Roland Tannous 1cefffa2d2 GRPO Fix - Support vllm pre-dequantized quantization states in fast_dequantize kernel (#2943) ... * Support pre-dequantized quantization states in fast_dequantize kernel * has_nested_quant conditional set to only * Update utils.py * Update utils.py --------- Co-authored-by: Daniel Han <danielhanchen@gmail.com>		2025-07-14 02:41:15 -07:00
..
dataprep	Convert torch.bfloat16, torch.float16, etc. to vLLM valid dtypes (#2811)	2025-06-30 16:39:36 -07:00
kernels	GRPO Fix - Support vllm pre-dequantized quantization states in fast_dequantize kernel (#2943)	2025-07-14 02:41:15 -07:00
models	fix dataloader_num_workers value error in GRPOTrainer (#2944)	2025-07-14 01:43:33 -07:00
registry	Qwen 3, Bug Fixes (#2445)	2025-04-30 22:38:39 -07:00
utils	Qwen 3, Bug Fixes (#2445)	2025-04-30 22:38:39 -07:00
__init__.py	Many bug fixes (#2927)	2025-07-10 07:03:48 -07:00
_auto_install.py	Flash Attention whls	2025-05-26 22:48:46 -07:00
chat_templates.py	fix issue with qwen3 template double quote escapes	2025-05-16 23:26:03 -06:00
save.py	reroute merge logic language models + comprehensive tests + eval kits (#2673)	2025-06-02 20:32:57 -07:00
tokenizer_utils.py	Many bug fixes (#2087)	2025-03-18 05:15:57 -07:00
trainer.py	Initialize parent class in UnslothTrainingArguments constructor	2025-07-03 15:47:58 +00:00