-
Web
Highlights
Pinned Loading
-
Kaggle_Workspace_FreeGPU
Kaggle_Workspace_FreeGPU PublicComfyUI (Flux2 GGUF) на бесплатных 2x Tesla T4 в Kaggle. Запуск в одну строку, самовосстановление окружения, мульти-GPU. Проект THE ANGEL AI.
Python 2
-
SageAttention-SM75-path
SageAttention-SM75-path PublicForked from thu-ml/SageAttention
Quantized Attention that achieves speedups of 2.1-3.1x and 2.7-5.1x compared to FlashAttention2 and xformers, respectively, without lossing end-to-end metrics across various models.
Python 1
-
ComfyUI-LTX2-MultiGPU
ComfyUI-LTX2-MultiGPU PublicHybrid Multi-GPU split loader for LTX 2.3 GGUF on 2x T4 (Kaggle). Replaces ComfyUI-MultiGPU DisTorch2 regex splitting with explicit block-aware sharding and forward-hook hidden-state transfer.
Python
If the problem persists, check the GitHub status page or contact support.