End of training
Browse files- README.md +43 -46
- adapter_model.safetensors +1 -1
README.md
CHANGED
@@ -1,60 +1,57 @@
|
|
1 |
-
---
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
tags:
|
6 |
-
- generated_from_trainer
|
7 |
-
|
8 |
-
-
|
9 |
-
|
10 |
-
---
|
11 |
|
12 |
-
|
13 |
-
should probably proofread and complete it, then remove this comment. -->
|
14 |
|
15 |
-
|
|
|
16 |
|
17 |
-
|
18 |
-
It achieves the following results on the evaluation set:
|
19 |
-
- eval_loss: 1.0495
|
20 |
-
- eval_runtime: 978.1933
|
21 |
-
- eval_samples_per_second: 8.966
|
22 |
-
- eval_steps_per_second: 1.121
|
23 |
-
- epoch: 0.6
|
24 |
-
- step: 1200
|
25 |
|
26 |
-
|
|
|
27 |
|
28 |
-
|
|
|
|
|
|
|
|
|
29 |
|
30 |
-
##
|
31 |
|
32 |
-
|
33 |
|
34 |
-
|
35 |
|
36 |
-
|
37 |
|
38 |
-
|
|
|
|
|
|
|
|
|
39 |
|
40 |
-
|
41 |
|
42 |
-
The following hyperparameters were used during training:
|
43 |
-
- learning_rate: 0.0005
|
44 |
-
- train_batch_size: 8
|
45 |
-
- eval_batch_size: 8
|
46 |
-
- seed: 42
|
47 |
-
- gradient_accumulation_steps: 2
|
48 |
-
- total_train_batch_size: 16
|
49 |
-
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
50 |
-
- lr_scheduler_type: cosine
|
51 |
-
- lr_scheduler_warmup_steps: 30
|
52 |
-
- training_steps: 2000
|
53 |
|
54 |
-
### Framework versions
|
55 |
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: ybelkada/falcon-7b-sharded-bf16
|
3 |
+
library_name: transformers
|
4 |
+
model_name: peft-starcoder-lora-a100
|
5 |
+
tags:
|
6 |
+
- generated_from_trainer
|
7 |
+
- trl
|
8 |
+
- sft
|
9 |
+
licence: license
|
10 |
+
---
|
11 |
|
12 |
+
# Model Card for peft-starcoder-lora-a100
|
|
|
13 |
|
14 |
+
This model is a fine-tuned version of [ybelkada/falcon-7b-sharded-bf16](https://huggingface.co/ybelkada/falcon-7b-sharded-bf16).
|
15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
16 |
|
17 |
+
## Quick start
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
18 |
|
19 |
+
```python
|
20 |
+
from transformers import pipeline
|
21 |
|
22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
23 |
+
generator = pipeline("text-generation", model="SoundsFun/peft-starcoder-lora-a100", device="cuda")
|
24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
25 |
+
print(output["generated_text"])
|
26 |
+
```
|
27 |
|
28 |
+
## Training procedure
|
29 |
|
30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/pasechnikm-mephi/huggingface/runs/ke1xns28)
|
31 |
|
32 |
+
This model was trained with SFT.
|
33 |
|
34 |
+
### Framework versions
|
35 |
|
36 |
+
- TRL: 0.12.0
|
37 |
+
- Transformers: 4.49.0
|
38 |
+
- Pytorch: 2.5.1
|
39 |
+
- Datasets: 3.4.0
|
40 |
+
- Tokenizers: 0.21.0
|
41 |
|
42 |
+
## Citations
|
43 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
44 |
|
|
|
45 |
|
46 |
+
Cite TRL as:
|
47 |
+
|
48 |
+
```bibtex
|
49 |
+
@misc{vonwerra2022trl,
|
50 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
51 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin GallouГ©dec},
|
52 |
+
year = 2020,
|
53 |
+
journal = {GitHub repository},
|
54 |
+
publisher = {GitHub},
|
55 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
56 |
+
}
|
57 |
+
```
|
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 261131840
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf78f7aa6af00dae63a39eed8db459f48adc33b3d3934d0a4dca159dfdae0c8e
|
3 |
size 261131840
|