kohya-ss / kohya-ss/sd-scripts

DeepSpeed+LoRA(full) == 'DeepSpeedEngine' object has no attribute 'text_model'

Open
#712 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
7.2k
Forks
1.2k
Avg merge
11m
Merged PRs (30d)
2

Description

When I try to use DeepSpeed with LoRA with full training (unet + text encoder) I get this message:
'DeepSpeedEngine' object has no attribute 'text_model'

So it seems I can't train LoRA using DeepSpeed using just the unet (https://github.com/kohya-ss/sd-scripts/issues/711) or with the unet with the text encoder.

Any help would be appreciated :)

Contributor guide

No contributing guide indexed for this repository

Research direction

No files, tests, or entry points are named. Reproduce the reported DeepSpeed and LoRA full-training configuration, then trace where `text_model` is accessed on the `DeepSpeedEngine` and compare with issue #711; done means training with the UNet and text encoder no longer raises this attribute error.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.