| Quantization made by Richard Erkhov. | |
| [Github](https://github.com/RichardErkhov) | |
| [Discord](https://discord.gg/pvy7H8DZMG) | |
| [Request more models](https://github.com/RichardErkhov/quant_request) | |
| pygm-350m-experimental - AWQ | |
| - Model creator: https://huggingface.co/alpindale/ | |
| - Original model: https://huggingface.co/alpindale/pygm-350m-experimental/ | |
| Original model description: | |
| --- | |
| tags: | |
| - generated_from_trainer | |
| metrics: | |
| - accuracy | |
| model-index: | |
| - name: pygmalion-350m | |
| results: [] | |
| --- | |
| # pygmalion-350m | |
| This model is a fine-tuned version of [PygmalionAI/pygmalion-350m](https://huggingface.co/PygmalionAI/pygmalion-350m/) on a 2.4MB dataset. | |
| It achieves the following results on the evaluation set: | |
| - Loss: 2.2731 | |
| - Accuracy: 0.5187 | |
| ## Model description | |
| A proof-of-concept model based on PygmalionAI/pygmalion-350m, which was in turn based on OPT-350m. | |
| This model was fine-tuned purely for testing purposes. | |
| ## Fine-tuning process | |
| Fine-tuned on an A100-80GB with HF's `run_clm.py` script. It was run through 3 epochs with 8 batch size using 2.4MB dataset (split 75/25 between training and validation sets). | |
| ### 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: 3.0 | |
| ### Framework versions | |
| - Transformers 4.27.0.dev0 | |
| - Pytorch 1.13.1+cu117 | |
| - Datasets 2.10.0 | |
| - Tokenizers 0.13.2 | |