File "/usr/local/python/lib/python3.8/site-packages/diffusers/models/attention_processor.py", line 3074, in __call__ [rank0]: batch_size, key_tokens, _ = ( [rank0]: ValueError: too many values to unpack (expected 3)

[laoniandisko]

Describe the bug

添加完xformer加速之后出现了这个问题

Reproduction

def __init__():

    self.unet = pipeline.unet
    self.set_diffusers_xformers_flag(self.unet,True)



def set_diffusers_xformers_flag( model, valid):

    def fn_recursive_set_mem_eff(module: torch.nn.Module):
        if hasattr(module, "set_use_memory_efficient_attention_xformers"):
            module.set_use_memory_efficient_attention_xformers(valid)
            print("="*100)
            print(hasattr(module, "set_use_memory_efficient_attention_xformers"))

        for child in module.children():
            fn_recursive_set_mem_eff(child)
    fn_recursive_set_mem_eff(model)

 def forward():
    self.attn1(
            mda_norm_hidden_states,
            encoder_hidden_states=encoder_hidden_states if self.only_cross_attention else None,
            attention_mask=attention_mask,
            **cross_attention_kwargs,
        )

Logs

ps：The rest of the code is not convenient to provide

System Info

• 🤗 Diffusers version: 0.33.1
• Platform: Linux-5.4.241-1-tlinux4-0017.7-x86_64-with-glibc2.2.5
• Running on Google Colab?: No
• Python version: 3.8.12
• PyTorch version (GPU?): 2.4.0+cu118 (True)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Huggingface_hub version: 0.30.2
• Transformers version: 4.46.3
• Accelerate version: 1.0.1
• PEFT version: 0.13.2
• Bitsandbytes version: 0.42.0
• Safetensors version: 0.4.3
• xFormers version: 0.0.27.post2+cu118
• Accelerator: NVIDIA H20, 97871 MiB

Who can help?

No response