[GH-ISSUE #20] Not so good configuration with memory bugs #18

New issue

Open

opened 2026-10-30 21:03:34 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-30 21:03:34 +00:00

Owner

Originally created by @cgvirus on GitHub (Apr 30, 2026).
Original GitHub issue: https://github.com/1038lab/ComfyUI-QwenTTS/issues/20

Unfortunately this workflow :

Lacks attention choice mode like sageattention, which is fastest to process
"Unload model" although has a switch doesn't really unload it
GPU always goes OOM, even for a single line of text of 10 words. GPU is 12 GB.
IF goes OOM, GPU Vram doesn't offload. We have to restart comfyui.

This workflow is much more better, maybe take a look at this for further dev?:
https://github.com/flybirdxx/ComfyUI-Qwen-TTS

Thanks.

Originally created by @cgvirus on GitHub (Apr 30, 2026). Original GitHub issue: https://github.com/1038lab/ComfyUI-QwenTTS/issues/20 Unfortunately this workflow : - Lacks attention choice mode like sageattention, which is fastest to process - "Unload model" although has a switch doesn't really unload it - GPU always goes OOM, even for a single line of text of 10 words. GPU is 12 GB. - IF goes OOM, GPU Vram doesn't offload. We have to restart comfyui. This workflow is much more better, maybe take a look at this for further dev?: https://github.com/flybirdxx/ComfyUI-Qwen-TTS Thanks.

No labels

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

1038lab/ComfyUI-QwenTTS#18

No description provided.

Rows
Columns