[BUG] ImportError: cannot import name 'text_encoder_lora_state_dict' from 'diffusers.loaders' (/usr/local/lib/python3.10/dist-packages/diffusers/loaders/__init__.py) about autotrain-advanced

[VikJoJo]

+1 the problem is relevant

[hqnicolas]

@mahdifekih You can make it work just edditing the Colab file:

import os
os.environ["HF_TOKEN"] = str(huggingface_token)
os.environ["NGROK_AUTH_TOKEN"]` = str(ngrok_token)
os.environ["AUTOTRAIN_LOCAL"] = "1"
!pip install -U autotrain-advanced==0.7.5 > install_logs.txt
!autotrain app --share

Where

autotrain-advanced==0.7.5

[alexljenkins]

I downgraded to 0.7.2 and it worked, 0.7.5 did not.
!pip install -U autotrain-advanced==0.7.2 > install_logs.txt
I also tried upgrading the python version to 3.11 and the version conflict got resolved, but still had the text_encoder_lora_state_dict import error (so didn't work).

[abhishekkrthakur]

we updated diffusers yesterday and it could be related to that. will take a look! till then, you can downgrade autotrain

[abhishekkrthakur]

this issue is being looked into at this moment

[hqnicolas]

+1

[hqnicolas]

Cause:

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
tensorflow-metadata 1.14.0 requires protobuf<4.21,>=3.20.3, but you have protobuf 4.23.4 which is incompatible.

[hqnicolas]

i'm runing by this process:
https://www.youtube.com/watch?v=0ab2uOi2-3A
0.7.5 works fine here

[alexljenkins]

it actually has been working for me when I run the pip install command twice - the first time there's a conflict and the second it runs :/

[mahdifekih]

i'm runing by this process: https://www.youtube.com/watch?v=0ab2uOi2-3A 0.7.5 works fine here

I am going by the same process as @hqnicolas and it worked for me with autotrain-advanced==0.7.5
Bless you guys I am working on my thesis

[mahdifekih]

this issue is being looked into at this moment

Glad to hear that! Keep us posted please and thanks for the good work!

[abhishekkrthakur]

this is now fixed in version 0.7.11 and above. feel free to re-open if you still face any issue.