Aryan Kumar 32c7a18fb1 [bugfix]: merge a wider-than-base LoRA at its own precision under FSDP too ...

The FSDP (DTensor) merge path still cast the adapter to the BF16 base before multiplying, so an FP32 adapter merged
differently with FSDP on. Both paths now share one helper, which also moves the factors to the weight's device.

2026-10-30 11:35:48 -07:00

..

lora

[bugfix]: merge a wider-than-base LoRA at its own precision under FSDP too

2026-10-30 11:35:48 -07:00

quantization

[feat]: FastH3 OmniRef inference on consumer devices (RTX 5090 / PRO 6000 / DGX Spark / 4090) (#1980)

2026-10-19 14:57:36 -07:00

__init__.py

[Feature] Remove V1 folder (#642)

2025-07-24 22:43:12 -07:00

activation.py

[feat] pre-commit support 120 col num (#1167)

2026-03-11 20:19:30 -07:00

AGENTS.md

[docs] add hierarchical AGENTS.md per-directory guidance (#1278)

2026-05-02 03:28:18 -07:00

custom_op.py

[Feature] Remove V1 folder (#642)

2025-07-24 22:43:12 -07:00

fp4linear.py

[feat]: FastH3 OmniRef inference on consumer devices (RTX 5090 / PRO 6000 / DGX Spark / 4090) (#1980)

2026-10-19 14:57:36 -07:00

fp8linear.py

[bugfix] FP8: gate the ROCm _scaled_mm path on CDNA4 (gfx950) instead of the capability tuple (#1859)

2026-09-16 19:58:55 -07:00

layernorm.py

[perf] Wan kernel fusion: Triton-fused residual+LayerNorm+modulate inference path (#1708)

2026-10-07 04:08:07 -07:00

linear.py

[refactor]: linear/mlp FP4 path additions for Wan-2.1 (Attn-QAT 6/12) (#1390)

2026-06-08 14:51:57 -07:00

mlp.py

[refactor]: linear/mlp FP4 path additions for Wan-2.1 (Attn-QAT 6/12) (#1390)

2026-06-08 14:51:57 -07:00

mxfp8linear.py

[bugfix]: requantize MXFP8 weights after LoRA unmerge (#1876)

2026-10-07 01:38:02 -07:00

pdd.py

[feat] Add H3 PDD/DMPDD training and packed Ref2VA preprocessing (#1984)

2026-10-20 11:15:57 -07:00

rotary_embedding_3d.py

[feat] pre-commit support 120 col num (#1167)

2026-03-11 20:19:30 -07:00

rotary_embedding.py

[misc] Remove DreamX World, HunyuanGameCraft, GLM-Image, LongCat, TurboDiffusion, and Z-Image (#1926)

2026-10-14 23:16:32 -07:00

triton_fused_norm.py

[bugfix] Repair DreamVerse and failing CI gates (#1958)

2026-10-16 17:15:38 -07:00

utils.py

[feat] pre-commit support 120 col num (#1167)

2026-03-11 20:19:30 -07:00

visual_embedding.py

[feat] pre-commit support 120 col num (#1167)

2026-03-11 20:19:30 -07:00

vocab_parallel_embedding.py

[feat] pre-commit support 120 col num (#1167)

2026-03-11 20:19:30 -07:00