Use qwen image edit (2509 and 2511) to generate black images

### Custom Node Testing

- [ ] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

System: Linux
Python version:. 12.3 | packaged by Anaconda, Inc. | (main, May 6, 2024, 19:46:43) [GCC 11.2.0]
Embedded Python: No
Pytorch version: 2.8.0+cu128
GPU:5090
Comfyui version: 0.3.60
Startup parameters: main.by -- port 6006

### Actual Behavior

Using the official qwen image edit workflow for single image editing, there is a problem of generating all black images after multiple runs;
The frequency is unstable, and it may take more than 20 times for a black image to appear

### Steps to Reproduce

### Debug Logs

```powershell
got prompt
WARNING: object supporting the buffer API required
Using xformers attention in VAE
Using xformers attention in VAE
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
Using scaled fp8: fp8 matrix mult: False, scale input: False
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
Requested to load QwenImageTEModel_
loaded completely 24628.920025634765 7909.737449645996 True
Requested to load WanVAE
loaded completely 13457.641510009766 242.02829551696777 True
Requested to load QwenImageTEModel_
loaded completely 24628.920025634765 7909.737449645996 True
model weight dtype torch.float8_e4m3fn, manual cast: torch.bfloat16
model_type FLUX
Requested to load QwenImage
loaded completely 24268.93999130249 19483.948791503906 True

0%| | 0/8 [00:00