Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,17 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
base_model: Qwen/Qwen2.5-32B
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
@@ -9,8 +20,57 @@ model-index:
|
|
| 9 |
results: []
|
| 10 |
---
|
| 11 |
|
| 12 |
-
|
| 13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
|
| 16 |
<details><summary>See axolotl config</summary>
|
|
@@ -368,67 +428,4 @@ weight_decay: 0.1
|
|
| 368 |
# fsdp_mixed_precision: BF16 # Added
|
| 369 |
```
|
| 370 |
|
| 371 |
-
</details><br>
|
| 372 |
-
|
| 373 |
-
# EVA-Qwen2.5-32B-SFFT-v0.1
|
| 374 |
-
|
| 375 |
-
This model is a fine-tuned version of [Qwen/Qwen2.5-32B](https://huggingface.co/Qwen/Qwen2.5-32B) on the None dataset.
|
| 376 |
-
It achieves the following results on the evaluation set:
|
| 377 |
-
- Loss: 0.9476
|
| 378 |
-
|
| 379 |
-
## Model description
|
| 380 |
-
|
| 381 |
-
More information needed
|
| 382 |
-
|
| 383 |
-
## Intended uses & limitations
|
| 384 |
-
|
| 385 |
-
More information needed
|
| 386 |
-
|
| 387 |
-
## Training and evaluation data
|
| 388 |
-
|
| 389 |
-
More information needed
|
| 390 |
-
|
| 391 |
-
## Training procedure
|
| 392 |
-
|
| 393 |
-
### Training hyperparameters
|
| 394 |
-
|
| 395 |
-
The following hyperparameters were used during training:
|
| 396 |
-
- learning_rate: 5e-05
|
| 397 |
-
- train_batch_size: 1
|
| 398 |
-
- eval_batch_size: 1
|
| 399 |
-
- seed: 42
|
| 400 |
-
- distributed_type: multi-GPU
|
| 401 |
-
- num_devices: 8
|
| 402 |
-
- gradient_accumulation_steps: 8
|
| 403 |
-
- total_train_batch_size: 64
|
| 404 |
-
- total_eval_batch_size: 8
|
| 405 |
-
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 406 |
-
- lr_scheduler_type: cosine
|
| 407 |
-
- lr_scheduler_warmup_steps: 20
|
| 408 |
-
- num_epochs: 3
|
| 409 |
-
|
| 410 |
-
### Training results
|
| 411 |
-
|
| 412 |
-
| Training Loss | Epoch | Step | Validation Loss |
|
| 413 |
-
|:-------------:|:------:|:----:|:---------------:|
|
| 414 |
-
| 1.4144 | 0.0078 | 1 | 1.3757 |
|
| 415 |
-
| 1.0796 | 0.2498 | 32 | 0.9875 |
|
| 416 |
-
| 1.0367 | 0.4995 | 64 | 0.9437 |
|
| 417 |
-
| 0.9919 | 0.7493 | 96 | 0.9212 |
|
| 418 |
-
| 0.9305 | 0.9990 | 128 | 0.9097 |
|
| 419 |
-
| 0.6963 | 1.2427 | 160 | 0.9228 |
|
| 420 |
-
| 0.686 | 1.4922 | 192 | 0.9187 |
|
| 421 |
-
| 0.6656 | 1.7417 | 224 | 0.9127 |
|
| 422 |
-
| 0.6818 | 1.9912 | 256 | 0.9029 |
|
| 423 |
-
| 0.467 | 2.2391 | 288 | 0.9575 |
|
| 424 |
-
| 0.459 | 2.4879 | 320 | 0.9502 |
|
| 425 |
-
| 0.4957 | 2.7366 | 352 | 0.9487 |
|
| 426 |
-
| 0.493 | 2.9854 | 384 | 0.9476 |
|
| 427 |
-
|
| 428 |
-
|
| 429 |
-
### Framework versions
|
| 430 |
-
|
| 431 |
-
- Transformers 4.45.1
|
| 432 |
-
- Pytorch 2.4.0+cu121
|
| 433 |
-
- Datasets 2.21.0
|
| 434 |
-
- Tokenizers 0.20.2
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
license: apache-2.0
|
| 4 |
+
datasets:
|
| 5 |
+
- anthracite-org/kalo-opus-instruct-22k-no-refusal
|
| 6 |
+
- Nopm/Opus_WritingStruct
|
| 7 |
+
- Gryphe/Sonnet3.5-SlimOrcaDedupCleaned
|
| 8 |
+
- Gryphe/Sonnet3.5-Charcard-Roleplay
|
| 9 |
+
- Gryphe/ChatGPT-4o-Writing-Prompts
|
| 10 |
+
- Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
|
| 11 |
+
- Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
|
| 12 |
+
- nothingiisreal/Reddit-Dirty-And-WritingPrompts
|
| 13 |
+
- allura-org/Celeste-1.x-data-mixture
|
| 14 |
+
- cognitivecomputations/dolphin-2.9.3
|
| 15 |
base_model: Qwen/Qwen2.5-32B
|
| 16 |
tags:
|
| 17 |
- generated_from_trainer
|
|
|
|
| 20 |
results: []
|
| 21 |
---
|
| 22 |
|
| 23 |
+
# EVA Qwen2.5-32B v0.2
|
| 24 |
+
|
| 25 |
+
<p>
|
| 26 |
+
A RP/storywriting specialist model, full-parameter finetune of Qwen2.5-32B on mixture of synthetic and natural data.<br>
|
| 27 |
+
It uses Celeste 70B 0.1 data mixture, greatly expanding it to improve versatility, creativity and "flavor" of the resulting model.<br>
|
| 28 |
+
</p>
|
| 29 |
+
|
| 30 |
+
<p>Dedicated to Nev.</p>
|
| 31 |
+
|
| 32 |
+
<p><b>Version notes for 0.2</b>: Basically, reprocessed the whole dataset again, due to a severe mistake in previously used pipeline, while left the data poisoned with a lot of non-unicode characters. Now, no more weird generation artifacts, and more stability. Major kudos to Cahvay for his work on fixing this critical issue.</p>
|
| 33 |
+
|
| 34 |
+
<p>
|
| 35 |
+
<p>Prompt format is ChatML.</p><br>
|
| 36 |
+
<h3>Recommended sampler values:</h3>
|
| 37 |
+
<ul>
|
| 38 |
+
<li>Temperature: 1</li>
|
| 39 |
+
<li>Min-P: 0.05</li>
|
| 40 |
+
<li>Top-A: 0.2</li>
|
| 41 |
+
<li>Repetition Penalty: 1.03</li>
|
| 42 |
+
</ul>
|
| 43 |
+
|
| 44 |
+
<h3>Recommended SillyTavern presets (via CalamitousFelicitousness):</h3>
|
| 45 |
+
|
| 46 |
+
- [Context](https://huggingface.co/EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1/blob/main/%5BChatML%5D%20Roleplay-v1.9%20Context.json)
|
| 47 |
+
- [Instruct and System Prompt](https://huggingface.co/EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1/blob/main/%5BChatML%5D%20Roleplay-v1.9%20Instruct.json)
|
| 48 |
+
</p>
|
| 49 |
+
|
| 50 |
+
<p>
|
| 51 |
+
<br>
|
| 52 |
+
<h3>
|
| 53 |
+
Training data:
|
| 54 |
+
</h3>
|
| 55 |
+
<ul>
|
| 56 |
+
<li>Celeste 70B 0.1 data mixture minus Opus Instruct subset. See that model's <a href=https://huggingface.co/nothingiisreal/L3.1-70B-Celeste-V0.1-BF16>card</a> for details.</li>
|
| 57 |
+
<li>Kalomaze's Opus_Instruct_25k dataset, filtered for refusals.</li>
|
| 58 |
+
<li>A subset (1k rows) of ChatGPT-4o-WritingPrompts by Gryphe</li>
|
| 59 |
+
<li>A subset (2k rows) of Sonnet3.5-Charcards-Roleplay by Gryphe</li>
|
| 60 |
+
<li>Synthstruct and SynthRP datasets by Epiculous</li>
|
| 61 |
+
<li>A subset from Dolphin-2.9.3, including filtered version of not_samantha and a small subset of systemchat.</li>
|
| 62 |
+
</ul>
|
| 63 |
+
<h3>
|
| 64 |
+
Training time and hardware:
|
| 65 |
+
</h3>
|
| 66 |
+
<ul><li>7 hours on 8xH100 SXM, provided by <a href=https://featherless.ai/>FeatherlessAI</a></li></ul><br>
|
| 67 |
+
</p>
|
| 68 |
+
<p>Model was created by Kearm, Auri and Cahvay.</p>
|
| 69 |
+
<h4>Special thanks:</h4><ul>
|
| 70 |
+
<li><b>to Cahvay for his work on investigating and reprocessing the corrupted dataset, removing the single biggest source of data poisoning.</b></li>
|
| 71 |
+
<li><b>to <a href=https://featherless.ai/>FeatherlessAI</a> for generously providing 8xH100 SXM node for training of this model</b></li>
|
| 72 |
+
<li>to Gryphe, Lemmy, Kalomaze, Nopm, Epiculous and CogninitiveComputations for the data</li>
|
| 73 |
+
<li>and to Allura-org for support, feedback, beta-testing and doing quality control of EVA models.</li></ul>
|
| 74 |
|
| 75 |
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
|
| 76 |
<details><summary>See axolotl config</summary>
|
|
|
|
| 428 |
# fsdp_mixed_precision: BF16 # Added
|
| 429 |
```
|
| 430 |
|
| 431 |
+
</details><br>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|