Patches: Make lazyload work on quantized

i wanna watch youtube while my model is loading without locking up my system >:(
2025-06-05 21:59:24 +02:00 · 2023-07-17 16:47:31 -05:00
parent e8d84bb787
commit 23b95343bd
2 changed files with 2 additions and 3 deletions
--- a/modeling/inference_models/generic_hf_torch/class.py
+++ b/modeling/inference_models/generic_hf_torch/class.py
@@ -78,7 +78,6 @@ class model_backend(HFTorchInferenceModel):
        }
        
        if self.use_4_bit:
-            self.lazy_load = False
            tf_kwargs.update({
                "quantization_config":BitsAndBytesConfig(
                    load_in_4bit=True,