internlm2_7b_w_untokenized_dataset internlm2_chat_1_8b_dpo_full internlm2_chat_1_8b_dpo_full_varlenattn internlm2_chat_1_8b_dpo_full_varlenattn_jsonl_dataset internlm2_chat_1_8b_full_alpaca_e3 internlm2_chat_1_8b_orpo_full internlm2_chat_1_8b_orpo_full_varlenattn internlm2_...
目前我们也在支持强化学习范式,例如 RLHF 和 DPO 等特性。 (3) 高效训练 自动分发高性能算子(如 FlashAttention、Triton kernels 等)以加速训练吞吐 兼容DeepSpeed ,轻松应用各种 ZeRO 训练优化策略。 我们近期也在支持序列并行(基本完成),可以快速轻松的训练超长 token 的训练。如果没有序列并行,想利用 DeepSpeed+...
QLoRA LoRA Full parameter fine-tune DPO ORPO Reward Model 🛠️ Quick Start Installation It is recommended to build a Python-3.10 virtual environment using conda conda create --name xtuner-env python=3.10 -y conda activate xtuner-env Install XTuner via pip pip install -U xtuner or with ...
示例: -曲调:Medical-Llama3-v2,sfr-iterative dpo-lalama-3-8b,Meta-llama-3.1-8b-指令-轻松 -投影机:llava-llama-3-8b-v1_1-mmproj、llava-lrama-3.1-8b-mmproj-f16、minicpm-V-2-5-mmproj ##理论 -CLIP模型通常由视觉编码器和文本编码器组成。视觉编码器获取图像的特征,并将其转换为嵌入。文本...
示例: -曲调:Medical-Llama3-v2,sfr-iterative dpo-lalama-3-8b,Meta-llama-3.1-8b-指令-轻松 -投影机:llava-llama-3-8b-v1_1-mmproj、llava-lrama-3.1-8b-mmproj-f16、minicpm-V-2-5-mmproj ##理论 -CLIP模型通常由视觉编码器和文本编码器组成。视觉编码器获取图像的特征,并将其转换为嵌入。文本...
增加了中英双语的minicpm1b的qlora训练config 92d8107 增加了minicpm2b使用qlora训练dpo的config b91c19f 增加了minicpm2b在中文上全量训练的config fa38e32 增加了minicpm2b在中文下训练的loraconfig' 983f2d6 增加了中英双语下使用qlora训练minicpm2b的config 90f45bc ppp...
internlm2_chat_7b_dpo_qlora_varlenattn internlm2_chat_7b_full_finetune_custom_dataset_e1 internlm2_chat_7b_orpo_qlora_varlenattn_ultrafeedback_e5 internlm2_chat_7b_qlora_alpaca_e3 internlm2_chat_7b_qlora_code_alpaca_e3 internlm2_chat_7b_qlora_custom_sft_e1 ...
Added minicpm config file to support sft、qlora、lora、dpo by@LDLINGLINGLINGin#847 fix lint by@HIT-cwhin#856 bump version to 0.1.22 by@HIT-cwhin#855 New Contributors @Johnson-Wang @xu-songmade their first contribution in#787 @bychen7made their first contribution in#795 ...
- **\[2024/07\]** Support [DPO](https://github.com/InternLM/xtuner/tree/main/xtuner/configs/dpo), [ORPO](https://github.com/InternLM/xtuner/tree/main/xtuner/configs/orpo) and [Reward Model](https://github.com/InternLM/xtuner/tree/main/xtuner/configs/reward_model) training with pack...
[Feature] Support DPO, ORPO and Reward Model (InternLM#743) Jun 13, 2024 LICENSE [Docs] Improve Docs (InternLM#22) Aug 10, 2023 MANIFEST.in [Fix] Update setup.py and MANIFEST.in (InternLM#45) Aug 25, 2023 README.md Support internlm2.5 (InternLM#803) Jul 3, 2024 README_zh-CN...