Aryan Kumar 01be8e707a [misc]: serve the Regenerate-2K examples on TORCH_SDPA and drop narrating comments ...

FLASH_ATTN bypassed the cuDNN attention the refine stage selects; TORCH_SDPA keeps it.

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

..

benchmarks

[bugfix] Fix the CUDA device of Ray workers on multi-GPU nodes (#1932)

2026-10-07 15:00:23 -07:00

__init__.py

[Feature] Remove V1 folder (#642)

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

component_loader.py

[misc]: serve the Regenerate-2K examples on TORCH_SDPA and drop narrating comments

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

fsdp_load.py

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

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

h3_community_lora.py

[feat]: Community MiniMax-H3 LoRA loading with quantized side branch (#1973)

2026-10-19 14:29:23 -07:00

lora_patch.py

[feat] Add streaming and GPU-accelerated LoRA extraction (#1784)

2026-10-07 04:48:35 -07:00

text_encoder_quantization.py

[perf] Optimize MiniMax-H3 text encoder memory (#1732)

2026-08-21 14:57:06 -07:00

utils.py

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

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

weight_utils.py

[bugfix] Fix the CUDA device of Ray workers on multi-GPU nodes (#1932)

2026-10-07 15:00:23 -07:00