Where is the modeling_deepseek.py file
#15
by
ccocks-deca
- opened
Isn't it supposed to have a modeling_deepseek.py file? I thought it was required for inference with AutoModel
We recommend using SGLang or vLLM for inference. Both these engines do not require the modeling_deepseek.py file that is provided with the original DeepSeek models.
That being said: as our Chimera models and the DeepSeek family of V3, V3-0324, R1, R1-0528, ... all share a common architecture, the original modeling_deepseek.py file can be used with our Chimera models, if needed.