deepseek-ai

/

DeepSeek-Coder-V2-Lite-Instruct

like 280

For some reason this will not run for me using accelerate, no matter the config transformers will only allocate the model to a single GPU.

Is there a model specific reason for this?