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

@ -40,7 +40,7 @@ func (llm *LLM) Load(opts *pb.ModelOptions) error {
ggllmOpts = append(ggllmOpts, ggllm.SetNBatch(512))
}
model, err := ggllm.New(opts.Model, ggllmOpts...)
model, err := ggllm.New(opts.ModelFile, ggllmOpts...)
llm.falcon = model
return err
}