Quantization made by Richard Erkhov.
tinyllama-1.1b-mt-dpo-full_LR5e-8_BS16_rmsprop_2epochs - AWQ
- Model creator: https://huggingface.co/martimfasantos/
- Original model: https://huggingface.co/martimfasantos/tinyllama-1.1b-mt-dpo-full_LR5e-8_BS16_rmsprop_2epochs/
Original model description:
license: apache-2.0 base_model: martimfasantos/tinyllama-1.1b-mt-sft-full tags: - alignment-handbook - trl - dpo - generated_from_trainer - trl - dpo - generated_from_trainer datasets: - haoranxu/ALMA-R-Preference model-index: - name: tinyllama-1.1b-mt-dpo-full_LR5e-8_BS16_rmsprop_2epochs results: []
tinyllama-1.1b-mt-dpo-full_LR5e-8_BS16_rmsprop_2epochs
This model is a fine-tuned version of martimfasantos/tinyllama-1.1b-mt-sft-full on the haoranxu/ALMA-R-Preference 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-08
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2
Training results
Framework versions
- Transformers 4.41.2
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1