Skip to content

An error occurs during the download of the Z-Image pipeline where from_single_file cannot be found in Qwen3ForCausalLM. #10

@uppiy47g

Description

@uppiy47g

Describe the bug

When attempting to download the Z-Image pipeline, the process fails during the text encoder loading step.
The pipeline crashes with an AttributeError indicating that Qwen3ForCausalLM does not have the method from_single_file.

Error Log

A relevant excerpt from the log:

AttributeError: type object 'Qwen3ForCausalLM' has no attribute 'from_single_file'
text_encoder = Qwen3ForCausalLM.from_single_file(...)
(from ZImagePipeline.py, line 355)

The text encoder is qwen-3-4b.safetensors, which can be used in ComfyUI.

AmuseDownload-04-05-2026-22654_080604.txt

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No fields configured for Bug.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions