Closed Duemellon closed 5 months ago
May I ask what method you used for downloaded and which t5 model link?
The model folders provided in our link are as follows:
I used this link: https://huggingface.co/QQGYLab/ELLA/tree/main/models--google--flan-t5-xl--text_encoder
I clicked these arrows individually:
I recommend git clone as it's easier.
Or use hugginface_hub client.
I didn't know such a thing existed! It's under the stacked tri-dots.
OK, I'll do that.
Finally. After the Git Clone I then moved the files to the proper directories & the Blank in the Node filled in.
Thanks.
as per above...... no luck!
Downloaded JSON files have this prefix: models--google--flan-t5-xl--textencoder... json
example pic has this:
model structure in installation info README.MD has this, which suggests we trim the prefix down:
I have them installed here, as this:
And I get this message from ComfyUI: _Failed to validate prompt for output 12:
When I change the file names to remove the prefixes to be this:
close my browser & restart ComfyUI I get this: And when I click the arrow to select it becomes this:
And this when I try to run: Failed to validate prompt for output 12: _* T5TextEncoderLoader #ELLA 13: