tied weights
#5
by
shiyu-wangbyte
- opened
The model weights are not tied. Please use the tie_weights method before using the infer_auto_device function.
Anyone know the meaning of this warning?
I also got the same error.
I have the same problem. I tried to do:
model = LlamaForCausalLM.from_pretrained(
model_path, torch_dtype=torch.float16, device_map='auto',
)
model.tie_weights()
But it did persist.