Aryan KumarandAryan Kumar d954fad9ec [perf] MiniMax H3: faster FP4 attention padding, uint8 VAE pixel path and cuDNN FP4 GEMMs (#2046) ...

Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>

2026-11-01 22:23:10 -08:00

..

__init__.py

[docs]: modernize documentation build (#1503)

2026-06-28 02:02:09 -07:00

autoencoder_kl.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

common.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

cosmos25wanvae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

flux2_components.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

flux2vae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

gen3c_tokenizer_vae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

hunyuan15vae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

hunyuanvae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

hyworldvae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

kandinsky6_sr.py

[feat]: add Kandinsky-6.0 TI2VA and VSR pipelines (#1924)

2026-10-06 10:50:17 -07:00

lingbotworld2_wanvae.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

ltx2vae.py

[new-model] Add LTX-2.5 inference support (#1704)

2026-10-07 03:34:18 -07:00

minimax_h3_audio.py

[perf] Optimize MiniMax H3 VAE decoding (#1734)

2026-08-21 14:57:32 -07:00

minimax_h3_int8_convrot.py

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

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

minimax_h3_int8_kernels.py

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

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

minimax_h3_nvfp4_checkpoint.py

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

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

minimax_h3_nvfp4_decoder.py

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

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

minimax_h3_nvfp4_fused.py

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

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

minimax_h3_parallel.py

[perf] MiniMax H3: decode eight VAE spatial tiles per call by default (#2040)

2026-11-01 22:06:23 -08:00

minimax_h3_qk_norm_rope.py

[perf] MiniMax H3: fuse VAE decoder QK RMSNorm and partial RoPE (#2041)

2026-11-01 22:15:35 -08:00

minimax_h3_taeh3.py

[feat]: add opt-in CUDA TAEH3 preview decode for FastH3 (#1795)

2026-08-31 23:34:56 -07:00

minimax_h3_video.py

[perf] MiniMax H3: faster FP4 attention padding, uint8 VAE pixel path and cuDNN FP4 GEMMs (#2046)

2026-11-01 22:23:10 -08:00

oobleck.py

[bugfix] Fix environment-variable bugs in LTX-2 debug logging, HF token lookup, and attention backend reads (#1895)

2026-09-29 11:48:31 -07:00

qwen_image21.py

[new-model] Add initial Qwen-Image-2.1 inference (#1918)

2026-10-25 17:28:49 +00:00

sa_audio.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

wanvae.py

[refactor] Move Wan VAE into the Wan package (#1824)

2026-09-05 18:43:55 -07:00