feat: add initial AutoGPTQ backend implementation

This commit is contained in:
Ettore Di Giacinto 2023-08-07 22:39:10 +02:00
parent 91d49cfe9f
commit a843e64fc2
37 changed files with 660 additions and 148 deletions

View file

@ -71,7 +71,7 @@ func (llm *LLM) Load(opts *pb.ModelOptions) error {
llamaOpts = append(llamaOpts, llama.EnabelLowVRAM)
}
model, err := llama.New(opts.Model, llamaOpts...)
model, err := llama.New(opts.ModelFile, llamaOpts...)
llm.llama = model
return err
}