Fangzhou AiandCursor Agent 77e8b02c8d [ROCm][Perf] Add opt-in a4w4 (FP4 activation) MoE for DeepSeek V4.1 on AITER (#58819) ...

Signed-off-by: fai <fangzhouai@gmail.com>
Signed-off-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com>
Co-authored-by: Cursor Agent (Claude Sonnet 5) <cursoragent@cursor.com>

2026-09-30 15:07:44 +08:00

..

compressed_tensors

[Bugfix][Quantization] Stop sleep(level=2) from zeroing compressed-tensors KV scales (#57163)

2026-09-29 15:34:15 +00:00

inc

[MoE] Defer the TRTLLM-Gen top-k finalize on the modular path (#58635)

2026-09-25 15:38:29 +08:00

online

[ROCm][Perf] Add opt-in a4w4 (FP4 activation) MoE for DeepSeek V4.1 on AITER (#58819)

2026-09-30 15:07:44 +08:00

quark

[Bugfix][ROCm] AMD-Quark mixed-precision DeepSeek-V4.1 support (#57071)

2026-09-25 20:51:18 -05:00

turboquant

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

utils

[Bugfix][ROCm] AMD-Quark mixed-precision DeepSeek-V4.1 support (#57071)

2026-09-25 20:51:18 -05:00

__init__.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

auto_awq.py

[Chore] Use Transformers v5 names and drop redundant processor use_fast (#58550)

2026-09-24 14:14:10 +00:00

auto_gptq.py

[Quant] Use canonical N-first weight format for CT WNA16 MoE (#52798)

2026-09-28 11:22:44 -06:00

awq_triton.py

Convert formatting to use ruff instead of yapf + isort (#26247)

2025-10-05 07:06:22 -07:00

base_config.py

[Chore] Use Transformers v5 names and drop redundant processor use_fast (#58550)

2026-09-24 14:14:10 +00:00

experts_int8.py

[Quantization] Support online quantization with partially pre-quantized checkpoints (#51392)

2026-09-04 11:46:06 -07:00

fbgemm_fp8.py

[ROCm][FEAT] Integrate aiter gemm w8a8 ptpc (#33773)

2026-04-16 09:55:29 +08:00

fp8.py

[MoE] Defer the TRTLLM-Gen top-k finalize on the modular path (#58635)

2026-09-25 15:38:29 +08:00

fp_quant.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

humming.py

[Feature][Humming] Humming feature integration (#56685)

2026-09-20 08:11:01 -07:00

input_quant_fp8.py

[Bugfix][ROCm] Dispatch the QuantFP8 CUDA fallback on the class (#58136)

2026-09-22 17:49:11 +00:00

kv_cache.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

modelopt.py

[MoE] Defer the TRTLLM-Gen top-k finalize on the modular path (#58635)

2026-09-25 15:38:29 +08:00

moe_wna16.py

[Bugfix] Fix moe_wna16 w13 zero-point shard split for 8-bit asym GPTQ MoE (#58950)

2026-09-28 23:38:28 +00:00

mxfp4.py

[ROCm][Perf] Add opt-in a4w4 (FP4 activation) MoE for DeepSeek V4.1 on AITER (#58819)

2026-09-30 15:07:44 +08:00

qutlass_utils.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00

torchao.py

Add pydocstyle to the ruff rules (#52136)

2026-09-16 16:44:29 +00:00