Tyler Michael Smith d7a299edaa [Kernel] Remove scaled_fp8_quant kernel padding footgun (#6842)

2024-07-30 16:37:01 -04:00

..

compressed_tensors

[Bugfix] Fix empty (nullptr) channelwise scales when loading wNa16 using compressed tensors (#6798)

2024-07-25 15:05:09 -07:00

utils

[Kernel] Remove scaled_fp8_quant kernel padding footgun (#6842)

2024-07-30 16:37:01 -04:00

__init__.py

[Kernel][Core] Add AWQ support to the Marlin kernel (#6612)

2024-07-21 19:41:42 -04:00

aqlm.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

awq_marlin.py

[Bugfix] Fix awq_marlin and gptq_marlin flags (#6745)

2024-07-24 22:34:11 -07:00

awq.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

base_config.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

bitsandbytes.py

[bitsandbytes]: support read bnb pre-quantized model (#5753)

2024-07-23 23:45:09 +00:00

deepspeedfp.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

fbgemm_fp8.py

[Misc] Pass cutlass_fp8_supported correctly in fbgemm_fp8 (#6871)

2024-07-28 11:13:49 -04:00

fp8.py

[Bugfix] Allow vllm to still work if triton is not installed. (#6786)

2024-07-29 14:51:27 -07:00

gptq_marlin_24.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

gptq_marlin.py

[Bugfix] Fix awq_marlin and gptq_marlin flags (#6745)

2024-07-24 22:34:11 -07:00

gptq.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

kv_cache.py

[Bugfix] Fix kv_cache_dtype=fp8 without scales for FP8 checkpoints (#6761)

2024-07-25 09:46:15 -07:00

marlin.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00

schema.py

Enable scaled FP8 (e4m3fn) KV cache on ROCm (AMD GPU) (#3290)

2024-04-03 14:15:55 -07:00

squeezellm.py

[ Misc ] fbgemm checkpoints (#6559)

2024-07-20 09:36:57 -07:00