁SHE
- D
- ▁AT
- ▁ON
- LY
- ▁HIM
- ▁THEY
- ▁ALL
- ▁HAVE
- ▁BY
- ▁SO
- ▁THIS
- ▁MY
- ▁WHICH
- ▁ME
- ▁SAID
- ▁FROM
- ▁ONE
- Y
- E
- ▁WERE
- ▁WE
- ▁NO
- N
- ▁THERE
- ▁OR
- ER
- ▁AN
- ▁WHEN
- ▁ARE
- ▁THEIR
- ▁WOULD
- ▁IF
- ▁WHAT
- ▁THEM
- ▁WHO
- ▁OUT
- M
- ▁DO
- ▁WILL
- ▁UP
- ▁BEEN
- P
- R
- ▁MAN
- ▁THEN
- ▁COULD
- ▁MORE
- C
- ▁INTO
- ▁NOW
- ▁VERY
- ▁YOUR
- ▁SOME
- ▁LITTLE
- ES
- ▁TIME
- RE
- ▁CAN
- ▁LIKE
- LL
- ▁ABOUT
- ▁HAS
- ▁THAN
- ▁DID
- ▁UPON
- ▁OVER
- IN
- ▁ANY
- ▁WELL
- ▁ONLY
- B
- ▁SEE
- ▁GOOD
- ▁OTHER
- ▁TWO
- L
- ▁KNOW
- ▁GO
- ▁DOWN
- ▁BEFORE
- A
- AL
- ▁OUR
- ▁OLD
- ▁SHOULD
- ▁MADE
- ▁AFTER
- ▁GREAT
- ▁DAY
- ▁MUST
- ▁COME
- ▁HOW
- ▁SUCH
- ▁CAME
- LE
- ▁WHERE
- ▁US
- ▁NEVER
- ▁THESE
- ▁MUCH
- ▁DE
- ▁MISTER
- ▁WAY
- G
- ▁S
- ▁MAY
- ATION
- ▁LONG
- OR
- ▁AM
- ▁FIRST
- ▁BACK
- ▁OWN
- ▁RE
- ▁AGAIN
- ▁SAY
- ▁MEN
- ▁WENT
- ▁HIMSELF
- ▁HERE
- NESS
- ▁THINK
- V
- IC
- ▁EVEN
- ▁THOUGHT
- ▁HAND
- ▁JUST
- ▁O
- ▁UN
- VE
- ION
- ▁ITS
- 'ON'
- ▁MAKE
- ▁MIGHT
- ▁TOO
- K
- ▁AWAY
- ▁LIFE
- TH
- ▁WITHOUT
- ST
- ▁THROUGH
- ▁MOST
- ▁TAKE
- ▁DON
- ▁EVERY
- F
- O
- ▁SHALL
- ▁THOSE
- ▁EYES
- AR
- ▁STILL
- ▁LAST
- ▁HOUSE
- ▁HEAD
- ABLE
- ▁NOTHING
- ▁NIGHT
- ITY
- ▁LET
- ▁MANY
- ▁OFF
- ▁BEING
- ▁FOUND
- ▁WHILE
- EN
- ▁SAW
- ▁GET
- ▁PEOPLE
- ▁FACE
- ▁YOUNG
- CH
- ▁UNDER
- ▁ONCE
- ▁TELL
- AN
- ▁THREE
- ▁PLACE
- ▁ROOM
- ▁YET
- ▁SAME
- IL
- US
- U
- ▁FATHER
- ▁RIGHT
- EL
- ▁THOUGH
- ▁ANOTHER
- LI
- RI
- ▁HEART
- IT
- ▁PUT
- ▁TOOK
- ▁GIVE
- ▁EVER
- ▁E
- ▁PART
- ▁WORK
- ERS
- ▁LOOK
- ▁NEW
- ▁KING
- ▁MISSUS
- ▁SIR
- ▁LOVE
- ▁MIND
- ▁LOOKED
- W
- RY
- ▁ASKED
- ▁LEFT
- ET
- ▁LIGHT
- CK
- ▁DOOR
- ▁MOMENT
- RO
- ▁WORLD
- ▁THINGS
- ▁HOME
- UL
- ▁THING
- LA
- ▁WHY
- ▁MOTHER
- ▁ALWAYS
- ▁FAR
- FUL
- ▁WATER
- CE
- IVE
- UR
- ▁HEARD
- ▁SOMETHING
- ▁SEEMED
- I
- LO
- ▁BECAUSE
- OL
- ▁END
- ▁TOLD
- ▁CON
- ▁YES
- ▁GOING
- ▁GOT
- RA
- IR
- ▁WOMAN
- ▁GOD
- EST
- TED
- ▁FIND
- ▁KNEW
- ▁SOON
- ▁EACH
- ▁SIDE
- H
- TON
- MENT
- ▁OH
- NE
- Z
- LING
- ▁AGAINST
- TER
- ▁NAME
- ▁MISS
- ▁QUITE
- ▁WANT
- ▁YEARS
- ▁FEW
- ▁BETTER
- ENT
- ▁HALF
- ▁DONE
- ▁ALSO
- ▁BEGAN
- ▁HAVING
- ▁ENOUGH
- IS
- ▁LADY
- ▁WHOLE
- LESS
- ▁BOTH
- ▁SEEN
- ▁SET
- ▁WHITE
- ▁COURSE
- IES
- ▁VOICE
- ▁CALLED
- ▁D
- ▁EX
- ATE
- ▁TURNED
- ▁GAVE
- ▁C
- ▁POOR
- MAN
- UT
- NA
- ▁DEAR
- ISH
- ▁GIRL
- ▁MORNING
- ▁BETWEEN
- LED
- ▁NOR
- IA
- ▁AMONG
- MA
- ▁
- ▁SMALL
- ▁REST
- ▁WHOM
- ▁FELT
- ▁HANDS
- ▁MYSELF
- ▁HIGH
- ▁M
- ▁HOWEVER
- ▁HERSELF
- ▁P
- CO
- ▁STOOD
- ID
- ▁KIND
- ▁HUNDRED
- AS
- ▁ROUND
- ▁ALMOST
- TY
- ▁SINCE
- ▁G
- AM
- ▁LA
- SE
- ▁BOY
- ▁MA
- ▁PERHAPS
- ▁WORDS
- ATED
- ▁HO
- X
- ▁MO
- ▁SAT
- ▁REPLIED
- ▁FOUR
- ▁ANYTHING
- ▁TILL
- ▁UNTIL
- ▁BLACK
- TION
- ▁CRIED
- RU
- TE
- ▁FACT
- ▁HELP
- ▁NEXT
- ▁LOOKING
- ▁DOES
- ▁FRIEND
- ▁LAY
- ANCE
- ▁POWER
- ▁BROUGHT
- VER
- ▁FIRE
- ▁KEEP
- PO
- FF
- ▁COUNTRY
- ▁SEA
- ▁WORD
- ▁CAR
- ▁DAYS
- ▁TOGETHER
- ▁IMP
- ▁REASON
- KE
- ▁INDEED
- TING
- ▁MATTER
- ▁FULL
- ▁TEN
- TIC
- ▁LAND
- ▁RATHER
- ▁AIR
- ▁HOPE
- ▁DA
- ▁OPEN
- ▁FEET
- ▁EN
- ▁FIVE
- ▁POINT
- ▁CO
- OM
- ▁LARGE
- ▁B
- ▁CL
- ME
- ▁GONE
- ▁CHILD
- INE
- GG
- ▁BEST
- ▁DIS
- UM
- ▁HARD
- ▁LORD
- OUS
- ▁WIFE
- ▁SURE
- ▁FORM
- DE
- ▁DEATH
- ANT
- ▁NATURE
- ▁BA
- ▁CARE
- ▁BELIEVE
- PP
- ▁NEAR
- ▁RO
- ▁RED
- ▁WAR
- IE
- ▁SPEAK
- ▁FEAR
- ▁CASE
- ▁TAKEN
- ▁ALONG
- ▁CANNOT
- ▁HEAR
- ▁THEMSELVES
- CI
- ▁PRESENT
- AD
- ▁MASTER
- ▁SON
- ▁THUS
- ▁LI
- ▁LESS
- ▁SUN
- ▁TRUE
- IM
- IOUS
- ▁THOUSAND
- ▁MONEY
- ▁W
- ▁BEHIND
- ▁CHILDREN
- ▁DOCTOR
- AC
- ▁TWENTY
- ▁WISH
- ▁SOUND
- ▁WHOSE
- ▁LEAVE
- ▁ANSWERED
- ▁THOU
- ▁DUR
- ▁HA
- ▁CERTAIN
- ▁PO
- ▁PASSED
- GE
- TO
- ▁ARM
- ▁LO
- ▁STATE
- ▁ALONE
- TA
- ▁SHOW
- ▁NEED
- ▁LIVE
- ND
- ▁DEAD
- ENCE
- ▁STRONG
- ▁PRE
- ▁TI
- ▁GROUND
- SH
- TI
- ▁SHORT
- IAN
- UN
- ▁PRO
- ▁HORSE
- MI
- ▁PRINCE
- ARD
- ▁FELL
- ▁ORDER
- ▁CALL
- AT
- ▁GIVEN
- ▁DARK
- ▁THEREFORE
- ▁CLOSE
- ▁BODY
- ▁OTHERS
- ▁SENT
- ▁SECOND
- ▁OFTEN
- ▁CA
- ▁MANNER
- MO
- NI
- ▁BRING
- ▁QUESTION
- ▁HOUR
- ▁BO
- AGE
- ▁ST
- ▁TURN
- ▁TABLE
- ▁GENERAL
- ▁EARTH
- ▁BED
- ▁REALLY
- ▁SIX
- 'NO'
- IST
- ▁BECOME
- ▁USE
- ▁READ
- ▁SE
- ▁VI
- ▁COMING
- ▁EVERYTHING
- ▁EM
- ▁ABOVE
- ▁EVENING
- ▁BEAUTIFUL
- ▁FEEL
- ▁RAN
- ▁LEAST
- ▁LAW
- ▁ALREADY
- ▁MEAN
- ▁ROSE
- WARD
- ▁ITSELF
- ▁SOUL
- ▁SUDDENLY
- ▁AROUND
- RED
- ▁ANSWER
- ICAL
- ▁RA
- ▁WIND
- ▁FINE
- ▁WON
- ▁WHETHER
- ▁KNOWN
- BER
- NG
- ▁TA
- ▁CAPTAIN
- ▁EYE
- ▁PERSON
- ▁WOMEN
- ▁SORT
- ▁ASK
- ▁BROTHER
- ▁USED
- ▁HELD
- ▁BIG
- ▁RETURNED
- ▁STRANGE
- ▁BU
- ▁PER
- ▁FREE
- ▁EITHER
- ▁WITHIN
- ▁DOUBT
- ▁YEAR
- ▁CLEAR
- ▁SIGHT
- ▁GRA
- ▁LOST
- ▁KEPT
- ▁F
- PE
- ▁BAR
- ▁TOWN
- ▁SLEEP
- ARY
- ▁HAIR
- ▁FRIENDS
- ▁DREAM
- ▁FELLOW
- PER
- ▁DEEP
- QUE
- ▁BECAME
- ▁REAL
- ▁PAST
- ▁MAKING
- RING
- ▁COMP
- ▁ACT
- ▁BAD
- HO
- STER
- ▁YE
- ▁MEANS
- ▁RUN
- MEN
- ▁DAUGHTER
- ▁SENSE
- ▁CITY
- ▁SOMETIMES
- ▁TOWARDS
- ▁ROAD
- ▁SP
- ▁LU
- ▁READY
- ▁FOOT
- ▁COLD
- ▁SA
- ▁LETTER
- ▁ELSE
- ▁MAR
- ▁STA
- BE
- ▁TRUTH
- ▁LE
- BO
- ▁BUSINESS
- CHE
- ▁JOHN
- ▁SUBJECT
- ▁COURT
- ▁IDEA
- ILY
- ▁RIVER
- ATING
- ▁FAMILY
- HE
- ▁DIDN
- ▁GLAD
- ▁SEVERAL
- IAL
- ▁UNDERSTAND
- ▁SC
- ▁POSSIBLE
- ▁DIFFERENT
- ▁RETURN
- ▁ARMS
- ▁LOW
- ▁HOLD
- ▁TALK
- ▁RU
- ▁WINDOW
- ▁INTEREST
- ▁SISTER
- SON
- ▁SH
- ▁BLOOD
- ▁SAYS
- ▁CAP
- ▁DI
- ▁HUMAN
- ▁CAUSE
- NCE
- ▁THANK
- ▁LATE
- GO
- ▁CUT
- ▁ACROSS
- ▁STORY
- NT
- ▁COUNT
- ▁ABLE
- DY
- LEY
- ▁NUMBER
- ▁STAND
- ▁CHURCH
- ▁THY
- ▁SUPPOSE
- LES
- BLE
- OP
- ▁EFFECT
- BY
- ▁K
- ▁NA
- ▁SPOKE
- ▁MET
- ▁GREEN
- ▁HUSBAND
- ▁RESPECT
- ▁PA
- ▁FOLLOWED
- ▁REMEMBER
- ▁LONGER
- ▁AGE
- ▁TAKING
- ▁LINE
- ▁SEEM
- ▁HAPPY
- LAND
- EM
- ▁STAY
- ▁PLAY
- ▁COMMON
- ▁GA
- ▁BOOK
- ▁TIMES
- ▁OBJECT
- ▁SEVEN
- QUI
- DO
- UND
- ▁FL
- ▁PRETTY
- ▁FAIR
- WAY
- ▁WOOD
- ▁REACHED
- ▁APPEARED
- ▁SWEET
- ▁FALL
- BA
- ▁PASS
- ▁SIGN
- ▁TREE
- IONS
- ▁GARDEN
- ▁ILL
- ▁ART
- ▁REMAIN
- ▁OPENED
- ▁BRIGHT
- ▁STREET
- ▁TROUBLE
- ▁PAIN
- ▁CONTINUED
- ▁SCHOOL
- OUR
- ▁CARRIED
- ▁SAYING
- HA
- ▁CHANGE
- ▁FOLLOW
- ▁GOLD
- ▁SW
- ▁FEELING
- ▁COMMAND
- ▁BEAR
- ▁CERTAINLY
- ▁BLUE
- ▁NE
- CA
- ▁WILD
- ▁ACCOUNT
- ▁OUGHT
- UD
- ▁T
- ▁BREATH
- ▁WANTED
- ▁RI
- ▁HEAVEN
- ▁PURPOSE
- ▁CHARACTER
- ▁RICH
- ▁PE
- ▁DRESS
- OS
- FA
- ▁TH
- ▁ENGLISH
- ▁CHANCE
- ▁SHIP
- ▁VIEW
- ▁TOWARD
- AK
- ▁JOY
- ▁JA
- ▁HAR
- ▁NEITHER
- ▁FORCE
- ▁UNCLE
- DER
- ▁PLAN
- ▁PRINCESS
- DI
- ▁CHIEF
- ▁HAT
- ▁LIVED
- ▁AB
- ▁VISIT
- ▁MOR
- TEN
- ▁WALL
- UC
- ▁MINE
- ▁PLEASURE
- ▁SMILE
- ▁FRONT
- ▁HU
- ▁DEAL
- OW
- ▁FURTHER
- GED
- ▁TRIED
- DA
- VA
- ▁NONE
- ▁ENTERED
- ▁QUEEN
- ▁PAY
- ▁EL
- ▁EXCEPT
- ▁SHA
- ▁FORWARD
- ▁EIGHT
- ▁ADDED
- ▁PUBLIC
- ▁EIGHTEEN
- ▁STAR
- ▁HAPPENED
- ▁LED
- ▁WALKED
- ▁ALTHOUGH
- ▁LATER
- ▁SPIRIT
- ▁WALK
- ▁BIT
- ▁MEET
- LIN
- ▁FI
- LT
- ▁MOUTH
- ▁WAIT
- ▁HOURS
- ▁LIVING
- ▁YOURSELF
- ▁FAST
- ▁CHA
- ▁HALL
- ▁BEYOND
- ▁BOAT
- ▁SECRET
- ENS
- ▁CHAIR
- RN
- ▁RECEIVED
- ▁CAT
- RESS
- ▁DESIRE
- ▁GENTLEMAN
- UGH
- ▁LAID
- EVER
- ▁OCCASION
- ▁WONDER
- ▁GU
- ▁PARTY
- DEN
- ▁FISH
- ▁SEND
- ▁NEARLY
- ▁TRY
- CON
- ▁SEEMS
- RS
- ▁BELL
- ▁BRA
- ▁SILENCE
- IG
- ▁GUARD
- ▁DIE
- ▁DOING
- ▁TU
- ▁COR
- ▁EARLY
- ▁BANK
- ▁FIGURE
- IF
- ▁ENGLAND
- ▁MARY
- ▁AFRAID
- LER
- ▁FO
- ▁WATCH
- ▁FA
- ▁VA
- ▁GRE
- ▁AUNT
- PED
- ▁SERVICE
- ▁JE
- ▁PEN
- ▁MINUTES
- ▁PAN
- ▁TREES
- NED
- ▁GLASS
- ▁TONE
- ▁PLEASE
- ▁FORTH
- ▁CROSS
- ▁EXCLAIMED
- ▁DREW
- ▁EAT
- ▁AH
- ▁GRAVE
- ▁CUR
- PA
- URE
- CENT
- ▁MILES
- ▁SOFT
- ▁AGO
- ▁POSITION
- ▁WARM
- ▁LENGTH
- ▁NECESSARY
- ▁THINKING
- ▁PICTURE
- ▁PI
- SHIP
- IBLE
- ▁HEAVY
- ▁ATTENTION
- ▁DOG
- ABLY
- ▁STANDING
- ▁NATURAL
- ▁APPEAR
- OV
- ▁CAUGHT
- VO
- ISM
- ▁SPRING
- ▁EXPERIENCE
- ▁PAT
- OT
- ▁STOPPED
- ▁REGARD
- ▁HARDLY
- ▁SELF
- ▁STRENGTH
- ▁GREW
- ▁KNIGHT
- ▁OPINION
- ▁WIDE
- ▁INSTEAD
- ▁SOUTH
- ▁TRANS
- ▁CORNER
- ▁LEARN
- ▁ISLAND
- ▁MI
- ▁THIRD
- ▁STE
- ▁STRAIGHT
- ▁TEA
- ▁BOUND
- ▁SEEING
- ▁JU
- ▁DINNER
- ▁BEAUTY
- ▁PEACE
- AH
- ▁REP
- ▁SILENT
- ▁CRE
- ALLY
- RIC
- ▁STEP
- ▁VER
- ▁JO
- GER
- ▁SITTING
- ▁THIRTY
- ▁SAVE
- ENED
- ▁GLANCE
- ▁REACH
- ▁ACTION
- ▁SAL
- ▁SAD
- ▁STONE
- ITIES
- ▁FRENCH
- ▁STRUCK
- ▁PAPER
- ▁WHATEVER
- ▁SUB
- ▁DISTANCE
- ▁WRONG
- ▁KNOWLEDGE
- ▁SAFE
- ▁SNOW
- ▁MUSIC
- ▁FIFTY
- RON
- ▁ATTEMPT
- ▁GOVERNMENT
- TU
- ▁CROWD
- ▁BESIDES
- ▁LOVED
- ▁BOX
- ▁DIRECTION
- ▁TRAIN
- ▁NORTH
- ▁THICK
- ▁GETTING
- AV
- ▁FLOOR
- ▁COMPANY
- ▁BLOW
- ▁PLAIN
- TRO
- ▁BESIDE
- ▁ROCK
- ▁IMMEDIATELY
- FI
- ▁SHADOW
- ▁SIT
- ORS
- ILE
- ▁DRINK
- ▁SPOT
- ▁DANGER
- ▁AL
- ▁SAINT
- ▁SLOWLY
- ▁PALACE
- IER
- ▁RESULT
- ▁PETER
- ▁FOREST
- ▁BELONG
- ▁SU
- ▁PAR
- RIS
- ▁TEARS
- ▁APPEARANCE
- ▁GATE
- BU
- ITION
- ▁QUICKLY
- ▁QUIET
- ▁LONDON
- ▁START
- ▁BROWN
- TRA
- KIN
- ▁CONSIDER
- ▁BATTLE
- ▁ANNE
- ▁PIECE
- ▁DIED
- ▁SUCCESS
- ▁LIPS
- ▁FILLED
- ▁FORGET
- ▁POST
- IFIED
- ▁MARGARET
- ▁FOOD
- HAM
- ▁PLEASANT
- ▁FE
- ▁EXPRESSION
- ▁POCKET
- ▁FRESH
- ▁WEAR
- TRI
- ▁BROKEN
- ▁LAUGHED
- GING
- ▁FOLLOWING
- WN
- IP
- ▁TOUCH
- ▁YOUTH
- ATIVE
- ▁LEG
- ▁WEEK
- ▁REMAINED
- ▁EASY
- NER
- RK
- ▁ENTER
- ▁FIGHT
- ▁PLACED
- ▁TRAVEL
- ▁SIMPLE
- ▁GIRLS
- ▁WAITING
- ▁STOP
- ▁WAVE
- AU
- ▁WISE
- ▁CAMP
- TURE
- UB
- ▁VE
- ▁OFFICE
- ▁GRAND
- ▁FIT
- ▁JUDGE
- UP
- MENTS
- ▁QUICK
- HI
- ▁FLO
- RIES
- VAL
- ▁COMFORT
- ▁PARTICULAR
- ▁STARTED
- ▁SUIT
- ▁NI
- ▁PALE
- ▁IMPOSSIBLE
- ▁HOT
- ▁CONVERSATION
- ▁SCENE
- ▁BOYS
- ▁WIN
- ▁BRE
- ▁SOCIETY
- ▁OUTSIDE
- ▁WRITE
- ▁EFFORT
- ▁TALKING
- ▁FORTUNE
- ▁NINE
- ▁WA
- ▁SINGLE
- ▁RULE
- ▁PORT
- ▁WINTER
- ▁CAST
- ▁CRA
- ▁HAPPEN
- ▁CRO
- ▁SHUT
- NING
- ▁GUN
- ▁NOBLE
- ▁BEGIN
- ▁PATH
- ▁SKY
- ▁WONDERFUL
- ▁SUDDEN
- ▁ARMY
- ▁CHE
- ▁WORTH
- ▁MOUNTAIN
- ▁MIN
- AG
- ▁FLU
- ▁GRACE
- ▁CHAPTER
- ▁BELOW
- ▁RING
- ▁TURNING
- ▁IRON
- ▁TOP
- ▁AFTERNOON
- ORY
- ▁EVIL
- ▁TRUST
- ▁BOW
- ▁TRI
- ▁SAIL
- ▁CONTENT
- ▁HORSES
- ITE
- ▁SILVER
- AP
- ▁LAD
- ▁RUNNING
- ▁HILL
- ▁BEGINNING
- ▁MAD
- ▁HABIT
- GRA
- ▁CLOTHES
- ▁MORROW
- ▁CRY
- ▁FASHION
- ▁PRESENCE
- ▁Z
- FE
- ▁ARRIVED
- ▁QUARTER
- ▁PERFECT
- ▁WO
- ▁TRA
- ▁USUAL
- ▁NECK
- ▁MARRIED
- ▁SEAT
- ▁WI
- ▁GAR
- ▁SAND
- ▁SHORE
- ▁GIVING
- NY
- ▁PROBABLY
- ▁MINUTE
- ▁EXPECT
- ▁DU
- ▁SHOT
- ▁INSTANT
- ▁DEGREE
- ▁COLOR
- ▁WEST
- RT
- ▁MARCH
- ▁BIRD
- ▁SHOWED
- ▁GREATER
- ▁SERIOUS
- ▁CARRY
- ▁COVERED
- ▁FORMER
- ▁LOUD
- ▁MOVED
- ▁MASS
- ▁SEEK
- ▁CHO
- GEN
- ▁ROMAN
- IB
- ▁MOON
- ▁BOARD
- ▁STREAM
- ▁EASILY
- ▁WISHED
- ▁SEARCH
- ▁COULDN
- ▁MONTHS
- ▁SICK
- LIE
- ▁DUTY
- ▁TWELVE
- ▁FAINT
- ▁STRANGER
- ▁SURPRISE
- ▁KILL
- ▁LEAVING
- ▁JOURNEY
- ▁SCARCELY
- ▁RAISED
- ▁SPEAKING
- ▁TERRIBLE
- ▁TOM
- ▁FIELD
- ▁GAME
- ▁QUA
- ▁PROMISE
- ▁LIE
- ▁CONDITION
- ▁TRO
- ▁PERSONAL
- ▁TALL
- ▁STICK
- ▁THREW
- ▁MARRY
- ▁VAN
- ▁BURN
- ▁ACCORDING
- ▁RISE
- ▁ATTACK
- ▁SWORD
- ▁GUESS
- ▁THOUGHTS
- ▁THIN
- ▁THROW
- ▁CALM
- SIDE
- ▁VILLAGE
- ▁DEN
- ▁ANXIOUS
- ▁MER
- GI
- ▁EXPECTED
- ▁BALL
- ▁ESPECIALLY
- ▁CHARGE
- ▁MEASURE
- ISE
- ▁NICE
- ▁TRYING
- ▁ALLOW
- ▁SHARP
- ▁BREAD
- ▁HONOUR
- ▁HONOR
- ▁ENTIRELY
- ▁BILL
- ▁BRI
- ▁WRITTEN
- ▁AR
- ▁BROKE
- ▁KILLED
- ▁MARK
- ▁VEN
- ▁LADIES
- ▁LEARNED
- ▁FLOWERS
- PLE
- ▁FORTY
- ▁OFFER
- ▁HAPPINESS
- ▁PRAY
- ▁CLASS
- ▁FER
- ▁PRINCIPLE
- GU
- ▁BOOKS
- ▁SHAPE
- ▁SUMMER
- ▁JACK
- ▁DRAW
- ▁GOLDEN
- ▁DECIDED
- ▁LEAD
- ▁UNLESS
- ▁HARM
- ▁LISTEN
- HER
- ▁SHOOK
- ▁INFLUENCE
- ▁PERFECTLY
- ▁MARRIAGE
- ▁BROAD
- ▁ESCAPE
- ▁STATES
- ▁MIDDLE
- ▁PLANT
- ▁MIL
- ▁MOVEMENT
- ▁NOISE
- ▁ENEMY
- ▁HISTORY
- ▁BREAK
- ROUS
- ▁UNDERSTOOD
- ▁LATTER
- FER
- ▁COMES
- ▁MERELY
- ▁SIMPLY
- WI
- ▁IMAGINE
- ▁LOWER
- ▁CONDUCT
- ▁BORN
- WA
- ▁YARD
- ▁KA
- ▁CLOSED
- ▁NOTE
- GA
- ▁STRA
- RAN
- ▁EXIST
- EV
- ▁SPEECH
- ▁BITTER
- JO
- ▁MAKES
- ▁GRASS
- ▁REPLY
- ▁CHANGED
- ▁MON
- ▁LYING
- ▁DANCE
- ▁FINALLY
- ▁AMERICAN
- ▁ENJOY
- ▁CONTAIN
- ▁MEANT
- USE
- ▁OBSERVED
- THER
- ▁LAUGH
- ▁AFTERWARDS
- ▁BEAT
- ▁RACE
- ▁EQUAL
- ▁RAIN
- PS
- ▁STEPS
- ▁BENEATH
- ▁TAIL
- ▁TASTE
- IO
- EY
- ▁CHAR
- ▁GE
- GN
- TIN
- ▁GROW
- ▁TE
- IANS
- ▁MOVE
- ▁REPEATED
- ▁DRIVE
- TUR
- ▁SI
- CLOCK
- ▁BRAVE
- ▁MADAME
- ▁LOT
- ▁CASTLE
- ▁HI
- AND
- ▁FUTURE
- ▁RELATION
- ▁SORRY
- ▁HEALTH
- ▁DICK
- ▁R
- ▁BUILDING
- ▁EDGE
- ▁BLESS
- ▁SPITE
- WE
- ▁MIS
- ▁PRISONER
- ▁ALLOWED
- ▁PH
- ▁CATCH
- MER
- ETH
- ▁COAT
- ▁COMPLETE
- ▁WOULDN
- ▁CREATURE
- ▁YELLOW
- ▁IMPORTANT
- ▁ADD
- ▁PASSING
- ▁DARKNESS
- ▁CARRIAGE
- ▁MILL
- ▁FIFTEEN
- NCY
- ▁HUNG
- ▁OB
- ▁PLEASED
- ▁SPREAD
- ▁CURIOUS
- ▁WORSE
- ▁CIRCUMSTANCES
- ▁GI
- LAR
- ▁CAL
- ▁HY
- ▁MERE
- ▁JANE
- ▁EAST
- BI
- ▁CUP
- ▁BLIND
- ▁PASSION
- ▁DISCOVERED
- ▁NOTICE
- ▁REPORT
- ▁SPACE
- ▁PRESENTLY
- ▁SORROW
- ▁PACK
- ▁DIN
- CY
- ▁DRY
- ▁ANCIENT
- ▁DRESSED
- ▁COVER
- ▁VO
- ▁EXISTENCE
- ▁EXACTLY
- ▁BEAST
- ▁PROPER
- ▁DROPPED
- ▁CLEAN
- ▁COLOUR
- ▁HOST
- ▁CHAMBER
- ▁FAITH
- LET
- ▁DETERMINED
- ▁PRIEST
- ▁STORM
- ▁SKIN
- ▁DARE
- ▁PERSONS
- ▁PICK
- ▁NARROW
- ▁SUPPORT
- ▁PRIVATE
- ▁SMILED
- ▁COUSIN
- ▁DRAWING
- ▁ATTEND
- ▁COOK
- ▁PREVENT
- ▁VARIOUS
- ▁BLA
- ▁FIXED
- ▁WEAK
- THE
- ▁HOLE
- ▁BOTTOM
- ▁NOBODY
- ADE
- ▁LEGS
- ITCH
- ▁INDIVIDUAL
- ▁EARS
- LIKE
- ▁ADVANTAGE
- ▁FRANCE
- ▁BON
- ▁WINE
- ▁LIVES
- OD
- ▁WALLS
- ▁TIRED
- ▁SHOP
- ▁ANIMAL
- ▁CRU
- ▁WROTE
- ▁ROYAL
- ▁CONSIDERED
- ▁MORAL
- ▁COMPANION
- ▁LOSE
- ▁ISN
- ▁BAG
- ▁LAKE
- ▁INTER
- ▁COM
- ▁LETTERS
- ▁LUCK
- ▁EAR
- ▁GERMAN
- ▁PET
- ▁SAKE
- ▁DROP
- ▁PAID
- ▁BREAKFAST
- ▁LABOR
- ▁DESERT
- ▁DECLARED
- ▁HUM
- ▁STUDY
- ▁INSTANCE
- ONE
- ▁SOMEWHAT
- ▁CLOTH
- ▁SPECIAL
- ▁COLONEL
- ▁SONG
- ▁MAIN
- ▁VALUE
- ▁PROUD
- ▁EXPRESS
- ▁NATION
- ▁HANDSOME
- ▁CONFESS
- ▁PU
- ▁PASSAGE
- ▁PERIOD
- ▁CUSTOM
- ▁HURT
- ▁SHOULDER
- ▁CHRIST
- ZA
- ▁RECEIVE
- ▁DIFFICULT
- ▁DEPEND
- ▁MEETING
- ▁CHI
- ▁GEN
- LIGHT
- ▁BELIEVED
- ▁SOCIAL
- ▁DIFFICULTY
- ▁GREATEST
- ▁DRAWN
- ▁GRANT
- ▁BIRDS
- ▁ANGRY
- ▁HEAT
- UFF
- ▁DUE
- ▁PLACES
- ▁SIN
- ▁COURAGE
- ▁EVIDENTLY
- ▁GENTLE
- ▁CRUEL
- ▁GEORGE
- ▁GRI
- ▁SERVANT
- ▁U
- ▁PURE
- OOK
- ▁KNOWS
- ▁KNOWING
- LF
- ▁WRITING
- ▁REMEMBERED
- ▁CU
- ▁HOLDING
- ▁TENDER
- ▁QUI
- ▁BURST
- ▁SURELY
- IGN
- ▁VALLEY
- ▁FU
- ▁BUTTER
- ▁SPOKEN
- ▁STORE
- ▁DISC
- ▁CHRISTIAN
- ▁PARIS
- ▁HENRY
- ▁FINISHED
- ▁PROVE
- ▁FOOL
- ▁SOLDIERS
- ▁LANGUAGE
- ▁INSIDE
- ▁BAN
- ▁FALLEN
- ROW
- ▁MAL
- ▁BABY
- ▁SITUATION
- ▁WATCHED
- ANS
- ▁RUIN
- ▁GENTLEMEN
- ▁FRO
- ▁FANCY
- ▁ACCEPT
- ▁SEASON
- ▁OURSELVES
- ▁SAN
- ▁SPEED
- IZED
- ▁COOL
- ▁SERVE
- ▁VESSEL
- ▁WILLIAM
- ▁OBLIGED
- ▁GROUP
- FORM
- ▁GOES
- UOUS
- ▁LEAVES
- ▁PECULIAR
- ▁NEWS
- ▁VAIN
- ▁EVERYBODY
- ▁PIN
- UG
- ▁FORGOTTEN
- ▁FRA
- GAN
- ▁CAREFULLY
- ▁FLASH
- UCH
- ▁FUR
- ▁MURDER
- ▁DELIGHT
- ▁WAITED
- ▁RENDER
- ▁PROPERTY
- ▁NOTICED
- ▁ROLL
- ▁KNOCK
- ▁EARNEST
- KI
- ▁HONEST
- ▁PROMISED
- ▁BAL
- AW
- ▁WALKING
- ANG
- ▁SQUARE
- ▁QUIETLY
- ▁CLOUD
- WOOD
- ▁FORMED
- ▁HIGHER
- ▁BUILT
- ▁FATE
- ▁TEACH
- MY
- ▁FALSE
- ▁YORK
- ▁DUST
- ▁CLIMB
- ▁FOND
- ▁GROWN
- ▁DESCEND
- ▁RAG
- ▁FRUIT
- ▁GENERALLY
- ▁OFFERED
- ▁ER
- ▁NURSE
- POSE
- ▁SPENT
- ▁JOIN
- ▁STATION
- ▁MEANING
- ▁SMOKE
- HOOD
- ▁ROUGH
- JU
- ▁LIKELY
- ▁SURFACE
- ▁KE
- ▁MONTH
- ▁POSSESSION
- ▁TONGUE
- ▁DUKE
- ▁NOSE
- ▁LAUGHING
- ▁WEATHER
- ▁WHISPERED
- ▁SYSTEM
- ▁LAWS
- DDLE
- ▁TOUCHED
- ▁TRADE
- LD
- ▁SURPRISED
- RIN
- ▁ARCH
- ▁WEALTH
- FOR
- ▁TEMPER
- ▁FRANK
- ▁GAL
- ▁BARE
- ▁OPPORTUNITY
- ▁CLAIM
- ▁ANIMALS
- ▁REV
- ▁COST
- ▁WASH
- ZE
- ▁CORN
- ▁OPPOSITE
- ▁POLICE
- ▁IDEAS
- LON
- ▁KEY
- ▁READING
- ▁COLLECT
- CHED
- ▁H
- ▁CROWN
- ▁TAR
- ▁SWIFT
- ▁SHOULDERS
- ▁ICE
- ▁GRAY
- ▁SHARE
- ▁PREPARED
- ▁GRO
- ▁UND
- ▁TER
- ▁EMPTY
- CING
- ▁SMILING
- ▁AVOID
- ▁DIFFERENCE
- ▁EXPLAIN
- ▁POUR
- ▁ATTRACT
- ▁OPENING
- ▁WHEEL
- ▁MATERIAL
- ▁BREAST
- ▁SUFFERING
- ▁DISTINCT
- ▁BOOT
- ▁ROW
- ▁FINGERS
- HAN
- ▁ALTOGETHER
- ▁FAT
- ▁PAPA
- ▁BRAIN
- ▁ASLEEP
- ▁GREY
- ▁SUM
- ▁GAS
- ▁WINDOWS
- ▁ALIVE
- ▁PROCEED
- ▁FLOWER
- ▁LEAP
- ▁PUR
- ▁PIECES
- ▁ALTER
- ▁MEMORY
- IENT
- ▁FILL
- ▁CLO
- ▁THROWN
- ▁KINGDOM
- ▁RODE
- IUS
- ▁MAID
- ▁DIM
- ▁BAND
- ▁VIRTUE
- ▁DISH
- ▁GUEST
- ▁LOSS
- ▁CAUSED
- ▁MOTION
- ▁POT
- ▁MILLION
- ▁FAULT
- ▁LOVELY
- ▁HERO
- PPING
- ▁UNITED
- ▁SPI
- SOME
- BRA
- ▁MOUNTAINS
- ▁NU
- ▁SATISFIED
- ▁DOLLARS
- ▁LOVER
- ▁CONCEAL
- ▁VAST
- ▁PULL
- ▁HATH
- ▁RUSH
- ▁J
- ▁DESPAIR
- EX
- ▁HEIGHT
- ▁CE
- ▁BENT
- ▁PITY
- ▁RISING
- ATH
- ▁PRIDE
- ▁HURRY
- KA
- ▁SETTLED
- ▁JUSTICE
- ▁LIFTED
- PEN
- ▁SOLDIER
- ▁FINDING
- ▁REMARK
- ▁REGULAR
- ▁STRUGGLE
- ▁MACHINE
- ▁SING
- ▁HURRIED
- ▁SUFFICIENT
- ▁REPRESENT
- ▁DOUBLE
- ▁ALARM
- ▁SUPPER
- ▁DREADFUL
- ▁FORE
- ATOR
- ▁STOCK
- ▁TIN
- ▁EXAMPLE
- ▁ROOF
- ▁FLOW
- ▁SUPPOSED
- ▁PRESERV
- ▁L
- ▁LISTENED
- OC
- ▁STO
- ▁SECURE
- ▁FRIGHTENED
- ▁DISTURB
- ▁EMOTION
- ▁SERVANTS
- ▁YO
- ▁BUY
- ▁FORCED
- ▁KITCHEN
- ▁TERROR
- ▁STAIRS
- ▁SIXTY
- KER
- ▁ORDINARY
- ▁DIRECTLY
- ▁HEADS
- ▁METHOD
- ▁FORGIVE
- ▁AWFUL
- ▁REFLECT
- ▁GREATLY
- ▁TALKED
- ▁RIDE
- STONE
- ▁FAVOUR
- ▁WELCOME
- ▁SEIZED
- OU
- ▁CONTROL
- ▁ORDERED
- ▁ANGEL
- ▁USUALLY
- ▁POET
- ▁BOLD
- LINE
- ▁ADVENTURE
- ▁WATCHING
- ▁FOLK
- ▁MISTRESS
- IZE
- ▁GROWING
- ▁CAVE
- ▁EVIDENCE
- ▁FINGER
- ▁SEVENTEEN
- ▁MOVING
- EOUS
- ▁DOESN
- ▁COW
- ▁TYPE
- ▁BOIL
- ▁TALE
- ▁DELIVER
- ▁FARM
- ▁MONSIEUR
- ▁GATHERED
- ▁FEELINGS
- ▁RATE
- ▁REMARKED
- ▁PUTTING
- ▁MAT
- ▁CONTRARY
- ▁CRIME
- ▁PLA
- ▁COL
- ▁NEARER
- TES
- ▁CIVIL
- ▁SHAME
- ▁LOOSE
- ▁DISCOVER
- ▁FLAT
- ▁TWICE
- ▁FAIL
- VIS
- ▁UNC
- EA
- ▁EUROPE
- ▁PATIENT
- ▁UNTO
- ▁SUFFER
- ▁PAIR
- ▁TREASURE
- OSE
- ▁EAGER
- ▁FLY
- ▁N
- ▁VAL
- ▁DAN
- ▁SALT
- ▁BORE
- BBE
- ▁ARTHUR
- ▁AFFAIRS
- ▁SLOW
- ▁CONSIST
- ▁DEVIL
- LAN
- ▁AFFECTION
- ▁ENGAGED
- ▁KISS
- ▁YA
- ▁OFFICER
- IFICATION
- ▁LAMP
- ▁PARTS
- HEN
- ▁MILK
- ▁PROCESS
- ▁GIFT
- ▁PULLED
- ▁HID
- ▁RAY
- ▁EXCELLENT
- ▁IMPRESSION
- ▁AUTHORITY
- ▁PROVED
- ▁TELLING
- TTE
- ▁TOWER
- ▁CONSEQUENCE
- ▁FAVOR
- ▁FLEW
- ▁CHARLES
- ISTS
- ▁ADDRESS
- ▁FAMILIAR
- ▁LIMIT
- ▁CONFIDENCE
- ▁RARE
- ▁WEEKS
- ▁WOODS
- ▁INTENTION
- ▁DIRECT
- ▁PERFORM
- ▁SOLEMN
- ▁DISTANT
- ▁IMAGE
- ▁PRESIDENT
- ▁FIRM
- ▁INDIAN
- ▁RANK
- ▁LIKED
- ▁AGREE
- ▁HOUSES
- ▁WIL
- ▁MATTERS
- ▁PRISON
- ▁MODE
- ▁MAJOR
- ▁WORKING
- ▁SLIP
- ▁WEIGHT
- ▁AWARE
- ▁BUSY
- ▁LOOKS
- ▁WOUND
- ▁THOR
- ▁BATH
- ▁EXERCISE
- ▁SIMILAR
- ▁WORE
- ▁AMOUNT
- ▁QUESTIONS
- ▁VIOLENT
- ▁EXCUSE
- ▁ASIDE
- ▁TUR
- ▁DULL
- OF
- ▁EMPEROR
- ▁NEVERTHELESS
- ▁SHOUT
- ▁EXPLAINED
- ▁SIZE
- ▁ACCOMPLISH
- FORD
- CAN
- ▁MISTAKE
- ▁INSTANTLY
- ▁SMOOTH
- ▁STRIKE
- ▁BOB
- ISED
- ▁HORROR
- ▁SCIENCE
- ▁PROTEST
- ▁MANAGE
- ▁OBEY
- ▁NECESSITY
- ▁SPLENDID
- ▁PRESS
- ▁INTERESTING
- ▁RELIGION
- ▁UNKNOWN
- ▁FIERCE
- ▁DISAPPEARED
- ▁HOLY
- ▁HATE
- ▁PLAYED
- ▁LIN
- ▁NATURALLY
- ▁DROVE
- ▁LOUIS
- TIES
- ▁BRAND
- INESS
- RIE
- ▁SHOOT
- ▁CONSENT
- ▁SEATED
- ▁LINES
- GUE
- ▁AGREED
- ▁CIRCLE
- ▁STIR
- ▁STREETS
- ▁TASK
- ▁RID
- ▁PRODUCED
- ▁ACCIDENT
- ▁WITNESS
- ▁LIBERTY
- ▁DETAIL
- ▁MINISTER
- ▁POWERFUL
- ▁SAVAGE
- ▁SIXTEEN
- ▁PRETEND
- ▁COAST
- ▁SQU
- ▁UTTER
- ▁NAMED
- ▁CLEVER
- ▁ADMIT
- ▁COUPLE
- ▁WICKED
- ▁MESSAGE
- ▁TEMPLE
- ▁STONES
- ▁YESTERDAY
- ▁HILLS
- DAY
- ▁SLIGHT
- ▁DIAMOND
- ▁POSSIBLY
- ▁AFFAIR
- ▁ORIGINAL
- ▁HEARING
- ▁WORTHY
- ▁SELL
- NEY
- ICK
- ▁COTTAGE
- ▁SACRIFICE
- ▁PROGRESS
- ▁SHOCK
- ▁DESIGN
- ▁SOUGHT
- ▁PIT
- ▁SUNDAY
- ▁OTHERWISE
- ▁CABIN
- ▁PRAYER
- ▁DWELL
- ▁GAIN
- ▁BRIDGE
- ▁PARTICULARLY
- ▁YIELD
- ▁TREAT
- RIGHT
- ▁OAK
- ▁ROPE
- WIN
- ▁ORDERS
- ▁SUSPECT
- ▁EDWARD
- AB
- ▁ELEVEN
- ▁TEETH
- ▁OCCURRED
- DDING
- ▁AMERICA
- ▁FALLING
- ▁LION
- ▁DEPART
- ▁KEEPING
- ▁DEMAND
- ▁PAUSED
- ▁CEASED
- INA
- ▁FUN
- ▁CHEER
- ▁PARDON
- ▁NATIVE
- LUS
- LOW
- ▁DOGS
- ▁REQUIRED
- ILITY
- ▁ELECT
- ▁ENTERTAIN
- ITUDE
- ▁HUGE
- ▁CARRYING
- ▁BLU
- ▁INSIST
- ▁SATISFACTION
- ▁HUNT
- ▁COUNTENANCE
- ▁UPPER
- ▁MAIDEN
- ▁FAILED
- ▁JAMES
- ▁FOREIGN
- ▁GATHER
- ▁TEST
- BOARD
- ▁TERMS
- ▁SILK
- ▁BEG
- ▁BROTHERS
- ▁PAGE
- ▁KNEES
- ▁SHOWN
- ▁PROFESSOR
- ▁MIGHTY
- ▁DEFI
- ▁CHARM
- ▁REQUIRE
- ▁LOG
- MORE
- ▁PROOF
- ▁POSSESSED
- ▁SOFTLY
- ▁UNFORTUNATE
- ▁PRICE
- ▁SEVERE
- ▁SINGING
- ▁STAGE
- ▁FREEDOM
- ▁SHOUTED
- ▁FARTHER
- ▁MAJESTY
- ▁PREVIOUS
- ▁GUIDE
- ▁MATCH
- ▁CHEST
- ▁INTENDED
- ▁BI
- ▁EXCITEMENT
- ▁OFFICERS
- ▁SUR
- ▁SHAKE
- ▁SENTIMENT
- ▁GENTLY
- ▁SUCCEEDED
- ▁MENTION
- ▁LOCK
- ▁ACQUAINTANCE
- ▁IMAGINATION
- ▁PHYSICAL
- ▁LEADING
- ▁SLAVE
- ▁CART
- ▁POINTED
- ▁STEAM
- ▁SHADE
- ▁PIPE
- ▁BASE
- ▁INVENT
- ▁ALAS
- ▁WORKED
- ▁REGRET
- ▁BUR
- ▁FAITHFUL
- ▁MENTIONED
- ▁RECORD
- ▁COMPLAIN
- ▁SUPERIOR
- ▁BAY
- ▁PAL
- EMENT
- UE
- ▁SEVENTY
- ▁HOTEL
- ▁SHEEP
- ▁MEAL
- ▁ADVICE
- ▁HIDDEN
- ▁DEMANDED
- ▁CONSCIOUS
- ▁BROW
- ▁POSSESS
- ▁FOURTH
- ▁EVENTS
- ▁FRI
- ▁PRAISE
- ▁ADVANCED
- ▁RESOLVED
- ▁STUFF
- ▁CHEERFUL
- ▁BIRTH
- ▁GRIEF
- ▁AFFORD
- ▁FAIRY
- ▁WAKE
- ▁SIDES
- ▁SUBSTANCE
- ▁ARTICLE
- ▁LEVEL
- ▁MIST
- ▁JOINED
- ▁PRACTICAL
- ▁CLEARLY
- ▁TRACE
- ▁AWAKE
- ▁OBSERVE
- ▁BASKET
- ▁LACK
- VILLE
- ▁SPIRITS
- ▁EXCITED
- ▁ABANDON
- ▁SHINING
- ▁FULLY
- ▁CALLING
- ▁CONSIDERABLE
- ▁SPRANG
- ▁MILE
- ▁DOZEN
- ▁PEA
- ▁DANGEROUS
- ▁WIT
- ▁JEW
- ▁POUNDS
- ▁FOX
- ▁INFORMATION
- ▁LIES
- ▁DECK
- NNY
- ▁PAUL
- ▁STARS
- ▁ANGER
- ▁SETTLE
- ▁WILLING
- ▁ADAM
- ▁FACES
- ▁SMITH
- ▁IMPORTANCE
- ▁STRAIN
- WAR
- ▁SAM
- ▁FEATHER
- ▁SERVED
- ▁AUTHOR
- ▁PERCEIVED
- ▁FLAME
- ▁DIVINE
- ▁TRAIL
- ▁ANYBODY
- ▁SIGH
- ▁DELICATE
- KY
- ▁FOLD
- ▁HAVEN
- ▁DESIRED
- ▁CURIOSITY
- ▁PRACTICE
- ▁CONSIDERATION
- ▁ABSOLUTELY
- ▁CITIZEN
- ▁BOTTLE
- ▁INTERESTED
- ▁MEAT
- ▁OCCUPIED
- ▁CHOOSE
- ▁THROAT
- ETTE
- ▁CANDLE
- ▁DAWN
- ▁PROTECT
- ▁SENTENCE
- IED
- ▁ROCKS
- ▁PORTION
- ▁APPARENTLY
- ▁PRESENTED
- ▁TIGHT
- ▁ACTUALLY
- ▁DYING
- ▁HAM
- ▁DAILY
- ▁SUFFERED
- ▁POLITICAL
- ▁BODIES
- ▁MODERN
- ▁COMPLETELY
- ▁SOONER
- TAN
- ▁PROP
- ▁ADVANCE
- ▁REFUSED
- ▁FARMER
- ▁POLITE
- ▁THUNDER
- ▁BRIEF
- ▁ELSIE
- ▁SAILOR
- ▁SUGGESTED
- ▁PLATE
- ▁AID
- ▁FLESH
- ▁WEEP
- ▁BUCK
- ▁ANTI
- ▁OCEAN
- ▁SPEND
- WELL
- ▁ODD
- ▁GOVERNOR
- ▁ENTRANCE
- ▁SUSPICION
- ▁STEPPED
- ▁RAPIDLY
- ▁CHECK
- ▁HIDE
- ▁FLIGHT
- ▁CLUB
- ▁ENTIRE
- ▁INDIANS
- ASH
- ▁CAPITAL
- ▁MAMMA
- HAR
- ▁CORRECT
- ▁CRACK
- ▁SENSATION
- ▁WORST
- ▁PACE
- ▁MIDST
- ▁AUGUST
- ▁PROPORTION
- ▁INNOCENT
- LINESS
- ▁REGARDED
- ▁DRIVEN
- ORD
- ▁HASTE
- ▁EDUCATION
- ▁EMPLOY
- ▁TRULY
- ▁INSTRUMENT
- ▁MAG
- ▁FRAME
- ▁FOOLISH
- ▁TAUGHT
- ▁HANG
- ▁ARGUMENT
- ▁NINETEEN
- ▁ELDER
- ▁NAY
- ▁NEEDED
- ▁NEIGHBOR
- ▁INSTRUCT
- ▁PAPERS
- ▁REWARD
- ▁EQUALLY
- ▁FIELDS
- ▁DIG
- HIN
- ▁CONDITIONS
- JA
- ▁SPAR
- ▁REQUEST
- ▁WORN
- ▁REMARKABLE
- ▁LOAD
- ▁WORSHIP
- ▁PARK
- ▁KI
- ▁INTERRUPTED
- ▁SKILL
- ▁TERM
- LAC
- ▁CRITIC
- ▁DISTRESS
- ▁BELIEF
- ▁STERN
- IGHT
- ▁TRACK
- ▁HUNTING
- ▁JEWEL
- ▁GRADUALLY
- ▁GLOW
- ▁RUSHED
- ▁MENTAL
- ▁VISITOR
- ▁PICKED
- ▁BEHOLD
- ▁EXPRESSED
- ▁RUB
- ▁SKI
- ARTAGNAN
- ▁MOREOVER
- ▁OPERATION
- ▁CAREFUL
- ▁KEEN
- ▁ASSERT
- ▁WANDER
- ▁ENEMIES
- ▁MYSTERIOUS
- ▁DEPTH
- ▁PREFER
- ▁CROSSED
- ▁CHARMING
- ▁DREAD
- ▁FLOUR
- ▁ROBIN
- ▁TRE
- ▁RELIEF
- ▁INQUIRED
- ▁APPLE
- ▁HENCE
- ▁WINGS
- ▁CHOICE
- ▁JUD
- OO
- ▁SPECIES
- ▁DELIGHTED
- IUM
- ▁RAPID
- ▁APPEAL
- ▁FAMOUS
- ▁USEFUL
- ▁HELEN
- ▁NEWSPAPER
- ▁PLENTY
- ▁BEARING
- ▁NERVOUS
- ▁PARA
- ▁URGE
- ▁ROAR
- ▁WOUNDED
- ▁CHAIN
- ▁PRODUCE
- ▁REFLECTION
- ▁MERCHANT
- ▁QUARREL
- ▁GLORY
- ▁BEGUN
- ▁BARON
- CUS
- ▁QUEER
- ▁MIX
- ▁GAZE
- ▁WHISPER
- ▁BURIED
- ▁DIV
- ▁CARD
- ▁FREQUENTLY
- ▁TIP
- ▁KNEE
- ▁REGION
- ▁ROOT
- ▁LEST
- ▁JEALOUS
- CTOR
- ▁SAVED
- ▁ASKING
- ▁TRIP
- QUA
- ▁UNION
- HY
- ▁COMPANIONS
- ▁SHIPS
- ▁HALE
- ▁APPROACHED
- ▁HARRY
- ▁DRUNK
- ▁ARRIVAL
- ▁SLEPT
- ▁FURNISH
- HEAD
- ▁PIG
- ▁ABSENCE
- ▁PHIL
- ▁HEAP
- ▁SHOES
- ▁CONSCIOUSNESS
- ▁KINDLY
- ▁EVIDENT
- ▁SCAR
- ▁DETERMIN
- ▁GRASP
- ▁STEAL
- ▁OWE
- ▁KNIFE
- ▁PRECIOUS
- ▁ELEMENT
- ▁PROCEEDED
- ▁FEVER
- ▁LEADER
- ▁RISK
- ▁EASE
- ▁GRIM
- ▁MOUNT
- ▁MEANWHILE
- ▁CENTURY
- OON
- ▁JUDGMENT
- ▁AROSE
- ▁VISION
- ▁SPARE
- ▁EXTREME
- ▁CONSTANT
- ▁OBSERVATION
- ▁THRUST
- ▁DELAY
- ▁CENT
- ▁INCLUD
- ▁LIFT
- ▁ADMIRE
- ▁ISSUE
- ▁FRIENDSHIP
- ▁LESSON
- ▁PRINCIPAL
- ▁MOURN
- ▁ACCEPTED
- ▁BURNING
- ▁CAPABLE
- ▁EXTRAORDINARY
- ▁SANG
- ▁REMOVED
- ▁HOPED
- ▁HORN
- ▁ALICE
- ▁MUD
- ▁APARTMENT
- ▁FIGHTING
- ▁BLAME
- ▁TREMBLING
- ▁SOMEBODY
- ▁ANYONE
- ▁BRIDE
- ▁READER
- ▁ROB
- ▁EVERYWHERE
- ▁LABOUR
- ▁RECALL
- ▁BULL
- ▁HIT
- ▁COUNCIL
- ▁POPULAR
- ▁CHAP
- ▁TRIAL
- ▁DUN
- ▁WISHES
- ▁BRILLIANT
- ▁ASSURED
- ▁FORGOT
- ▁CONTINUE
- ▁ACKNOWLEDG
- ▁RETREAT
- ▁INCREASED
- ▁CONTEMPT
- ▁GRANDFATHER
- ▁SYMPATHY
- ▁GHOST
- ▁STRETCHED
- ▁CREATURES
- ▁CAB
- ▁HIND
- ▁PLAYING
- ▁MISERABLE
- ▁MEMBERS
- ▁KINDNESS
- ▁HIGHEST
- ▁PRIM
- ▁KISSED
- ▁DESERVE
- ▁HUT
- ▁BEGGED
- ▁EIGHTY
- ▁CLOSELY
- ▁WONDERED
- ▁MILITARY
- ▁REMIND
- ▁ACCORDINGLY
- ▁LARGER
- ▁MAINTAIN
- ▁ENGINE
- ▁MOTIVE
- ▁DESTROY
- ▁STRIP
- ▁HANS
- ▁AHEAD
- ▁INFINITE
- ▁PROMPT
- ▁INFORMED
- TTLE
- ▁PEER
- ▁PRESSED
- ▁TRAP
- ▁SOMEWHERE
- ▁BOUGHT
- ▁VISIBLE
- ▁ASHAMED
- ▁TEAR
- ▁NEIGHBOUR
- ▁CONSTITUTION
- ▁INTELLIGENCE
- ▁PROFESSION
- ▁HUNGRY
- RIDGE
- ▁SMELL
- ▁STORIES
- ▁LISTENING
- ▁APPROACH
- ▁STRING
- ▁EXPLANATION
- ▁IMMENSE
- ▁RELIGIOUS
- ▁THROUGHOUT
- ▁HOLLOW
- ▁AWAIT
- ▁FLYING
- ▁SCREAM
- ▁ACTIVE
- ▁RUM
- ▁PRODUCT
- ▁UNHAPPY
- ▁VAGUE
- ARIES
- ▁ELIZABETH
- ▁STUPID
- ▁DIGNITY
- ▁ISABEL
- GAR
- ▁BRO
- ▁PITCH
- ▁COMRADE
- ▁STIFF
- ▁RECKON
- ▁SOLD
- ▁SPARK
- ▁STRO
- ▁CRYING
- ▁MAGIC
- ▁REPEAT
- PORT
- ▁MARKED
- ▁COMFORTABLE
- ▁PROJECT
- ▁BECOMING
- ▁PARENTS
- ▁SHELTER
- ▁STOLE
- ▁HINT
- ▁NEST
- ▁TRICK
- ▁THOROUGHLY
- ▁HOSPITAL
- ▁WEAPON
- ▁ROME
- ▁STYLE
- ▁ADMITTED
- ▁SAFETY
- FIELD
- ▁UNDERSTANDING
- ▁TREMBLE
- ▁PRINT
- ▁SLAVES
- ▁WEARY
- ▁ARTIST
- ▁CREDIT
- BURG
- ▁CONCLUSION
- ▁SELDOM
- ▁UNUSUAL
- ▁CLOUDS
- ▁UNABLE
- ▁GAY
- ▁HANGING
- ▁SCR
- ▁BOWED
- ▁DAVID
- ▁VOL
- ▁PUSHED
- ▁ESCAPED
- MOND
- ▁WARN
- ▁BETRAY
- ▁EGGS
- ▁PLAINLY
- ▁EXHIBIT
- ▁DISPLAY
- ▁MEMBER
- ▁GRIN
- ▁PROSPECT
- ▁BRUSH
- ▁BID
- ▁SUCCESSFUL
- ▁EXTENT
- ▁PERSUADE
- ▁MID
- ▁MOOD
- ▁ARRANGED
- ▁UNIVERSAL
- ▁JIM
- ▁SIGNAL
- ▁WHILST
- ▁PHILIP
- ▁WOLF
- RATE
- ▁EAGERLY
- ▁BILLY
- ▁RETURNING
- ▁CONSCIENCE
- ▁FORTUNATE
- ▁FEMALE
- ▁GLEAM
- ▁HASTILY
- ▁PROVIDED
- ▁OBTAIN
- ▁INSTINCT
- ▁CONCERNED
- ▁CONCERNING
- ▁SOMEHOW
- ▁PINK
- ▁RAGE
- ▁ACCUSTOMED
- ▁UNCONSCIOUS
- ▁ADVISE
- ▁BRANCHES
- ▁TINY
- ▁REFUSE
- ▁BISHOP
- ▁SUPPLY
- ▁PEASANT
- ▁LAWYER
- ▁WASTE
- ▁CONNECTION
- ▁DEVELOP
- ▁CORRESPOND
- ▁PLUM
- ▁NODDED
- ▁SLIPPED
- ▁EU
- ▁CONSTANTLY
- CUM
- MMED
- ▁FAIRLY
- HOUSE
- ▁KIT
- ▁RANG
- ▁FEATURES
- ▁PAUSE
- ▁PAINFUL
- ▁JOE
- ▁WHENCE
- ▁LAUGHTER
- ▁COACH
- ▁CHRISTMAS
- ▁EATING
- ▁WHOLLY
- ▁APART
- ▁SUPER
- ▁REVOLUTION
- ▁LONELY
- ▁CHEEKS
- ▁THRONE
- ▁CREW
- ▁ATTAIN
- ▁ESTABLISHED
- TIME
- ▁DASH
- ▁FRIENDLY
- ▁OPERA
- ▁EARL
- ▁EXHAUST
- ▁CLIFF
- ▁REVEAL
- ▁ADOPT
- ▁CENTRE
- ▁MERRY
- ▁SYLVIA
- ▁IDEAL
- ▁MISFORTUNE
- ▁FEAST
- ▁ARAB
- ▁NUT
- ▁FETCH
- ▁FOUGHT
- ▁PILE
- ▁SETTING
- ▁SOURCE
- ▁PERSIST
- ▁MERCY
- ▁BARK
- ▁LUC
- ▁DEEPLY
- ▁COMPARE
- ▁ATTITUDE
- ▁ENDURE
- ▁DELIGHTFUL
- ▁BEARD
- ▁PATIENCE
- ▁LOCAL
- ▁UTTERED
- ▁VICTORY
- ▁TREATED
- ▁SEPARATE
- ▁WAG
- ▁DRAGG
- ▁TITLE
- ▁TROOPS
- ▁TRIUMPH
- ▁REAR
- ▁GAINED
- ▁SINK
- ▁DEFEND
- ▁TIED
- ▁FLED
- ▁DARED
- ▁INCREASE
- ▁POND
- ▁CONQUER
- ▁FOREHEAD
- ▁FAN
- ▁ANXIETY
- ▁ENCOUNTER
- ▁SEX
- ▁HALT
- ▁SANK
- ▁CHEEK
- ▁HUMBLE
- ▁WRITER
- ▁EMPLOYED
- ▁DISTINGUISHED
- ▁RAISE
- ▁WHIP
- ▁GIANT
- ▁RANGE
- ▁OBTAINED
- ▁FLAG
- ▁MAC
- ▁JUMPED
- ▁DISCOVERY
- ▁NATIONAL
- ▁COMMISSION
- ▁POSITIVE
- ▁LOVING
- ▁EXACT
- ▁MURMURED
- ▁GAZED
- ▁REFER
- ▁COLLEGE
- ▁ENCOURAGE
- ▁NOVEL
- ▁CLOCK
- ▁MORTAL
- ▁ROLLED
- ▁RAT
- IZING
- ▁GUILTY
- ▁VICTOR
- WORTH
- ▁PRA
- ▁APPROACHING
- ▁RELATIVE
- ▁ESTATE
- ▁UGLY
- ▁METAL
- ▁ROBERT
- ▁TENT
- ▁ADMIRATION
- ▁FOURTEEN
- ▁BARBAR
- ▁WITCH
- ELLA
- ▁CAKE
- ▁SHONE
- ▁MANAGED
- ▁VOLUME
- ▁GREEK
- ▁DANCING
- ▁WRETCHED
- ▁CONDEMN
- ▁MAGNIFICENT
- ▁CONSULT
- J
- ▁ORGAN
- ▁FLEET
- ▁ARRANGEMENT
- ▁INCIDENT
- ▁MISERY
- ▁ARROW
- ▁STROKE
- ▁ASSIST
- ▁BUILD
- ▁SUCCEED
- ▁DESPERATE
- ▁WIDOW
- UDE
- ▁MARKET
- ▁WISDOM
- ▁PRECISE
- ▁CURRENT
- ▁SPOIL
- ▁BADE
- ▁WOODEN
- ▁RESIST
- ▁OBVIOUS
- ▁SENSIBLE
- FALL
- ▁ADDRESSED
- ▁GIL
- ▁COUNSEL
- ▁PURCHASE
- ▁SELECT
- ▁USELESS
- ▁STARED
- ▁ARREST
- ▁POISON
- ▁FIN
- ▁SWALLOW
- ▁BLOCK
- ▁SLID
- ▁NINETY
- ▁SPORT
- ▁PROVIDE
- ▁ANNA
- ▁LAMB
- ▁INTERVAL
- ▁JUMP
- ▁DESCRIBED
- ▁STRIKING
- ▁PROVISION
- ▁PROPOSED
- ▁MELANCHOLY
- ▁WARRIOR
- ▁SUGGEST
- ▁DEPARTURE
- ▁BURDEN
- ▁LIMB
- ▁TROUBLED
- ▁MEADOW
- ▁SACRED
- ▁SOLID
- ▁TRU
- ▁LUCY
- ▁RECOVER
- ▁ENERGY
- ▁POWDER
- ▁RESUMED
- ▁INTENSE
- ▁BRITISH
- ▁STRAW
- ▁AGREEABLE
- ▁EVERYONE
- ▁CONCERN
- ▁VOYAGE
- ▁SOUTHERN
- ▁BOSOM
- ▁UTTERLY
- ▁FEED
- ▁ESSENTIAL
- ▁CONFINE
- ▁HOUSEHOLD
- ▁EXTREMELY
- ▁WONDERING
- ▁LIST
- ▁PINE
- PHA
- ▁EXPERIMENT
- ▁JOSEPH
- ▁MYSTERY
- ▁RESTORE
- ▁BLUSH
- FOLD
- ▁CHOSEN
- ▁INTELLECT
- ▁CURTAIN
- OLOGY
- ▁MOUNTED
- ▁LAP
- ▁EPI
- ▁PUNISH
- ▁WEDDING
- ▁RECOGNIZED
- ▁DRIFT
- ▁PREPARATION
- ▁RESOLUTION
- ▁OPPRESS
- ▁FIX
- ▁VICTIM
- OGRAPH
- ▁SUMMON
- ▁JULIA
- ▁FLOOD
- ▁WAL
- ULATION
- ▁SLIGHTLY
- ▁LODGE
- ▁WIRE
- ▁CONFUSION
- ▁UNEXPECTED
- ▁CONCEIVE
- ▁PRIZE
- ▁JESUS
- ▁ADDITION
- ▁RUDE
- ▁FATAL
- ▁CARELESS
- ▁PATCH
- ▁KO
- ▁CATHERINE
- ▁PARLIAMENT
- ▁PROFOUND
- ▁ALOUD
- ▁RELIEVE
- ▁PUSH
- ABILITY
- ▁ACCOMPANIED
- ▁SOVEREIGN
- ▁SINGULAR
- ▁ECHO
- ▁COMPOSED
- ▁SHAKING
- ATORY
- ▁ASSISTANCE
- ▁TEACHER
- ▁HORRIBLE
- ▁STRICT
- ▁VERSE
- ▁PUNISHMENT
- ▁GOWN
- ▁MISTAKEN
- ▁VARI
- ▁SWEPT
- ▁GESTURE
- ▁BUSH
- ▁STEEL
- ▁AFFECTED
- ▁DIRECTED
- ▁SURROUNDED
- ▁ABSURD
- ▁SUGAR
- ▁SCRAP
- ▁IMMEDIATE
- ▁SADDLE
- ▁TY
- ▁ARISE
- ▁SIGHED
- ▁EXCHANGE
- ▁IMPATIENT
- ▁SNAP
- ▁EMBRACE
- ▁DISEASE
- ▁PROFIT
- ▁RIDING
- ▁RECOVERED
- ▁GOVERN
- ▁STRETCH
- ▁CONVINCED
- ▁LEANING
- ▁DOMESTIC
- ▁COMPLEX
- ▁MANIFEST
- ▁INDULGE
- ▁GENIUS
- ▁AGENT
- ▁VEIL
- ▁DESCRIPTION
- ▁INCLINED
- ▁DECEIVE
- ▁DARLING
- ▁REIGN
- HU
- ▁ENORMOUS
- ▁RESTRAIN
- ▁DUTIES
- BURY
- TTERED
- ▁POLE
- ▁ENABLE
- ▁EXCEPTION
- ▁INTIMATE
- ▁COUNTESS
- ▁TRIBE
- ▁HANDKERCHIEF
- ▁MIDNIGHT
- ▁PROBLEM
- ▁TRAMP
- ▁OIL
- CAST
- ▁CRUSH
- ▁DISCUSS
- ▁RAM
- ▁TROT
- ▁UNRE
- ▁WHIRL
- ▁LOCKED
- ▁HORIZON
- ▁OFFICIAL
- ▁SCHEME
- ▁DROWN
- ▁PIERRE
- ▁PERMITTED
- ▁CONNECTED
- ▁ASSURE
- ▁COCK
- ▁UTMOST
- ▁DEVOTED
- ▁RELI
- ▁SUFFICIENTLY
- ▁INTELLECTUAL
- ▁CARPET
- ▁OBJECTION
- ▁AFTERWARD
- ▁REALITY
- ▁NEGRO
- ▁RETAIN
- ▁ASCEND
- ▁CEASE
- ▁KATE
- ▁MARVEL
- KO
- ▁BOND
- MOST
- ▁COAL
- GATE
- ▁IGNORANT
- ▁BREAKING
- ▁TWIN
- ▁ASTONISHMENT
- ▁COFFEE
- ▁JAR
- ▁CITIES
- ▁ORIGIN
- ▁EXECUT
- ▁FINAL
- ▁INHABITANTS
- ▁STABLE
- ▁CHIN
- ▁PARTIES
- ▁PLUNGE
- ▁GENEROUS
- ▁DESCRIBE
- ▁ANNOUNCED
- ▁MERIT
- ▁REVERE
- ▁ERE
- ACIOUS
- ZI
- ▁DISAPPOINT
- ▁SUGGESTION
- ▁DOUBTLESS
- ▁TRUNK
- ▁STAMP
- ▁JOB
- ▁APPOINTED
- ▁DIVIDED
- ▁ACQUAINTED
- CHI
- ▁ABSOLUTE
- ▁FEARFUL
- ▁PRIVILEGE
- ▁CRAFT
- ▁STEEP
- ▁HUNTER
- ▁FORBID
- ▁MODEST
- ▁ENDEAVOUR
- ▁SWEEP
- ▁BEHELD
- ▁ABSORB
- ▁CONSTRUCT
- ▁EMPIRE
- ▁EXPEDITION
- ▁ERECT
- ▁OFFEND
- ▁INTEND
- ▁PERMIT
- ▁DESTROYED
- ▁CONTRACT
- ▁THIRST
- ▁WAGON
- ▁EVA
- ▁GLOOM
- ▁ATMOSPHERE
- ▁RESERVE
- ▁VOTE
- ▁GER
- ▁NONSENSE
- ▁PREVAIL
- ▁QUALITY
- ▁CLASP
- ▁CONCLUDED
- ▁RAP
- ▁KATY
- ▁ETERNAL
- ▁MUTTERED
- ▁NEGLECT
- ▁SQUIRE
- ▁CREEP
- LOCK
- ▁ELECTRIC
- ▁HAY
- ▁EXPENSE
- ▁SCORN
- ▁RETIRED
- ▁STOUT
- ▁MURMUR
- ▁SHARPLY
- ▁DISTRICT
- ▁LEAF
- ▁FAILURE
- WICK
- ▁JEAN
- ▁NUMEROUS
- ▁INFANT
- ▁REALIZED
- ▁TRAVELLER
- ▁HUNGER
- ▁JUNE
- ▁MUN
- ▁RECOMMEND
- ▁CREP
- ZZLE
- ▁RICHARD
- WORK
- ▁MONTE
- ▁PREACH
- ▁PALM
- AVI
- ▁ANYWHERE
- ▁DISPOSITION
- ▁MIRROR
- ▁VENTURE
- ▁POUND
- ▁CIGAR
- ▁INVITED
- ▁BENCH
- ▁PROTECTION
- ▁BENEFIT
- ▁THOMAS
- ▁CLERK
- ▁REPROACH
- ▁UNIFORM
- ▁GENERATION
- ▁SEAL
- ▁COMPASS
- ▁WARNING
- ▁EXTENDED
- ▁DIFFICULTIES
- ▁MAYBE
- ▁GROAN
- ▁AFFECT
- ▁COMB
- ▁EARN
- ▁WESTERN
- ▁IDLE
- ▁SCORE
- ▁TAP
- ▁ASTONISHED
- ▁INTRODUCED
- ▁LEISURE
- ▁LIEUTENANT
- ▁VIOLENCE
- ▁FIRMLY
- ▁MONSTER
- ▁UR
- ▁PROPERLY
- ▁TWIST
- ▁PIRATE
- ▁ROBBER
- ▁BATTER
- ▁WEPT
- ▁LEANED
- ▁FOG
- ▁ORNAMENT
- ▁ANDREW
- ▁BUSHES
- ▁REPUBLIC
- ▁CONFIDENT
- ▁LEAN
- ▁DART
- ▁STOOP
- ▁CURL
- ▁COUNTER
- ▁NORTHERN
- ▁PEARL
- ▁NEAREST
- ▁FRANCIS
- ▁WANDERING
- ▁FREQUENT
- ▁STARTLED
- ▁STATEMENT
- ▁OCCUR
- ▁BLOOM
- ▁NERVE
- ▁INSPECT
- ▁INDUCE
- ▁FLATTER
- ▁DATE
- ▁AMBITION
- ▁SLOPE
- ▁MALE
- ▁MADAM
- ▁MONK
- ▁RENT
- ▁CONFIRM
- ▁INVESTIGAT
- ▁RABBIT
- ▁REGIMENT
- ▁SUBMIT
- ▁SPELL
- ▁FURIOUS
- ▁RAIL
- ▁BESTOW
- ▁RALPH
- ▁SCATTERED
- ▁COMPELLED
- ▁THREAD
- ▁CHILL
- ▁DENY
- ▁PRONOUNC
- ▁MANKIND
- ▁CATTLE
- ▁EXECUTION
- ▁REBEL
- ▁SUPREME
- ▁VALUABLE
- ▁LIKEWISE
- ▁CONVEY
- ▁TIDE
- ▁GLOOMY
- ▁COIN
- ▁ACTUAL
- ▁TAX
- ▁PROVINCE
- ▁GRATEFUL
- ▁SPIRITUAL
- ▁VANISHED
- ▁DIANA
- ▁HAUNT
- ▁DRAGON
- ▁CRAWL
- ▁CHINA
- ▁GRATITUDE
- ▁NEAT
- ▁FINISH
- ▁INTENT
- ▁FRIGHT
- ▁EMBARRASS
- ▁THIRTEEN
- ▁RUTH
- ▁SLIGHTEST
- ▁DEVELOPMENT
- ▁INTERVIEW
- ▁SPECTACLE
- ▁BROOK
- VIE
- ▁WEAKNESS
- ▁AUDIENCE
- ▁CONSEQUENTLY
- ▁ABROAD
- ▁ASPECT
- ▁PAINTED
- ▁RELEASE
- ▁INSULT
- ▁SOOTH
- ▁DISAPPOINTMENT
- ▁EMERG
- ▁BRIG
- ▁ESTEEM
- ▁INVITATION
- ▁PASSENGER
- ▁PUBLISH
- ▁PIANO
- ▁IRISH
- ▁DESK
- ▁BEATEN
- ▁FIFTH
- ▁IMPULSE
- ▁SWEAR
- ▁EATEN
- ▁PURPLE
- ▁COMMITTED
- ▁COUNTRIES
- ▁PERCEIVE
- ISON
- ▁CELEBRAT
- ▁GRANDMOTHER
- ▁SHUDDER
- ▁SUNSHINE
- ▁SPANISH
- ▁HITHERTO
- ▁MARILLA
- ▁SNAKE
- ▁MOCK
- ▁INTERFERE
- ▁WALTER
- ▁AMID
- ▁MARBLE
- ▁MISSION
- TERIOR
- ▁DRIVING
- ▁FURNITURE
- ▁STEADY
- ▁CIRCUMSTANCE
- ▁INTERPRET
- ▁ENCHANT
- ▁ERROR
- ▁CONVICTION
- ▁HELPLESS
- ▁MEDICINE
- ▁QUALITIES
- ▁ITALIAN
- ▁HASTENED
- ▁OCCASIONALLY
- ▁PURSUED
- ▁HESITATED
- ▁INDEPENDENT
- ▁OLIVER
- ▁LINGER
- UX
- ▁EXAMINED
- ▁REPENT
- ▁PHYSICIAN
- ▁CHASE
- ▁BELOVED
- ▁ATTACHED
- ▁FLORENCE
- ▁HONEY
- ▁MOUSE
- ▁CRIES
- ▁BAKE
- ▁POEM
- ▁DESTRUCTION
- ▁FULFIL
- ▁MESSENGER
- ▁TRISTRAM
- ▁FANCIED
- ▁EXCESS
- ▁CURSE
- ▁CHU
- ▁QUANTITY
- ▁THORNTON
- ▁CREATED
- ▁CONTINUALLY
- ▁LIGHTNING
- ▁BORNE
- ▁TOTAL
- ▁DISPOSED
- ▁RIFLE
- ▁POLLY
- ▁GOAT
- ▁BACKWARD
- ▁VIRGINIA
- ▁KICK
- ▁PERIL
- ▁QUO
- ▁GLORIOUS
- ▁MULTITUDE
- ▁LEATHER
- ▁ABSENT
- ▁DEMON
- ▁DEBT
- ▁TORTURE
- ▁ACCORD
- ▁MATE
- ▁CATHOLIC
- ▁PILL
- ▁LIBRARY
- ▁PURSUIT
- ▁SHIRT
- ▁DEAREST
- ▁COLLAR
- ▁BEACH
- ▁ROBE
- ▁DECLARE
- ▁BRANCH
- ▁TEMPT
- ▁STEADILY
- ▁DISGUST
- ▁SILLY
- ▁ARRIVE
- ▁DRANK
- ▁LEVI
- ▁COMMUNICAT
- ▁RACHEL
- ▁WASHINGTON
- ▁RESIGN
- ▁MEANTIME
- ▁LACE
- ▁ENGAGEMENT
- ▁QUIVER
- ▁SEPARATED
- ▁DISCUSSION
- ▁VENTURED
- ▁SURROUNDING
- ▁POLISH
- ▁NAIL
- ▁SWELL
- ▁JOKE
- ▁LINCOLN
- ▁STUDENT
- ▁GLITTER
- ▁RUSSIAN
- ▁READILY
- ▁CHRIS
- ▁POVERTY
- ▁DISGRACE
- ▁CHEESE
- ▁HEAVILY
- ▁SCALE
- ▁STAFF
- ▁ENTREAT
- ▁FAREWELL
- ▁LUNCH
- ▁PEEP
- ▁MULE
- ▁SOMEONE
- ▁DISAPPEAR
- ▁DECISION
- ▁PISTOL
- ▁PUN
- ▁SPUR
- ▁ASSUMED
- ▁EXTEND
- ▁ENTHUSIASM
- ▁DEFINITE
- ▁UNDERTAKE
- ▁COMMITTEE
- ▁SIMON
- ▁FENCE
- ▁APPLIED
- ▁RELATED
- ▁VICE
- ▁UNPLEASANT
- ▁PROBABLE
- ▁PROCURE
- ▁FROWN
- ▁CLOAK
- ▁HUMANITY
- ▁FAMILIES
- ▁PHILOSOPHER
- ▁DWARF
- ▁OVERCOME
- ▁DEFEAT
- ▁FASTENED
- ▁MARSH
- ▁CLASSES
- ▁TOMB
- ▁GRACIOUS
- ▁REMOTE
- ▁CELL
- ▁SHRIEK
- ▁RESCUE
- ▁POOL
- ▁ORGANIZ
- ▁CHOSE
- ▁CUTTING
- ▁COWARD
- ▁BORDER
- ▁DIRTY
- ▁MONKEY
- ▁HOOK
- ▁CHUCK
- ▁EMILY
- ▁JEST
- ▁PLAC
- ▁WEIGH
- ▁ASSOCIATE
- ▁GLIMPSE
- ▁STUCK
- ▁BOLT
- ▁MURDERER
- ▁PONY
- ▁DISTINGUISH
- ▁INSTITUTION
- ▁CUNNING
- ▁COMPLIMENT
- ▁APPETITE
- ▁REPUTATION
- ▁FEEBLE
- ▁KIN
- ▁SERIES
- ▁GRACEFUL
- ▁PLATFORM
- ▁BREEZE
- ▁PHRASE
- ▁CLAY
- MONT
- ▁RATTL
- ▁OPPOSITION
- ▁LANE
- ▁BOAST
- ▁GROWTH
- ▁INCLINATION
- ▁BEHAVE
- ▁SUSAN
- ▁DISTINCTION
- ▁DISLIKE
- ▁NICHOLAS
- ▁SATISFY
- ▁DRAMA
- ▁ELBOW
- ▁GAZING
- ▁CONSUM
- ▁SPIN
- ▁OATH
- ▁CHANNEL
- ▁CHARACTERISTIC
- ▁SPEAR
- ▁SLAIN
- ▁SAUCE
- ▁FROG
- ▁CONCEPTION
- ▁TIMID
- ▁ZEAL
- ▁APPARENT
- SHIRE
- ▁CENTER
- ▁VARIETY
- ▁DUSK
- ▁APT
- ▁COLUMN
- ▁REVENGE
- ▁RIVAL
- ▁IMITAT
- ▁PASSIONATE
- ▁SELFISH
- ▁NORMAN
- ▁REPAIR
- ▁THRILL
- ▁TREATMENT
- ▁ROSA
- ▁MARTIN
- ▁INDIFFERENT
- ▁THITHER
- ▁GALLANT
- ▁PEPPER
- ▁RECOLLECT
- ▁VINE
- ▁SCARCE
- ▁SHIELD
- ▁MINGLED
- CLOSE
- ▁HARSH
- ▁BRICK
- ▁HUMOR
- ▁MISCHIEF
- ▁TREMENDOUS
- ▁FUNCTION
- ▁SMART
- ▁SULTAN
- ▁DISMISS
- ▁THREATENED
- ▁CHEAP
- ▁FLOCK
- ▁ENDEAVOR
- ▁WHISK
- ▁ITALY
- ▁WAIST
- ▁FLUTTER
- ▁SMOKING
- ▁MONARCH
- ▁AFRICA
- ▁ACCUSE
- ▁HERBERT
- ▁REFRESH
- ▁REJOICE
- ▁PILLOW
- ▁EXPECTATION
- ▁POETRY
- ▁HOPELESS
- ▁PERISH
- ▁PHILOSOPHY
- ▁WHISTLE
- ▁BERNARD
- ▁LAMENT
- ▁IMPROVE
- ▁SUP
- ▁PERPLEX
- ▁FOUNTAIN
- ▁LEAGUE
- ▁DESPISE
- ▁IGNORANCE
- ▁REFERENCE
- ▁DUCK
- ▁GROVE
- ▁PURSE
- ▁PARTNER
- ▁PROPHET
- ▁SHIVER
- ▁NEIGHBOURHOOD
- ▁REPRESENTATIVE
- SAIL
- ▁WIP
- ▁ACQUIRED
- ▁CHIMNEY
- ▁DOCTRINE
- ▁MAXIM
- ▁ANGLE
- ▁MAJORITY
- ▁AUTUMN
- ▁CONFUSED
- ▁CRISTO
- ▁ACHIEVE
- ▁DISGUISE
- ▁REDUCED
- ▁EARLIER
- ▁THEATRE
- ▁DECIDE
- MINATED
- OLOGICAL
- ▁OCCUPATION
- ▁VIGOROUS
- ▁CONTINENT
- ▁DECLINE
- ▁COMMUNITY
- ▁MOTIONLESS
- ▁HATRED
- ▁COMMUNICATION
- ▁BOWL
- ▁COMMENT
- ▁APPROVE
- ▁CEREMONY
- ▁CRIMINAL
- ▁SCIENTIFIC
- ▁DUCHESS
- ▁VIVID
- ▁SHIFT
- ▁AVAIL
- ▁DAMP
- ▁JOHNSON
- ▁SLENDER
- ▁CONTRAST
- ▁AMUSEMENT
- ▁PLOT
- ▁LYN
- ▁ASSOCIATION
- ▁SNATCH
- ▁UNCERTAIN
- ▁PRESSURE
- ▁PERCH
- ▁APPLY
- ▁PLANET
- ▁NOTWITHSTANDING
- ▁SWUNG
- ▁STIRRED
- ▁ATTENDANT
- ▁ENJOYMENT
- ▁WORRY
- ▁ALBERT
- ▁NAKED
- ▁TALENT
- ▁MARIAN
- ▁REFORM
- ▁DELIBERATE
- ▁INTELLIGENT
- ▁SENSITIVE
- ▁YONDER
- ▁PUPIL
- ▁FRIGHTFUL
- ▁DOUBTFUL
- ▁STANDARD
- ▁MAGISTRATE
- ▁SHEPHERD
- ▁STOMACH
- ▁DEPOSIT
- ▁RENEW
- ▁HEDGE
- ▁FRANCS
- ▁POSSIBILITY
- ▁RESEMBLE
- ▁FATIGUE
- ▁PORTRAIT
- ▁FAVORITE
- ▁CREAM
- ▁BURG
- ▁SECRETARY
- ▁DIVERS
- ▁ACTIVITY
- ▁SPECULAT
- ▁HUMOUR
- ▁FITTED
- ▁EXTERNAL
- ▁CETERA
- ▁WRAPPED
- ▁WHIT
- ▁FRED
- ▁EXAMINATION
- ▁LODGING
- ▁OWING
- ▁JAW
- ▁CROW
- ▁BALANCE
- ▁PUFF
- ▁TENDERNESS
- ▁PORTHOS
- ▁ANCHOR
- ▁INTERRUPT
- ▁NECESSARILY
- ▁PERPETUAL
- ▁AGONY
- ▁POPE
- ▁SCHOLAR
- ▁SCOTLAND
- ▁SUPPRESS
- ▁WRATH
- ▁WRECK
- ▁EXCEED
- ▁PERFECTION
- ▁INDIA
- ▁TRADITION
- ▁SECTION
- ▁EASTERN
- ▁DOORWAY
- ▁WIVES
- ▁CONVENTION
- ▁ANNOUNC
- ▁EGYPT
- ▁CONTRADICT
- ▁SCRATCH
- ▁CENTRAL
- ▁GLOVE
- ▁WAX
- ▁PREPARE
- ▁ACCOMPANY
- ▁INCREASING
- ▁LIBERAL
- ▁RAISING
- ▁ORANGE
- ▁SHOE
- ▁ATTRIBUTE
- ▁LITERATURE
- ▁PUZZLED
- ▁WITHDRAW
- ▁WHITHER
- ▁HAWK
- ▁MOONLIGHT
- ▁EXAMINE
- ▁HAPPILY
- ▁PRECEDE
- ▁DETECTIVE
- ▁INCHES
- ▁SOLITARY
- ▁DUTCH
- ▁NAPOLEON
- ▁UNEASY
- ▁CARDINAL
- ▁BLEW
- ▁FOWL
- ▁DECORAT
- ▁CHILDHOOD
- ▁TORMENT
- ▁LOSING
- ▁PERMISSION
- ▁BLANK
- ▁UPSTAIRS
- ▁CAPACITY
- ▁TRIFLE
- ▁FOLLY
- ▁RECOGNIZE
- ▁REMOVE
- ▁VENGEANCE
- ▁ENTERPRISE
- ▁BEDROOM
- ▁ANYHOW
- ▁INQUIRY
- ▁ASHES
- ▁DRAG
- ▁HUSH
- ▁AWKWARD
- ▁SATURDAY
- ▁GENUINE
- ▁SURVIV
- ▁SKIRT
- ▁AFFECTIONATE
- ▁TANG
- ▁MUTUAL
- ▁DISPUTE
- ▁EAGLE
- ▁INCOME
- ▁BIND
- ▁FAME
- ▁IMPROVEMENT
- ROVING
- ▁DIFFER
- ▁AWOKE
- ▁SLEEVE
- ▁SOLITUDE
- ▁FAVOURITE
- JI
- ▁DETECT
- ▁COMPREHEND
- ▁PREPARING
- ▁SERPENT
- ▁SUMMIT
- ▁KNOT
- ▁KNIT
- ▁COPY
- ▁STOPPING
- ▁FADED
- ▁HIDEOUS
- ▁JULIE
- STEAD
- ▁SHINE
- ▁CONFLICT
- ▁PROPOSITION
- ▁REFUGE
- ▁GALLERY
- ▁BUNDLE
- ▁AXE
- ▁SLAVERY
- ▁MASK
- ▁ALYOSHA
- ▁LADDER
- ▁DEPARTMENT
- ▁DISCHARGE
- ▁DEPRESS
- ▁GALLOP
- ▁SCARLET
- ▁KITTY
- ▁RECEIVING
- ▁SURRENDER
- ▁SUSTAIN
- ▁TWILIGHT
- ▁CONGRESS
- ▁IRELAND
- ▁FUNNY
- ▁LEND
- ▁CONSTITUTE
- ▁FUNERAL
- ▁CRYSTAL
- ▁SPAIN
- ▁EXCEEDINGLY
- ▁DAMN
- ▁COMMUN
- ▁CIVILIZATION
- ▁PREJUDICE
- ▁PORCH
- ▁ASSISTANT
- ▁INDUSTRY
- ▁TUMBLE
- ▁DEFENCE
- ▁HITHER
- ▁SMOT
- ▁COLONI
- ▁AMAZEMENT
- ▁MARGUERITE
- ▁MIRACLE
- ▁INHERIT
- ▁BEGGAR
- ▁ENVELOPE
- ▁INDIGNATION
- ▁NATASHA
- ▁PROPOSAL
- ▁FRAGMENT
- ▁ROUSED
- ▁ROAST
- ENCIES
- ▁COMMENCED
- ▁RESOURCE
- ▁POPULATION
- ▁QUOTH
- ▁PURSUE
- ▁EDUCAT
- ▁AFFLICT
- ▁CONTACT
- ▁CRIMSON
- ▁DIVISION
- ▁DISORDER
- ▁COPPER
- ▁SOLICIT
- ▁MODERATE
- ▁DRUM
- ▁SWIM
- ▁SALUTE
- ▁ASSUME
- ▁MUSCLE
- ▁OVERWHELM
- ▁SHAKESPEARE
- ▁STRUGGLING
- ▁TRANQUIL
- ▁CHICKEN
- ▁TREAD
- ▁CLAW
- ▁BIBLE
- ▁RIDGE
- ▁THREAT
- ▁VELVET
- ▁EXPOSED
- ▁IDIOT
- ▁BARREL
- ▁PENNY
- ▁TEMPTATION
- ▁DANGLARS
- ▁CENTURIES
- ▁DISTRIBUT
- ▁REJECT
- ▁RETORTED
- ▁CONCENTRAT
- ▁CORDIAL
- ▁MOTOR
- ▁CANNON
- KEEP
- ▁WRETCH
- ▁ASSURANCE
- ▁THIEF
- ▁SURVEY
- ▁VITAL
- ▁RAILWAY
- ▁JACKSON
- ▁CRASH
- ▁GROWL
- ▁COMBAT
- ▁RECOLLECTION
- ▁SECURITY
- ▁JACOB
- ▁CLUTCH
- ▁BLANKET
- ▁NANCY
- ▁CELLAR
- ▁CONVENIENT
- ▁INDIGNANT
- ▁COARSE
- ▁WORM
- ▁SCREEN
- ▁TRANSPORT
- ▁BULLET
- ▁APPRECIATE
- ▁DEVOTION
- ▁INVISIBLE
- ▁DRIED
- ▁MIXTURE
- ▁CANDID
- ▁PERFORMANCE
- ▁RIPE
- ▁EXQUISITE
- ▁BARGAIN
- ▁TOBACCO
- ▁LOYAL
- ▁MOULD
- ▁ATTENTIVE
- ▁DOROTHY
- ▁BRUTE
- ▁ESTABLISHMENT
- ▁ABILITY
- ▁INHABIT
- ▁OBSCURE
- ▁BORROW
- ▁ESSENCE
- ▁DISMAY
- ▁FLEE
- ▁BLADE
- ▁PLUCK
- ▁COFFIN
- ▁SUNSET
- ▁STEPHEN
- ▁ECONOMIC
- ▁HOLIDAY
- ▁MECHANICAL
- ▁COTTON
- ▁AWAKENED
- ▁SEIZE
- ▁RIDICULOUS
- ▁SANCHO
- ▁HESITATION
- ▁CORPSE
- ▁SAVING
- HOLD
- FOOT
- ▁ELDEST
- ▁DESPITE
- ▁EDITH
- ▁CHERISH
- ▁RESISTANCE
- ▁WILSON
- ▁ARGUE
- ▁INQUIRE
- ▁APPREHENSION
- ▁AVENUE
- ▁DRAKE
- ▁PROPOSE
- HURST
- ▁INFERIOR
- ▁STAIRCASE
- ▁WHEREFORE
- ▁CARLYLE
- ▁COUCH
- ▁ROUTE
- ▁POLITICS
- ▁TOMORROW
- ▁THRONG
- ▁NAUGHT
- ▁SUNLIGHT
- ▁INDIFFERENCE
- ▁OBEDIENCE
- ▁RECEPTION
- ▁VEGETABLE
- ▁IMPERFECT
- ▁RESIDENCE
- ▁TURKEY
- ▁VIOLET
- ▁SARAH
- ▁ALTAR
- ▁GRIEVE
- ▁JERK
- ▁ENSU
- ▁MAGICIAN
- ▁BLOSSOM
- ▁LANTERN
- ▁RESOLUTE
- ▁THOUGHTFULLY
- ▁FORTNIGHT
- ▁TRUMPET
- ▁VALJEAN
- ▁UNWILLING
- ▁LECTURE
- ▁WHEREUPON
- ▁HOLLAND
- ▁CHANGING
- ▁CREEK
- ▁SLICE
- ▁NORMAL
- ▁ANNIE
- ▁ACCENT
- ▁FREDERICK
- ▁DISAGREEABLE
- ▁RUBBED
- ▁DUMB
- ▁ESTABLISH
- ▁IMPORT
- ▁AFFIRM
- ▁MATTHEW
- ▁BRISK
- ▁CONVERT
- ▁BENDING
- ▁IVAN
- ▁MADEMOISELLE
- ▁MICHAEL
- ▁EASIER
- ▁JONES
- ▁FACING
- ▁EXCELLENCY
- ▁LITERARY
- ▁GOSSIP
- ▁DEVOUR
- ▁STAGGER
- ▁PENCIL
- ▁AVERAGE
- ▁HAMMER
- ▁TRIUMPHANT
- ▁PREFERRED
- ▁APPLICATION
- ▁OCCUPY
- ▁AUTHORITIES
- BURN
- ▁ASCERTAIN
- ▁CORRIDOR
- ▁DELICIOUS
- ▁PRACTISE
- ▁UNIVERSE
- ▁SHILLING
- ▁CONTEST
- ▁ASHORE
- ▁COMMIT
- ▁ADMINISTRATION
- ▁STUDIED
- ▁RIGID
- ▁ADORN
- ▁ELSEWHERE
- ▁INNOCENCE
- ▁JOURNAL
- ▁LANDSCAPE
- ▁TELEGRAPH
- ▁ANGRILY
- ▁CAMPAIGN
- ▁UNJUST
- ▁CHALLENGE
- ▁TORRENT
- ▁RELATE
- ▁ASSEMBLED
- ▁IMPRESSED
- ▁CANOE
- ▁CONCLUD
- ▁QUIXOTE
- ▁SATISFACTORY
- ▁NIECE
- ▁DEAF
- ▁RAFT
- ▁JIMMY
- ▁GLID
- ▁REGULAT
- ▁CHATTER
- ▁GLACIER
- ▁ENVY
- ▁STATUE
- ▁BOSTON
- ▁RICHMOND
- ▁DENIED
- ▁FANNY
- ▁SOLOMON
- ▁VULGAR
- ▁STALK
- ▁REPLACE
- ▁SPOON
- ▁BASIN
- ▁FEATURE
- ▁CONVICT
- ▁ARCHITECT
- ▁ADMIRAL
- ▁RIBBON
- ▁PERMANENT
- ▁APRIL
- ▁JOLLY
- ▁NEIGHBORHOOD
- ▁IMPART
- BOROUGH
- CAMP
- ▁HORRID
- ▁IMMORTAL
- ▁PRUDENCE
- ▁SPANIARD
- ▁SUPPOSING
- ▁TELEPHONE
- ▁TEMPERATURE
- ▁PENETRATE
- ▁OYSTER
- ▁APPOINTMENT
- ▁EGYPTIAN
- ▁DWELT
- ▁NEPHEW
- ▁RAILROAD
- ▁SEPTEMBER
- ▁DEVICE
- ▁WHEAT
- ▁GILBERT
- ▁ELEGANT
- ▁ADVERTISE
- ▁RATIONAL
- ▁TURTLE
- ▁BROOD
- ▁ASSEMBLY
- ▁CULTIVATE
- ▁EDITOR
- ▁SPECIMEN
- ▁UNDOUBTEDLY
- ▁WHALE
- ▁DROPPING
- ▁BALLOON
- ▁MEDICAL
- COMB
- ▁COMPOSITION
- ▁FOOTSTEPS
- ▁LAUNCELOT
- ▁DISCOURSE
- ▁ERRAND
- ▁CONVERSE
- ▁ADVANCING
- ▁DOWNSTAIRS
- ▁TUMULT
- ▁CORRUPT
- ▁SUFFICE
- ▁ANGUISH
- ▁SHAGGY
- ▁RETIRE
- ▁TIMBER
- ▁BLAZE
- ▁ABSTRACT
- ▁EMBROIDER
- ▁PHOTOGRAPH
- ▁PROSPERITY
- ▁TERRIBLY
- ▁TERRITORY
- ▁THRESHOLD
- ▁PAVEMENT
- ▁INJURED
- ▁LIMP
- ▁AGITATION
- ▁RASCAL
- ▁PRESUME
- ▁OBSERVING
- ▁OBSTACLE
- ▁SIMPLICITY
- ▁SLUMBER
- ▁SUPPLIED
- ▁COMBINATION
- ▁DRAIN
- ▁WILDERNESS
- ▁BELIEVING
- ▁VILLAIN
- ▁RECKLESS
- ▁INJURY
- ▁CLAPP
- ▁FRIDAY
- ▁HERCULES
- ▁KENNEDY
- ▁SYMPTOM
- ▁SLEDGE
- ▁CEILING
- ▁LEMON
- ▁PLAGUE
- ▁MONDAY
- ▁CANVAS
- ▁IMPATIENCE
- ▁UNCOMFORTABLE
- ▁ACCESS
- ▁FROZEN
- ▁SENATOR
- ▁FRANZ
- ▁SWIMMING
- ▁BARRIER
- ▁ADJUST
- ▁COMPARISON
- ▁PROCLAIM
- ▁WRINKL
- ▁OVERLOOK
- ▁MITYA
- ▁GUILT
- ▁PERCEPTION
- ▁PRECAUTION
- ▁SPECTATOR
- ▁SURPRISING
- ▁DISTRACT
- ▁DISDAIN
- ▁BONNET
- ▁MAGNET
- ▁PROFESS
- ▁CONFOUND
- ▁NARRATIVE
- ▁STRUCTURE
- ▁SKETCH
- ▁ULTIMATE
- ▁GLOBE
- ▁INSECT
- FICIENCY
- ▁ORCHARD
- ▁AMIABLE
- ▁DESCENT
- ▁INDEPENDENCE
- ▁MANUFACTURE
- ▁SPRINKLE
- ▁NIGHTINGALE
- ▁CUSHION
- ▁EMINENT
- ▁SCOTT
- ▁ARRAY
- ▁COSETTE
- ▁WAVING
- ▁EXTRACT
- ▁IRREGULAR
- ▁PERSECUT
- ▁DERIVED
- ▁WITHDREW
- ▁CAUTION
- ▁SUSPICIOUS
- ▁MEMORIES
- ▁NOWHERE
- ▁SUBTLE
- ▁THOROUGH
- Q
- ▁APPROPRIATE
- ▁SLAUGHTER
- ▁YOURSELVES
- ▁THUMB
- ▁TWAS
- ▁ABODE
- ▁BIDDING
- ▁CONSPICUOUS
- ▁REBECCA
- ▁SERGEANT
- ▁APRON
- ▁ANTICIPATE
- ▁DISCIPLINE
- ▁GLANCING
- ▁PILGRIM
- ▁SULLEN
- ▁CONTRIBUTE
- ▁PRAIRIE
- ▁CARVED
- ▁COMMERCE
- ▁EXCLAMATION
- ▁MUSCULAR
- ▁NOVEMBER
- ▁PHENOMENA
- ▁SYMBOL
- ▁UMBRELLA
- ▁DIMINISH
- ▁PARLOUR
- ▁THREATENING
- ▁STUMP
- ▁EXTENSIVE
- ▁PLEASING
- ▁REMEMBRANCE
- ▁COMBINED
- ▁SHERIFF
- ▁SHAFT
- ▁LAURA
- ▁INTERCOURSE
- ▁STRICKEN
- ▁SUPPLIES
- ▁LANDLORD
- ▁SHRINK
- ▁PRICK
- ▁CAESAR
- ▁DRUG
- ▁BEWILDERED
- ▁NAUTILUS
- ▁BRUTAL
- ▁COMMERCIAL
- ▁MAGGIE
- ▁SPHERE
- ▁VIRGIN
- ▁BRETHREN
- ▁DESTINY
- ▁POLICY
- ▁TERRIFIED
- ▁HOUSEKEEPER
- ▁CRAZY
- ▁ARDENT
- ▁DISCERN
- ▁WRAP
- ▁MARQUIS
- ▁RUSSIA
- MOUTH
- ▁BRITAIN
- ▁HARBOUR
- ▁CONCERT
- ▁DONKEY
- ▁DAMAGE
- ▁SLIM
- ABOUT
- ▁LUXURY
- ▁MONSTROUS
- ▁TENDENCY
- ▁PARADISE
- ▁CULTURE
- ▁JULIUS
- ▁RAOUL
- ▁REMEDY
- ▁DECAY
- ▁SCOLD
- ▁SPLIT
- ▁ASSAULT
- ▁DECEMBER
- ▁MOSCOW
- ▁EXPLORE
- ▁TROUSERS
- ▁WRIST
- PIECE
- ▁MUSKET
- ▁VALENTINE
- ▁TYRANT
- ▁ABRAHAM
- ▁MEDIUM
- ▁ARTIFICIAL
- ▁FACULTY
- ▁OBLIGATION
- ▁RESEMBLANCE
- ▁INQUIRIES
- ▁DETAIN
- ▁SWARM
- ▁PLEDGE
- ▁ADMIRABLE
- ▁DEFECT
- ▁SUPERINTEND
- ▁PATRIOT
- ▁CLUNG
- ▁DISMAL
- ▁RECIT
- ▁IGNOR
- ▁AMELIA
- ▁JUSTIFY
- ▁ELEPHANT
- ▁ESTIMATE
- ▁KNELT
- ▁SERVING
- ▁WHIM
- ▁SHRILL
- ▁STUDIO
- ▁TEXT
- ▁ALEXANDER
- ▁WROUGHT
- ▁ABUNDANT
- ▁SITUATED
- ▁REGAIN
- ▁FIERY
- ▁SNEER
- ▁SWEAT
- ▁GLARE
- ▁NIGH
- ▁ESCORT
- ▁INEVITABLE
- ▁PSMITH
- ▁RELUCTANT
- ▁PRECEDING
- ▁RESORT
- ▁OUTRAGE
- ▁AMBASSADOR
- ▁CONSOLATION
- ▁RECOGNITION
- ▁REMORSE
- ▁BEHALF
- ▁FORMIDABLE
- ▁GRAVITY
- ▁DIVIDE
- ▁CONFRONT
- ▁GIGANTIC
- ▁OCTOBER
- ▁FLANK
- ▁SLEW
- ▁CLARA
- ▁FILM
- ▁BULK
- ▁POMP
- ▁ELEANOR
- ▁EMPHASIS
- ▁JAPANESE
- ▁CAVALRY
- ▁EXCLUSIVE
- ▁PERFUME
- ▁BRONZE
- ▁FEDERAL
- ▁LIQUID
- ▁RUBBING
- ▁OVEN
- DOLPH
- ▁CONVULS
- ▁DEPRIVED
- ▁RESPONSIBILITY
- ▁SIGNIFICANT
- ▁WAISTCOAT
- ▁CLUSTER
- ▁MARTHA
- ▁REVERSE
- ▁ATTORNEY
- ▁DROOP
- ▁SKILFUL
- ▁HABITUAL
- ▁PUMP
- ▁INTERVEN
- ▁OWL
- ▁CONJECTURE
- ▁FANTASTIC
- ▁RESPONSIBLE
- ▁DESTINED
- ▁DOCUMENT
- ▁THEREUPON
- ▁GODDESS
- ▁PACIFIC
- ▁WARRANT
- ▁COSTUME
- ▁BRIDLE
- ▁CALIFORNIA
- ▁DEMOCRATIC
- ▁EUSTACE
- ▁SQUIRREL
- ▁UNCOMMON
- ▁MARVELLOUS
- ▁PLOUGH
- ▁TRAGEDY
- ▁VAULT
- ▁HESITATE
- ▁REFRAIN
- ▁ADMIRING
- ▁CORPORAL
- ▁ENTITLED
- ▁SHREWD
- ▁SQUEEZ
- ▁ACCURATE
- ▁TEMPEST
- ▁MONUMENT
- ▁SIEGE
- ▁CHINESE
- ▁RAVEN
- ▁LOUNG
- ▁ASSASSIN
- ▁INFLICT
- ▁AGITATED
- ▁DESIRABLE
- ▁EARLIEST
- ▁LAUNCH
- ▁PILOT
- ▁PULSE
- ▁MUTE
- LEIGH
- ▁LIQUOR
- ▁SCARECROW
- ▁SKULL
- ▁DESOLATE
- ▁SUBLIME
- ▁SERENE
- ▁RECESS
- ▁WAKING
- ▁CHARLOTTE
- ▁CIRCULAR
- ▁INJUSTICE
- ▁PINOCCHIO
- ▁PRISCILLA
- ▁THYSELF
- ▁OCCURRENCE
- ▁CASUAL
- ▁FRANTIC
- ▁LEGEND
- ▁FERTIL
- ▁BACKGROUND
- ▁DELICACY
- ▁ESTRALLA
- ▁MANUSCRIPT
- ▁RESPONSE
- ▁UNIVERSITY
- ▁WOLVES
- ▁SCANDAL
- ▁STUMBLE
- ▁HOARSE
- ▁BODILY
- ▁CONVENT
- ▁EXAMINING
- ▁INCAPABLE
- ▁PERCEIVING
- ▁PHILADELPHIA
- ▁SUBSEQUENT
- ▁THIEVES
- ▁ACCUMULAT
- ▁DAMSEL
- ▁SCOTCH
- ▁UNDERNEATH
- ▁NOBILITY
- ▁SMASH
- ▁REVOLT
- ▁ENGAGE
- ▁CATHEDRAL
- ▁CHAMPION
- ▁DESPATCH
- ▁ETERNITY
- ▁JANUARY
- ▁PLEADED
- ▁PROBABILITY
- ▁JIMMIE
- ▁PARALLEL
- ▁FISHERMAN
- ▁JERRY
- ▁SWORE
- ▁DRAUGHT
- ▁OPPONENT
- ▁PRIMITIVE
- ▁SIGNIFICANCE
- ▁SUBSTANTIAL
- ▁AMAZED
- ▁DUNBAR
- ▁COMMEND
- ▁CONTEMPLATE
- ▁TESTIMONY
- ▁IMPERIAL
- ▁ADAPT
- ▁JUICE
- ▁CALAMIT
- CULAR
- ▁CHATEAU
- ▁PHOENIX
- ▁PRUDENT
- ▁SOLUTION
- ▁VILLEFORT
- ▁REACTION
- ▁RELAX
- ▁YU
- ▁PROHIBIT
- ▁DISTRUST
- ▁PLUNDER
- ▁WELFARE
- ▁NAVIGAT
- ▁PARLOR
- ▁LAZY
- ▁DETACH
- OMETER
- ▁PRIV
- ▁DISCOURAGE
- ▁OBSTINATE
- ▁REJOICING
- ▁SERMON
- ▁VEHICLE
- ▁FANCIES
- ▁ENLIGHTEN
- ▁ACUTE
- ▁ILLUSION
- ▁ANTHEA
- ▁MARTIAN
- ▁EXCITE
- ▁GENEROSITY
- OLOGIST
- ▁AMAZING
- ▁UNWORTHY
- ▁INTERNAL
- ▁INCENSE
- ▁VIBRAT
- ▁ADHERE
- ROACH
- ▁FEBRUARY
- ▁MEXICAN
- ▁POTATOES
- ▁INCESSANT
- ▁INTERPOSED
- ▁PARCEL
- ▁VEXED
- ▁PROMOTE
- MIDST
- ▁ARISTOCRAT
- ▁CYRIL
- ▁EMBARK
- ▁ABUNDANCE
- ▁LITERALLY
- ▁SURGEON
- ▁TERRACE
- ▁ATLANTIC
- ▁MARTYR
- ▁SPECK
- ▁SENATE
- ▁LOAF
- ▁ADMINISTER
- ▁APPREHEND
- ▁SUBDUED
- ▁TEMPORARY
- ▁DOMINION
- ▁ELABORATE
- ▁DIGNIFIED
- ▁ELIZA
- ▁SPLASH
- ▁CONSEIL
- ▁DEXTER
- ▁UNSEEN
- ▁TRAGIC
- VOCATION
- ▁GRATIFY
- ▁BACHELOR
- ▁DEFENSE
- ▁EXCURSION
- ▁FACULTIES
- ▁PROPRIETOR
- ▁SYMPATHETIC
- ▁UNNECESSARY
- ▁RADIANT
- ▁VACANT
- ▁OUNCE
- ▁SCREW
- ▁PHENOMENON
- ▁PROMINENT
- ▁WORRIED
- ▁STUDIES
- ▁CLIMATE
- ▁KEITH
- ▁ARAMIS
- ▁BLISS
- ▁CONTINUAL
- ▁SURPASS
- ▁HEBREW
- ▁IDENTITY
- ▁PROVOKE
- ▁TEMPERAMENT
- ▁CHARIOT
- ▁HARBOR
- ▁NINTH
- ▁PRIOR
- ▁DESIROUS
- ▁JERUSALEM
- ▁UNDERTAKING
- ▁EDISON
- ▁MIRTH
- ▁SCOUT
- ▁APPARATUS
- ▁ILLUSTRATION
- ▁INTELLIGIBLE
- ▁INVARIABLY
- ▁PIERCED
- ▁REVIEW
- ▁FLICKER
- ▁HAZARD
- ▁REVELATION
- ▁DIXON
- ▁EXCITING
- ▁GOSPEL
- ▁CONSTANCE
- ▁OVERTAKE
- ▁GUINEA
- ▁ALADDIN
- ▁CHICAGO
- ▁TULLIVER
- ▁HAMILTON
- ▁GARRISON
- ▁DISCIPLE
- ▁INTENSITY
- ▁TRAITOR
- ▁CHANCELLOR
- ▁PROVERB
- ▁DAGGER
- ▁FORESEE
- ▁CONFIDE
- ▁GLIMMER
- ▁CHAUVELIN
- ▁ILLUSTRATE
- ▁VOLUNTEER
- ▁JUNGLE
- ▁STREAK
- ▁SUNRISE
- ▁DISSOLV
- ▁QUEST
- ▁AWHILE
- ▁FELICITY
- ▁LEGISLATURE
- ▁LEONORA
- ▁MAGAZINE
- ▁PITIFUL
- ▁COLONY
- ▁SHAWL
- ▁ARRIVING
- ▁FUNDAMENTAL
- ▁CARPENTER
- ▁OVERFLOW
- ▁EXPAND
- ▁HARVEST
- ▁FEMININE
- ▁INNUMERABLE
- ▁SCRAMBLE
- ▁TWENTIETH
- ▁TRIFLING
- ▁GHASTL
- ▁CONQUEST
- ▁DANIEL
- ▁FACILIT
- ▁FORSAKE
- ▁BEHAVIOUR
- ▁GORGEOUS
- ▁PRODUCING
- ▁HAPPIER
- ▁PROMISING
- ▁RAINBOW
- ▁INSTINCTIVELY
- ▁DECREE
- ▁EYEBROWS
- ▁IRRESISTIBLE
- ▁PHARAOH
- ▁SCROOGE
- ▁UNNATURAL
- ▁CRUMBS
- ▁REFINED
- ▁DREARY
- ▁TRENCH
- ▁CONVINCE
- ▁FRINGE
- ▁EXTREMITY
- ▁INTIMACY
- ▁SCOUNDREL
- ▁SUFFRAGE
- ▁UNEASINESS
- ▁BARRICADE
- ▁CIRCULAT
- ▁SAMUEL
- ▁BRUCE
- ▁DARCY
- <sos/eos>
init: null
input_size: null
ctc_conf:
dropout_rate: 0.0
ctc_type: builtin
reduce: true
ignore_nan_grad: true
joint_net_conf: null
model_conf:
ctc_weight: 0.3
lsm_weight: 0.1
length_normalized_loss: false
use_preprocessor: true
token_type: bpe
bpemodel: data/en_token_list/bpe_unigram5000/bpe.model
non_linguistic_symbols: null
cleaner: null
g2p: null
speech_volume_normalize: null
rir_scp: null
rir_apply_prob: 1.0
noise_scp: null
noise_apply_prob: 1.0
noise_db_range: '13_15'
frontend: default
frontend_conf:
n_fft: 512
hop_length: 256
fs: 16k
specaug: specaug
specaug_conf:
apply_time_warp: true
time_warp_window: 5
time_warp_mode: bicubic
apply_freq_mask: true
freq_mask_width_range:
- 0
- 27
num_freq_mask: 2
apply_time_mask: true
time_mask_width_ratio_range:
- 0.0
- 0.05
num_time_mask: 10
normalize: global_mvn
normalize_conf:
stats_file: exp/asr_stats_raw_en_bpe5000_sp/train/feats_stats.npz
preencoder: null
preencoder_conf: {}
encoder: conformer
encoder_conf:
output_size: 512
attention_heads: 8
linear_units: 2048
num_blocks: 12
dropout_rate: 0.1
positional_dropout_rate: 0.1
attention_dropout_rate: 0.1
input_layer: conv2d
normalize_before: true
macaron_style: true
rel_pos_type: latest
pos_enc_layer_type: rel_pos
selfattention_layer_type: rel_selfattn
activation_type: swish
use_cnn_module: true
cnn_module_kernel: 31
postencoder: null
postencoder_conf: {}
decoder: transformer
decoder_conf:
attention_heads: 8
linear_units: 2048
num_blocks: 6
dropout_rate: 0.1
positional_dropout_rate: 0.1
self_attention_dropout_rate: 0.1
src_attention_dropout_rate: 0.1
required:
- output_dir
- token_list
version: 0.10.7a1
distributed: true
```
</details>
### Citing ESPnet
```BibTex
@inproceedings{watanabe2018espnet,
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
title={{ESPnet}: End-to-End Speech Processing Toolkit},
year={2018},
booktitle={Proceedings of Interspeech},
pages={2207--2211},
doi={10.21437/Interspeech.2018-1456},
url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
}
```
or arXiv:
```bibtex
@misc{watanabe2018espnet,
title={ESPnet: End-to-End Speech Processing Toolkit},
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
year={2018},
eprint={1804.00015},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
hyechanjun/interview-question-remake
|
hyechanjun
| 2022-03-07T17:57:47Z | 5 | 1 |
transformers
|
[
"transformers",
"pytorch",
"bart",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2022-03-07T17:48:30Z |
---
datasets:
- "INTERVIEW: NPR Media Dialog Transcripts"
---
# AI Interviewer Question-Asking Model
For a Senior Project at Calvin University
Created by: Hyechan Jun, Ha-Ram Koo, and Advait Scaria
This model is fine-tuned on facebook/bart-base to generate sequences ending in a question mark (?). It is a remake of an earlier model that had errors in its training and validation datasets.
|
Subsets and Splits
Filtered Qwen2.5 Distill Models
Identifies specific configurations of models by filtering cards that contain 'distill', 'qwen2.5', '7b' while excluding certain base models and incorrect model ID patterns, uncovering unique model variants.
Filtered Model Cards Count
Finds the count of entries with specific card details that include 'distill', 'qwen2.5', '7b' but exclude certain base models, revealing valuable insights about the dataset's content distribution.
Filtered Distill Qwen 7B Models
Filters for specific card entries containing 'distill', 'qwen', and '7b', excluding certain strings and patterns, to identify relevant model configurations.
Filtered Qwen-7b Model Cards
The query performs a detailed filtering based on specific keywords and excludes certain entries, which could be useful for identifying a specific subset of cards but does not provide deeper insights or trends.
Filtered Qwen 7B Model Cards
The query filters for specific terms related to "distilled" or "distill", "qwen", and "7b" in the 'card' column but excludes certain base models, providing a limited set of entries for further inspection.
Qwen 7B Distilled Models
The query provides a basic filtering of records to find specific card names that include keywords related to distilled Qwen 7b models, excluding a particular base model, which gives limited insight but helps in focusing on relevant entries.
Qwen 7B Distilled Model Cards
The query filters data based on specific keywords in the modelId and card fields, providing limited insight primarily useful for locating specific entries rather than revealing broad patterns or trends.
Qwen 7B Distilled Models
Finds all entries containing the terms 'distilled', 'qwen', and '7b' in a case-insensitive manner, providing a filtered set of records but without deeper analysis.
Distilled Qwen 7B Models
The query filters for specific model IDs containing 'distilled', 'qwen', and '7b', providing a basic retrieval of relevant entries but without deeper analysis or insight.
Filtered Model Cards with Distill Qwen2.
Filters and retrieves records containing specific keywords in the card description while excluding certain phrases, providing a basic count of relevant entries.
Filtered Model Cards with Distill Qwen 7
The query filters specific variations of card descriptions containing 'distill', 'qwen', and '7b' while excluding a particular base model, providing limited but specific data retrieval.
Distill Qwen 7B Model Cards
The query filters and retrieves rows where the 'card' column contains specific keywords ('distill', 'qwen', and '7b'), providing a basic filter result that can help in identifying specific entries.