microsoft / Tutel
Tutel MoE: Optimized Mixture-of-Experts Library, Support GLM/Kimi/Gpt/DeepSeek/Qwen using FP8/NVFP4/MXFP4
- Repository stars
- 1,029All-time #39,660
- Fork
- 111
- Last push
- updated 2 days ago
- Repository created
- since 2021
About
Tutel MoE: Optimized Mixture-of-Experts Library, Support GLM/Kimi/Gpt/DeepSeek/Qwen using FP8/NVFP4/MXFP4Summarised by AI from the README; the original prevails
Author description: Tutel MoE: Optimized Mixture-of-Experts Library, Support GLM/Kimi/Gpt/DeepSeek/Qwen using FP8/NVFP4/MXFP4
Basic listing based on official public information; functionality has not been independently evaluated.
- Pricing
- Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.
