mirror of
https://github.com/KoboldAI/KoboldAI-Client.git
synced 2025-06-05 21:59:24 +02:00
4-bit on Colab
This commit is contained in:
@@ -82,7 +82,7 @@ class model_backend(HFTorchInferenceModel):
|
||||
"low_cpu_mem_usage": True,
|
||||
}
|
||||
|
||||
if self.use_4_bit:
|
||||
if self.use_4_bit or utils.koboldai_vars.colab_arg:
|
||||
tf_kwargs.update({
|
||||
"quantization_config":BitsAndBytesConfig(
|
||||
load_in_4bit=True,
|
||||
|
Reference in New Issue
Block a user