Openrlhf

Changes
- Default to using `torch.cuda.device_count()` for `tp_size` in `batch_inference` tongyx361
- Improved description of `tqdm` tongyx361
- Fixed loading dataset from local text files tongyx361
- Added support for Llama3.1 xiaoxigua999
- Added `--packing_samples` support for all HF models (SFT/DPO/RM training) xiaoxigua999
- Added `--nll_loss_coef` (for chosen response) support for DPO xiaoxigua999

Changes
- Added support for `--packing_samples` in DPO/RM training (xiaoxigua999)
- Updated `reward_dataset` to correctly handle `prompt_key` (Nickydusk)
- Updated versions of Transformers and DeepSpeed (openllmai0)

Changes
- Refactored the `parser.parse_args()` and added `--train_split` and `--test_split` openllmai0
- Added support for running with `openrlhf.cli.train_ppo` as the module name openllmai0
- Fixed PyPI workflows (you can now use `pip install openrlhf`) hijkzzz

Changes
- Fixed Qwen2 + FlashAttention2 openllmai0
- Fixed Right Padding in DPO and KTO openllmai0
- Fixed default input_key for Iterative DPO openllmai0
- Use `cosine_with_min_lr` openllmai0
- New `OpenRLHF` Logo hijkzzz

Changes
- Refactored the KTO Trainer openllmai0
- Fixed issues with KTO/DPO datasets openllmai0
- Added SFT Packing feature openllmai0
- Supported vLLM 0.5.1 (via Gloo) openllmai0

Changes
- Refactored README.md and scripts openllmai0
- Cleaned Dataset Classes openllmai0
- Fixed [tie_word_embeddings save_model](https://github.com/OpenLLMAI/OpenRLHF/commit/79d1d5873212c1baa0552042bb242482f7066a34) openllmai0
- Added [Read the Docs](https://openrlhf.readthedocs.io/) hijkzzz