File size: 126 Bytes
5fa1a76
1
If the model was quantized with the device_map parameter, make sure to move the entire model to a GPU or CPU before saving it.