automaticcat
automaticcat
> @tikikun, isnt this it? https://github.com/pytorch/pytorch/blob/efc6e8457a221c6e70265fe895f8bc418d73aa0f/torch/_inductor/autotune_process.py#L508 this is a method from a class
But it seems that would make the reconstruction part not viable
Or should we have headless jan or jan cli?
hi @0xSage @louis-jan this issue can easily be mitigated by constrained the context len to be below the chosen context len to be able to send. On the llama cpp...
Research input: - Market has tendency to consolidate on GGUF or Huggingface config file -> model already has its own config - What we want in the description is not...
Should be able to extract chat template from gguf, not good ux
Any update on this feature everyone?