ValueError: Key down_blocks.0.attentions.0.transformer_blocks.0.attn1.processor is invalid, expected torch.Tensor but received <class 'dict'>

[Sandeep-Narahari]

Describe the bug

After completing the training I got this error while training custom diffusion

Reproduction

!accelerate launch train_custom_diffusion.py \ --pretrained_model_name_or_path="CompVis/stable-diffusion-v1-4" \ --output_dir=/content/output \ --concepts_list=/content/meta.json \ --with_prior_preservation --real_prior --prior_loss_weight=1.0 \ --resolution=512 \ --train_batch_size=1 \ --learning_rate=1e-5 \ --lr_warmup_steps=0 \ --max_train_steps=500 \ --num_class_images=200 \ --scale_lr --hflip \ --enable_xformers_memory_efficient_attention \ --modifier_token "<new1>+<new2>"

Logs

No response

System Info

Google collab
Python 3.10.12
diffusers Version: 0.21.0.dev0
clip-retrieval Version: 2.37.0

Who can help?

@sayakpaul @yiyixuxu @willi

[sayakpaul]

Could you provide the Colab Notebook you're using for this?

[canberk17]

@sayakpaul, I believe this is related to the issue discussed here: #4704 .@Sandeep-Narahari, could you try adding
--no_safe_serialization to your arguments and see if that'll do the trick?

[Sandeep-Narahari]

Yeah correct got it , working now , Thank you