diffusers

[gguf][torch.compile time] Convert to plain tensor earlier in dequantize_gguf_tensor

#13166

Merged