2024 Release #96

danielhanchen · 2024-01-18T17:51:00Z

No description provided.

commit 35f2ab4a8b4deecbbbe9fbd95f4efde8694233db Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Feb 4 17:35:56 2024 +1100 2x faster inference (#151) * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py * Fast inference repatch * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update mistral.py * Update __init__.py * Fix inference * Update mistral.py * fast lm_head * Remove fast path * Update rope_embedding.py * Update loader.py * LlamaAttention_fast_forward_inference * if past_key_value is not None and q_len == 1: * revert inference * Update loader.py * past_key_value * Update llama.py * Update llama.py * Fix SDPA * Update llama.py * padding * Inference * Update llama.py * Revert * Update mistral.py * faster inference * inference * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * inference * Update llama.py * Update utils.py * faster inference * Update llama.py * revert * lm_head * Update llama.py * inference * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * faster inference * Update llama.py * fast inference * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * torch compile * past_key_values * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update utils.py * Update llama.py * fast inference + saving config.json * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * fast inference again * more temp matrices * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update mistral.py * Update llama.py * SDPA * attention_mask * New version * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update utils.py * Update utils.py commit 051a73b0e63d3ae3acd7c4d962349280f69bbdb0 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 31 04:03:37 2024 +1100 Hotfix - fix inference (#146) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py * Fast inference repatch * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update mistral.py * Update __init__.py * Fix inference * Update mistral.py * fast lm_head * Remove fast path * Update rope_embedding.py * Update loader.py * LlamaAttention_fast_forward_inference * if past_key_value is not None and q_len == 1: * revert inference * Update loader.py * past_key_value commit 05624642802c7f90dcc7aeea0e1c8d447cde006e Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 17:49:54 2024 +1100 Fix inference attention mask (#142) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py commit 206a9b65f090bd71ccaad7dd88b67ba2bfde0b58 Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 03:45:07 2024 +1100 Nightly (#140) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving commit 8faf469f028a05852b2dc29ec8df1f36998fab33 Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 02:52:39 2024 +1100 Fix saving issues (#139) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print commit 1ecc0185a5759c7a0c95dfc96aceea5023cebdfc Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 28 04:30:29 2024 +1100 1 more bug (#138) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py commit cd32ba76b71adf3317ede9de7d1cf6f30ad3bf0d Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 28 04:20:06 2024 +1100 Fix bugs + more accurate Swiglu (#137) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask commit 89daa0efcc38c7690abbb8170b5d9f3d364796ce Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 27 04:50:22 2024 +1100 Inference bug fix (#134) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py commit 87a7ef1049f6fca409a0673f51f4758e0aff248d Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 27 04:47:54 2024 +1100 More bug fixes (#133) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py commit 3d67790901696e953171f64b4bf9d980780051a0 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 26 04:19:17 2024 +1100 Fix bugs (#129) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py commit a833f403462e9cfc1f96b3b84d9da15d7d8db5ee Author: Daniel Han <danielhanchen@gmail.com> Date: Tue Jan 23 03:55:24 2024 +1100 2-4x faster native HF inference (#119) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving commit b370c9c8aacc31a7845404566dd95dfa8c0e3bac Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 21 22:20:22 2024 +1100 Hotfix (#118) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py commit 57a5b5a49da588b1db8e9a988cc985dc20393d34 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 05:00:37 2024 +1100 Update save.py commit 5145a61e69ab9b3035465f649e1c1e5aae749f8f Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 04:21:54 2024 +1100 Update save.py commit a7bd8d119c16433de4f8b6a36903ef7131f225e5 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 04:13:03 2024 +1100 Update save.py commit be4b97e7d89074b6dd1d2e984fa429051d328192 Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 21 03:43:49 2024 +1100 Fixed saving! (#113) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF * Fix quantization_method * Fix quantization_config * patch model * Update llama.py * Update llama.py * Update llama.py * Update save.py * Update save.py * tokenizer_save_settings * Update save.py * quantization and loftq * Update save.py * Update llama.py * Update save.py * upload_to_huggingface * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py commit abb462be71e8cf01ad989dca0efaa17441113651 Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 20 23:23:00 2024 +1100 Hotfix for Jan 2024 Release (#110) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF * Fix quantization_method * Fix quantization_config * patch model * Update llama.py * Update llama.py * Update llama.py * Update save.py * Update save.py * tokenizer_save_settings * Update save.py * quantization and loftq * Update save.py * Update llama.py * Update save.py commit 31e2d71720e64b854145d7779833b7d2d3d4177e Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 20 04:25:06 2024 +1100 Quick fixes (#106) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF commit 8846337e5c8c2f206a4ac8fe6d239f3d1221f7ac Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sat Jan 20 02:30:31 2024 +1100 Update _utils.py commit d378df87e5f3945474915a098c9aa58313465064 Merge: c1e7480 920e3c2 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 23:15:38 2024 +1100 Merge branch 'main' of https://github.com/unslothai/unsloth commit c1e7480ac2ad0e5efa05e84fe0997619ccdd86a4 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 23:15:20 2024 +1100 Revert quantization methods commit 920e3c2ea07a044addeb7c3fa8be6f0189cb7f84 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 22:57:22 2024 +1100 getattr issues (#103) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr commit fc25ab0df032f8ee5ea750f27c68d63f49d2d9a9 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 22:52:30 2024 +1100 Quick fixes (#101) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py commit b8b1eafda35d124046e11766aeeb6343957e0daf Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 04:51:19 2024 +1100 2024 Release (#96) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml commit 4112eb4a3df4c0911e36211b47381086c963b4e0 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 03:41:00 2024 +1100 Update pyproject.toml commit 59d74753362ff59e664cb6d650b564511e6e20f3 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 03:35:17 2024 +1100 Update pyproject.toml commit c1ac4d2707574868767345e76ebe49c8353f9057 Author: Daniel Han <danielhanchen@gmail.com> Date: Thu Jan 11 04:08:03 2024 +1100 Fix some bugs (#83) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs commit d3887c7fd93d9b910bf6ee3ab3c7fd485fc55e46 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 10 23:10:48 2024 +1100 Update README.md (#81) commit b5d94d9a0ad9532494e1b3c7badbb94fa92c50eb Author: shimmy <107991372+shimmyshimmer@users.noreply.github.com> Date: Wed Jan 10 23:10:23 2024 +1100 Discord button redo (#80) commit 01d7f58e11373ab07b9282a42bc14f542dbdabf0 Author: shimmy <107991372+shimmyshimmer@users.noreply.github.com> Date: Wed Jan 10 23:02:20 2024 +1100 Update logos (#79) * HF Perf Button * Update README.md Adding new buttons cleanup * Update README.md * Delete images/Discord.png * Delete images/try live demo green.png * new transparent logos * Revamping page * Revamp mainpage * Update README.md * Update README.md commit 9faaf5b388e025f8ffc302450a12ffb84e7e1750 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 10 20:03:01 2024 +1100 Create FUNDING.yml (#78) commit 82e6fece0b78011707090639823d2d7acf5a3864 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Wed Jan 10 01:02:44 2024 +1100 fix_tokenizer commit b52278199b7ae2764f242622275bb8a85ba7b721 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Tue Jan 9 23:40:43 2024 +1100 check_tokenizer

* HF Perf Button * Update README.md Adding new buttons cleanup * Update README.md * Delete images/Discord.png * Delete images/try live demo green.png * new transparent logos * Revamping page * Revamp mainpage * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * finetune button * Delete start free finetune button.png * free finetune button * Add files via upload * Update README.md * Update README.md * Add files via upload * Add files via upload * Update README.md * Add files via upload * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Update README.md * Squashed commit of the following: commit 35f2ab4a8b4deecbbbe9fbd95f4efde8694233db Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Feb 4 17:35:56 2024 +1100 2x faster inference (#151) * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py * Fast inference repatch * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update mistral.py * Update __init__.py * Fix inference * Update mistral.py * fast lm_head * Remove fast path * Update rope_embedding.py * Update loader.py * LlamaAttention_fast_forward_inference * if past_key_value is not None and q_len == 1: * revert inference * Update loader.py * past_key_value * Update llama.py * Update llama.py * Fix SDPA * Update llama.py * padding * Inference * Update llama.py * Revert * Update mistral.py * faster inference * inference * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * inference * Update llama.py * Update utils.py * faster inference * Update llama.py * revert * lm_head * Update llama.py * inference * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * faster inference * Update llama.py * fast inference * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * torch compile * past_key_values * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update utils.py * Update llama.py * fast inference + saving config.json * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * fast inference again * more temp matrices * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update mistral.py * Update llama.py * SDPA * attention_mask * New version * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update utils.py * Update utils.py commit 051a73b0e63d3ae3acd7c4d962349280f69bbdb0 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 31 04:03:37 2024 +1100 Hotfix - fix inference (#146) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py * Fast inference repatch * Update llama.py * Update utils.py * Update utils.py * Update utils.py * Update mistral.py * Update __init__.py * Fix inference * Update mistral.py * fast lm_head * Remove fast path * Update rope_embedding.py * Update loader.py * LlamaAttention_fast_forward_inference * if past_key_value is not None and q_len == 1: * revert inference * Update loader.py * past_key_value commit 05624642802c7f90dcc7aeea0e1c8d447cde006e Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 17:49:54 2024 +1100 Fix inference attention mask (#142) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving * Update llama.py * Update llama.py commit 206a9b65f090bd71ccaad7dd88b67ba2bfde0b58 Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 03:45:07 2024 +1100 Nightly (#140) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print * Mistral patch * Update mistral.py * Update save.py * saving commit 8faf469f028a05852b2dc29ec8df1f36998fab33 Author: Daniel Han <danielhanchen@gmail.com> Date: Mon Jan 29 02:52:39 2024 +1100 Fix saving issues (#139) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py * Update mistral.py * attention mask * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update dpo.py * Patch saving * Update save.py * Update save.py * patch_saving_functions * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * print commit 1ecc0185a5759c7a0c95dfc96aceea5023cebdfc Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 28 04:30:29 2024 +1100 1 more bug (#138) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask * Update save.py * Update save.py commit cd32ba76b71adf3317ede9de7d1cf6f30ad3bf0d Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 28 04:20:06 2024 +1100 Fix bugs + more accurate Swiglu (#137) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py * Works? * Update pyproject.toml * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Swiglu * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * attention_mask * Update llama.py * Update llama.py * labels * Update mistral.py * Update llama.py * attention mask commit 89daa0efcc38c7690abbb8170b5d9f3d364796ce Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 27 04:50:22 2024 +1100 Inference bug fix (#134) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Revert "Update llama.py" This reverts commit a208ec46e012cf470ecefe6268a66358215df7b6. * Update llama.py commit 87a7ef1049f6fca409a0673f51f4758e0aff248d Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 27 04:47:54 2024 +1100 More bug fixes (#133) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update llama.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update swiglu.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update fast_lora.py * Update save.py * Update fast_lora.py * Update utils.py * Update llama.py * Update fast_lora.py * Update swiglu.py * Update save.py * Update save.py * Update llama.py * Update llama.py * Update llama.py commit 3d67790901696e953171f64b4bf9d980780051a0 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 26 04:19:17 2024 +1100 Fix bugs (#129) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving * Update llama.py * hidden_states * q_len == 1 * q_len issue * Update mistral.py * Update mistral.py * incorrect inference * Update to transformers 4.37 * Graceful FA2 error + torch 2.1.1 * Update mapper.py * Update pyproject.toml * Fix saving and bnb-4bit * Update fast_lora.py * Update fast_lora.py * remove patching * Update llama.py * Update llama.py * Update swiglu.py * Repatch * Update fast_lora.py commit a833f403462e9cfc1f96b3b84d9da15d7d8db5ee Author: Daniel Han <danielhanchen@gmail.com> Date: Tue Jan 23 03:55:24 2024 +1100 2-4x faster native HF inference (#119) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * fast inference * Update llama.py * Update save.py * Update llama.py * Mistral correct RoPE scaling * Max sequence lengths * Apache 2 * fast_linear_forward * Update utils.py * Update utils.py * No print * Update utils.py * Update utils.py * inference * Update llama.py * Fast inference RoPE * Update llama.py * Update llama.py * RoPE * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * LoRA * Fast LoRA saving commit b370c9c8aacc31a7845404566dd95dfa8c0e3bac Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 21 22:20:22 2024 +1100 Hotfix (#118) * faster saving & inference * Update llama.py * Update save.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update mistral.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py * Update llama.py commit 57a5b5a49da588b1db8e9a988cc985dc20393d34 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 05:00:37 2024 +1100 Update save.py commit 5145a61e69ab9b3035465f649e1c1e5aae749f8f Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 04:21:54 2024 +1100 Update save.py commit a7bd8d119c16433de4f8b6a36903ef7131f225e5 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sun Jan 21 04:13:03 2024 +1100 Update save.py commit be4b97e7d89074b6dd1d2e984fa429051d328192 Author: Daniel Han <danielhanchen@gmail.com> Date: Sun Jan 21 03:43:49 2024 +1100 Fixed saving! (#113) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF * Fix quantization_method * Fix quantization_config * patch model * Update llama.py * Update llama.py * Update llama.py * Update save.py * Update save.py * tokenizer_save_settings * Update save.py * quantization and loftq * Update save.py * Update llama.py * Update save.py * upload_to_huggingface * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py commit abb462be71e8cf01ad989dca0efaa17441113651 Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 20 23:23:00 2024 +1100 Hotfix for Jan 2024 Release (#110) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF * Fix quantization_method * Fix quantization_config * patch model * Update llama.py * Update llama.py * Update llama.py * Update save.py * Update save.py * tokenizer_save_settings * Update save.py * quantization and loftq * Update save.py * Update llama.py * Update save.py commit 31e2d71720e64b854145d7779833b7d2d3d4177e Author: Daniel Han <danielhanchen@gmail.com> Date: Sat Jan 20 04:25:06 2024 +1100 Quick fixes (#106) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr * RSLoRA and LoftQ direct support * Update llama.py * Update llama.py * Update llama.py * Fix DPO + GGUF commit 8846337e5c8c2f206a4ac8fe6d239f3d1221f7ac Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Sat Jan 20 02:30:31 2024 +1100 Update _utils.py commit d378df87e5f3945474915a098c9aa58313465064 Merge: c1e7480 920e3c2 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 23:15:38 2024 +1100 Merge branch 'main' of https://github.com/unslothai/unsloth commit c1e7480ac2ad0e5efa05e84fe0997619ccdd86a4 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 23:15:20 2024 +1100 Revert quantization methods commit 920e3c2ea07a044addeb7c3fa8be6f0189cb7f84 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 22:57:22 2024 +1100 getattr issues (#103) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py * getattr commit fc25ab0df032f8ee5ea750f27c68d63f49d2d9a9 Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 22:52:30 2024 +1100 Quick fixes (#101) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml * Quick fixes * Update llama.py * Update llama.py * Update dpo.py * Update dpo.py * Update llama.py * Update save.py commit b8b1eafda35d124046e11766aeeb6343957e0daf Author: Daniel Han <danielhanchen@gmail.com> Date: Fri Jan 19 04:51:19 2024 +1100 2024 Release (#96) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs * Faster saving + other changes * Update llama.py * Saving modules * spelling * Update llama.py * Update save.py * Update save.py * Update loader.py * Update llama.py * patch saving * Update save.py * Update save.py * Update save.py * patch saving * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * original_model * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * Update save.py * saving to RAM leakage? * Update save.py * new_save_directory * Update save.py * Update save.py * Update save.py * Update save.py * Update pyproject.toml * Update pyproject.toml * Update pyproject.toml commit 4112eb4a3df4c0911e36211b47381086c963b4e0 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 03:41:00 2024 +1100 Update pyproject.toml commit 59d74753362ff59e664cb6d650b564511e6e20f3 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Fri Jan 19 03:35:17 2024 +1100 Update pyproject.toml commit c1ac4d2707574868767345e76ebe49c8353f9057 Author: Daniel Han <danielhanchen@gmail.com> Date: Thu Jan 11 04:08:03 2024 +1100 Fix some bugs (#83) * Fix tokenizer, dropout, bias for LoRA * Update loader.py * Fix LoRA downcasting * Update _utils.py * Saving to GGUF * fix * colab_quantize_to_gguf * move save modules * save module * Update __init__.py * Update save.py * Temp downgrade due to TRL issue * Fix up bugs commit d3887c7fd93d9b910bf6ee3ab3c7fd485fc55e46 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 10 23:10:48 2024 +1100 Update README.md (#81) commit b5d94d9a0ad9532494e1b3c7badbb94fa92c50eb Author: shimmy <107991372+shimmyshimmer@users.noreply.github.com> Date: Wed Jan 10 23:10:23 2024 +1100 Discord button redo (#80) commit 01d7f58e11373ab07b9282a42bc14f542dbdabf0 Author: shimmy <107991372+shimmyshimmer@users.noreply.github.com> Date: Wed Jan 10 23:02:20 2024 +1100 Update logos (#79) * HF Perf Button * Update README.md Adding new buttons cleanup * Update README.md * Delete images/Discord.png * Delete images/try live demo green.png * new transparent logos * Revamping page * Revamp mainpage * Update README.md * Update README.md commit 9faaf5b388e025f8ffc302450a12ffb84e7e1750 Author: Daniel Han <danielhanchen@gmail.com> Date: Wed Jan 10 20:03:01 2024 +1100 Create FUNDING.yml (#78) commit 82e6fece0b78011707090639823d2d7acf5a3864 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Wed Jan 10 01:02:44 2024 +1100 fix_tokenizer commit b52278199b7ae2764f242622275bb8a85ba7b721 Author: Daniel Han-Chen <danielhanchen@gmail.com> Date: Tue Jan 9 23:40:43 2024 +1100 check_tokenizer --------- Co-authored-by: Daniel Han <danielhanchen@gmail.com>

danielhanchen added 30 commits January 6, 2024 17:59

Fix tokenizer, dropout, bias for LoRA

1543548

Update loader.py

14bee60

Merge branch 'main' into nightly

82db08e

Fix LoRA downcasting

d78643d

Update _utils.py

b6a9841

Merge branch 'main' into nightly

1d0e1dc

Saving to GGUF

3c41946

fix

25b15bc

colab_quantize_to_gguf

4bd8f7b

move save modules

b3e0e0f

save module

bb93e31

Update __init__.py

29aaf66

Update save.py

605290f

Merge branch 'main' into nightly

4937926

Temp downgrade due to TRL issue

f342605

Merge branch 'main' into nightly

b332a8c

Merge branch 'main' into nightly

5dca87c

Fix up bugs

38f7c04

Merge branch 'main' into nightly

74037bb

Faster saving + other changes

7af483c

Update llama.py

5ad792a

Saving modules

2fc754f

spelling

48461a6

Update llama.py

ed3e1db

Update save.py

07202ff

Update save.py

7c1c87f

Update loader.py

44430d1

Update llama.py

2ceed5d

patch saving

e932893

Update save.py

08227b1

danielhanchen added 26 commits January 18, 2024 17:44

Update save.py

f2884c1

Update save.py

cc0ff3f

Update save.py

c549550

Update save.py

f755247

Update save.py

2fced37

Update save.py

ee7154f

Update save.py

5aaffb7

Update save.py

61e4c1c

Update save.py

9f11efb

Update save.py

43a025a

Update save.py

b90c385

Update save.py

573427b

Update save.py

2672abf

Update save.py

68670c6

Update save.py

2231bdd

Update save.py

074e79a

saving to RAM leakage?

90a2f9b

Update save.py

8e140af

new_save_directory

d3b4a24

Update save.py

f535904

Update save.py

386655e

Update save.py

e5e91f4

Update save.py

4a91ede

Update pyproject.toml

c53dee9

Update pyproject.toml

7196ef0

Update pyproject.toml

b2cebb8

danielhanchen merged commit b8b1eaf into main Jan 18, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

2024 Release #96

2024 Release #96

danielhanchen commented Jan 18, 2024

2024 Release #96

2024 Release #96

Conversation

danielhanchen commented Jan 18, 2024