Hugging Face's logo

deepseek-ai
/
DeepSeek-V3

Safetensors
deepseek_v3
custom_code
fp8
Model card Files Community
58
DeepSeek-V3
  • 3 contributors
 History: 15 commits
cuichenx's picture
cuichenx
`aux_loss_alpha` should be 1e-4 instead of 1e-3?
99b7681 verified