How is the refiner model different?

#1
by makisekurisu-jp - opened

HunyuanImage-2.1-GGUF is a quantization of the official original version, and HunyuanImage-2.1-Distilled-GGUF is a quantization of the distilled official version.
But what exactly is the “HunyuanImage-2.1-Refiner-GGUF” model, and how does it differ from the other two?
In addition, Tencent has released a fine-tuned Flux model called SRPO.
https://huggingface.co/tencent/SRPO
Would your organization consider creating a quantized version of this model?

refiner is an added step going over the almost finsihed image, improving quality at a late latent stage. SDXL has those, sometimes doing more harm than good, not seen them very much elsewhere.
It's optional for hunyuan and i can get it to work. A example workflow had a refiner node that i couldn't find yet.

refiner is an added step going over the almost finsihed image, improving quality at a late latent stage. SDXL has those, sometimes doing more harm than good, not seen them very much elsewhere.
It's optional for hunyuan and i can get it to work. A example workflow had a refiner node that i couldn't find yet.

We'll have to wait for an official release from Comfy. I haven't found any usable workflow that includes the refiner yet.

refiner is an added step going over the almost finsihed image, improving quality at a late latent stage. SDXL has those, sometimes doing more harm than good, not seen them very much elsewhere.
It's optional for hunyuan and i can get it to work. A example workflow had a refiner node that i couldn't find yet.

I can generate 2K images with the Q5KM base model, but using the refiner workflow always causes the Python program to crash—even when the refiner model is the GGUF-quantized Q4KM. My GPU has 12 GB of VRAM and my system RAM is 16 GB.

https://openart.ai/workflows/M1oiYEYRVhhXslYGSW7T

QuantStack org

Yeah i think refiner might have an issue, ill try to fix it today or tomorrow (;

QuantStack org

refiner is an added step going over the almost finsihed image, improving quality at a late latent stage. SDXL has those, sometimes doing more harm than good, not seen them very much elsewhere.
It's optional for hunyuan and i can get it to work. A example workflow had a refiner node that i couldn't find yet.

I can generate 2K images with the Q5KM base model, but using the refiner workflow always causes the Python program to crash—even when the refiner model is the GGUF-quantized Q4KM. My GPU has 12 GB of VRAM and my system RAM is 16 GB.

https://openart.ai/workflows/M1oiYEYRVhhXslYGSW7T

Do you have the latest nightly from comfyui? Since the refiner support was added later on, and i couldnt find any issue with the gguf

QuantStack org

Though i couldnt run inference, so if it is indeed the latest version

  1. Did the safetensors version work?
  2. Could you send the python tracelog?

Though i couldnt run inference, so if it is indeed the latest version

  1. Did the safetensors version work?
  2. Could you send the python tracelog?

workflow.png

workflow (1).png

## ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2025-09-20 10:28:36.420
** Platform: Windows
** Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr  8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
** Python executable: D:\ComfyUI\venv\Scripts\Python.exe
** ComfyUI Path: D:\ComfyUI
** ComfyUI Base Folder Path: D:\ComfyUI
** User directory: D:\ComfyUI\user
** ComfyUI-Manager config path: D:\ComfyUI\user\default\ComfyUI-Manager\config.ini
** Log path: D:\ComfyUI\user\comfyui.log

Prestartup times for custom nodes:
   0.0 seconds: D:\ComfyUI\custom_nodes\ComfyUI-Easy-Use
   0.0 seconds: D:\ComfyUI\custom_nodes\rgthree-comfy
  13.1 seconds: D:\ComfyUI\custom_nodes\ComfyUI-Manager

Checkpoint files will always be loaded safely.
Total VRAM 12282 MB, total RAM 16106 MB
pytorch version: 2.8.0+cu128
xformers version: 0.0.32.post2
Enabled fp16 accumulation.
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4070 Ti : cudaMallocAsync
Using sage attention
Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr  8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
ComfyUI version: 0.3.59
ComfyUI frontend version: 1.26.13
got prompt
Using xformers attention in VAE
Using xformers attention in VAE
VAE load device: cuda:0, offload device: cpu, dtype: torch.float16
Requested to load AutoencodingEngine
0 models unloaded.
loaded partially 128.0 127.8439998626709 0
Requested to load QwenImageTEModel_
loaded completely 9412.831000137328 7912.2333984375 True
gguf qtypes: F32 (713), Q3_K (482), Q5_K (2), Q4_K (38), BF16 (21)
model weight dtype torch.bfloat16, manual cast: None
model_type FLOW
unet missing: ['double_blocks.0.img_attn.qkv.weight', 'double_blocks.0.txt_attn.qkv.weight', 'double_blocks.1.img_attn.qkv.weight', 'double_blocks.1.txt_attn.qkv.weight', 'double_blocks.2.img_attn.qkv.weight', 'double_blocks.2.txt_attn.qkv.weight', 'double_blocks.3.img_attn.qkv.weight', 'double_blocks.3.txt_attn.qkv.weight', 'double_blocks.4.img_attn.qkv.weight', 'double_blocks.4.txt_attn.qkv.weight', 'double_blocks.5.img_attn.qkv.weight', 'double_blocks.5.txt_attn.qkv.weight', 'double_blocks.6.img_attn.qkv.weight', 'double_blocks.6.txt_attn.qkv.weight', 'double_blocks.7.img_attn.qkv.weight', 'double_blocks.7.txt_attn.qkv.weight', 'double_blocks.8.img_attn.qkv.weight', 'double_blocks.8.txt_attn.qkv.weight', 'double_blocks.9.img_attn.qkv.weight', 'double_blocks.9.txt_attn.qkv.weight', 'double_blocks.10.img_attn.qkv.weight', 'double_blocks.10.txt_attn.qkv.weight', 'double_blocks.11.img_attn.qkv.weight', 'double_blocks.11.txt_attn.qkv.weight', 'double_blocks.12.img_attn.qkv.weight', 'double_blocks.12.txt_attn.qkv.weight', 'double_blocks.13.img_attn.qkv.weight', 'double_blocks.13.txt_attn.qkv.weight', 'double_blocks.14.img_attn.qkv.weight', 'double_blocks.14.txt_attn.qkv.weight', 'double_blocks.15.img_attn.qkv.weight', 'double_blocks.15.txt_attn.qkv.weight', 'double_blocks.16.img_attn.qkv.weight', 'double_blocks.16.txt_attn.qkv.weight', 'double_blocks.17.img_attn.qkv.weight', 'double_blocks.17.txt_attn.qkv.weight', 'double_blocks.18.img_attn.qkv.weight', 'double_blocks.18.txt_attn.qkv.weight', 'double_blocks.19.img_attn.qkv.weight', 'double_blocks.19.txt_attn.qkv.weight', 'single_blocks.0.linear1.weight', 'single_blocks.0.linear2.weight', 'single_blocks.1.linear1.weight', 'single_blocks.1.linear2.weight', 'single_blocks.2.linear1.weight', 'single_blocks.2.linear2.weight', 'single_blocks.3.linear1.weight', 'single_blocks.3.linear2.weight', 'single_blocks.4.linear1.weight', 'single_blocks.4.linear2.weight', 'single_blocks.5.linear1.weight', 'single_blocks.5.linear2.weight', 'single_blocks.6.linear1.weight', 'single_blocks.6.linear2.weight', 'single_blocks.7.linear1.weight', 'single_blocks.7.linear2.weight', 'single_blocks.8.linear1.weight', 'single_blocks.8.linear2.weight', 'single_blocks.9.linear1.weight', 'single_blocks.9.linear2.weight', 'single_blocks.10.linear1.weight', 'single_blocks.10.linear2.weight', 'single_blocks.11.linear1.weight', 'single_blocks.11.linear2.weight', 'single_blocks.12.linear1.weight', 'single_blocks.12.linear2.weight', 'single_blocks.13.linear1.weight', 'single_blocks.13.linear2.weight', 'single_blocks.14.linear1.weight', 'single_blocks.14.linear2.weight', 'single_blocks.15.linear1.weight', 'single_blocks.15.linear2.weight', 'single_blocks.16.linear1.weight', 'single_blocks.16.linear2.weight', 'single_blocks.17.linear1.weight', 'single_blocks.17.linear2.weight', 'single_blocks.18.linear1.weight', 'single_blocks.18.linear2.weight', 'single_blocks.19.linear1.weight', 'single_blocks.19.linear2.weight', 'single_blocks.20.linear1.weight', 'single_blocks.20.linear2.weight', 'single_blocks.21.linear1.weight', 'single_blocks.21.linear2.weight', 'single_blocks.22.linear1.weight', 'single_blocks.22.linear2.weight', 'single_blocks.23.linear1.weight', 'single_blocks.23.linear2.weight', 'single_blocks.24.linear1.weight', 'single_blocks.24.linear2.weight', 'single_blocks.25.linear1.weight', 'single_blocks.25.linear2.weight', 'single_blocks.26.linear1.weight', 'single_blocks.26.linear2.weight', 'single_blocks.27.linear1.weight', 'single_blocks.27.linear2.weight', 'single_blocks.28.linear1.weight', 'single_blocks.28.linear2.weight', 'single_blocks.29.linear1.weight', 'single_blocks.29.linear2.weight', 'single_blocks.30.linear1.weight', 'single_blocks.30.linear2.weight', 'single_blocks.31.linear1.weight', 'single_blocks.31.linear2.weight', 'single_blocks.32.linear1.weight', 'single_blocks.32.linear2.weight', 'single_blocks.33.linear1.weight', 'single_blocks.33.linear2.weight', 'single_blocks.34.linear1.weight', 'single_blocks.34.linear2.weight', 'single_blocks.35.linear1.weight', 'single_blocks.35.linear2.weight', 'single_blocks.36.linear1.weight', 'single_blocks.36.linear2.weight', 'single_blocks.37.linear1.weight', 'single_blocks.37.linear2.weight', 'single_blocks.38.linear1.weight', 'single_blocks.38.linear2.weight', 'single_blocks.39.linear1.weight', 'single_blocks.39.linear2.weight']
unet unexpected: ['double_blocks.0.img_attn_k.bias', 'double_blocks.0.img_attn_k.weight', 'double_blocks.0.img_attn_q.bias', 'double_blocks.0.img_attn_q.weight', 'double_blocks.0.img_attn_v.bias', 'double_blocks.0.img_attn_v.weight', 'double_blocks.0.txt_attn_k.bias', 'double_blocks.0.txt_attn_k.weight', 'double_blocks.0.txt_attn_q.bias', 'double_blocks.0.txt_attn_q.weight', 'double_blocks.0.txt_attn_v.bias', 'double_blocks.0.txt_attn_v.weight', 'double_blocks.1.img_attn_k.bias', 'double_blocks.1.img_attn_k.weight', 'double_blocks.1.img_attn_q.bias', 'double_blocks.1.img_attn_q.weight', 'double_blocks.1.img_attn_v.bias', 'double_blocks.1.img_attn_v.weight', 'double_blocks.1.txt_attn_k.bias', 'double_blocks.1.txt_attn_k.weight', 'double_blocks.1.txt_attn_q.bias', 'double_blocks.1.txt_attn_q.weight', 'double_blocks.1.txt_attn_v.bias', 'double_blocks.1.txt_attn_v.weight', 'double_blocks.2.img_attn_k.bias', 'double_blocks.2.img_attn_k.weight', 'double_blocks.2.img_attn_q.bias', 'double_blocks.2.img_attn_q.weight', 'double_blocks.2.img_attn_v.bias', 'double_blocks.2.img_attn_v.weight', 'double_blocks.2.txt_attn_k.bias', 'double_blocks.2.txt_attn_k.weight', 'double_blocks.2.txt_attn_q.bias', 'double_blocks.2.txt_attn_q.weight', 'double_blocks.2.txt_attn_v.bias', 'double_blocks.2.txt_attn_v.weight', 'double_blocks.3.img_attn_k.bias', 'double_blocks.3.img_attn_k.weight', 'double_blocks.3.img_attn_q.bias', 'double_blocks.3.img_attn_q.weight', 'double_blocks.3.img_attn_v.bias', 'double_blocks.3.img_attn_v.weight', 'double_blocks.3.txt_attn_k.bias', 'double_blocks.3.txt_attn_k.weight', 'double_blocks.3.txt_attn_q.bias', 'double_blocks.3.txt_attn_q.weight', 'double_blocks.3.txt_attn_v.bias', 'double_blocks.3.txt_attn_v.weight', 'double_blocks.4.img_attn_k.bias', 'double_blocks.4.img_attn_k.weight', 'double_blocks.4.img_attn_q.bias', 'double_blocks.4.img_attn_q.weight', 'double_blocks.4.img_attn_v.bias', 'double_blocks.4.img_attn_v.weight', 'double_blocks.4.txt_attn_k.bias', 'double_blocks.4.txt_attn_k.weight', 'double_blocks.4.txt_attn_q.bias', 'double_blocks.4.txt_attn_q.weight', 'double_blocks.4.txt_attn_v.bias', 'double_blocks.4.txt_attn_v.weight', 'double_blocks.5.img_attn_k.bias', 'double_blocks.5.img_attn_k.weight', 'double_blocks.5.img_attn_q.bias', 'double_blocks.5.img_attn_q.weight', 'double_blocks.5.img_attn_v.bias', 'double_blocks.5.img_attn_v.weight', 'double_blocks.5.txt_attn_k.bias', 'double_blocks.5.txt_attn_k.weight', 'double_blocks.5.txt_attn_q.bias', 'double_blocks.5.txt_attn_q.weight', 'double_blocks.5.txt_attn_v.bias', 'double_blocks.5.txt_attn_v.weight', 'double_blocks.6.img_attn_k.bias', 'double_blocks.6.img_attn_k.weight', 'double_blocks.6.img_attn_q.bias', 'double_blocks.6.img_attn_q.weight', 'double_blocks.6.img_attn_v.bias', 'double_blocks.6.img_attn_v.weight', 'double_blocks.6.txt_attn_k.bias', 'double_blocks.6.txt_attn_k.weight', 'double_blocks.6.txt_attn_q.bias', 'double_blocks.6.txt_attn_q.weight', 'double_blocks.6.txt_attn_v.bias', 'double_blocks.6.txt_attn_v.weight', 'double_blocks.7.img_attn_k.bias', 'double_blocks.7.img_attn_k.weight', 'double_blocks.7.img_attn_q.bias', 'double_blocks.7.img_attn_q.weight', 'double_blocks.7.img_attn_v.bias', 'double_blocks.7.img_attn_v.weight', 'double_blocks.7.txt_attn_k.bias', 'double_blocks.7.txt_attn_k.weight', 'double_blocks.7.txt_attn_q.bias', 'double_blocks.7.txt_attn_q.weight', 'double_blocks.7.txt_attn_v.bias', 'double_blocks.7.txt_attn_v.weight', 'double_blocks.8.img_attn_k.bias', 'double_blocks.8.img_attn_k.weight', 'double_blocks.8.img_attn_q.bias', 'double_blocks.8.img_attn_q.weight', 'double_blocks.8.img_attn_v.bias', 'double_blocks.8.img_attn_v.weight', 'double_blocks.8.txt_attn_k.bias', 'double_blocks.8.txt_attn_k.weight', 'double_blocks.8.txt_attn_q.bias', 'double_blocks.8.txt_attn_q.weight', 'double_blocks.8.txt_attn_v.bias', 'double_blocks.8.txt_attn_v.weight', 'double_blocks.9.img_attn_k.bias', 'double_blocks.9.img_attn_k.weight', 'double_blocks.9.img_attn_q.bias', 'double_blocks.9.img_attn_q.weight', 'double_blocks.9.img_attn_v.bias', 'double_blocks.9.img_attn_v.weight', 'double_blocks.9.txt_attn_k.bias', 'double_blocks.9.txt_attn_k.weight', 'double_blocks.9.txt_attn_q.bias', 'double_blocks.9.txt_attn_q.weight', 'double_blocks.9.txt_attn_v.bias', 'double_blocks.9.txt_attn_v.weight', 'double_blocks.10.img_attn_k.bias', 'double_blocks.10.img_attn_k.weight', 'double_blocks.10.img_attn_q.bias', 'double_blocks.10.img_attn_q.weight', 'double_blocks.10.img_attn_v.bias', 'double_blocks.10.img_attn_v.weight', 'double_blocks.10.txt_attn_k.bias', 'double_blocks.10.txt_attn_k.weight', 'double_blocks.10.txt_attn_q.bias', 'double_blocks.10.txt_attn_q.weight', 'double_blocks.10.txt_attn_v.bias', 'double_blocks.10.txt_attn_v.weight', 'double_blocks.11.img_attn_k.bias', 'double_blocks.11.img_attn_k.weight', 'double_blocks.11.img_attn_q.bias', 'double_blocks.11.img_attn_q.weight', 'double_blocks.11.img_attn_v.bias', 'double_blocks.11.img_attn_v.weight', 'double_blocks.11.txt_attn_k.bias', 'double_blocks.11.txt_attn_k.weight', 'double_blocks.11.txt_attn_q.bias', 'double_blocks.11.txt_attn_q.weight', 'double_blocks.11.txt_attn_v.bias', 'double_blocks.11.txt_attn_v.weight', 'double_blocks.12.img_attn_k.bias', 'double_blocks.12.img_attn_k.weight', 'double_blocks.12.img_attn_q.bias', 'double_blocks.12.img_attn_q.weight', 'double_blocks.12.img_attn_v.bias', 'double_blocks.12.img_attn_v.weight', 'double_blocks.12.txt_attn_k.bias', 'double_blocks.12.txt_attn_k.weight', 'double_blocks.12.txt_attn_q.bias', 'double_blocks.12.txt_attn_q.weight', 'double_blocks.12.txt_attn_v.bias', 'double_blocks.12.txt_attn_v.weight', 'double_blocks.13.img_attn_k.bias', 'double_blocks.13.img_attn_k.weight', 'double_blocks.13.img_attn_q.bias', 'double_blocks.13.img_attn_q.weight', 'double_blocks.13.img_attn_v.bias', 'double_blocks.13.img_attn_v.weight', 'double_blocks.13.txt_attn_k.bias', 'double_blocks.13.txt_attn_k.weight', 'double_blocks.13.txt_attn_q.bias', 'double_blocks.13.txt_attn_q.weight', 'double_blocks.13.txt_attn_v.bias', 'double_blocks.13.txt_attn_v.weight', 'double_blocks.14.img_attn_k.bias', 'double_blocks.14.img_attn_k.weight', 'double_blocks.14.img_attn_q.bias', 'double_blocks.14.img_attn_q.weight', 'double_blocks.14.img_attn_v.bias', 'double_blocks.14.img_attn_v.weight', 'double_blocks.14.txt_attn_k.bias', 'double_blocks.14.txt_attn_k.weight', 'double_blocks.14.txt_attn_q.bias', 'double_blocks.14.txt_attn_q.weight', 'double_blocks.14.txt_attn_v.bias', 'double_blocks.14.txt_attn_v.weight', 'double_blocks.15.img_attn_k.bias', 'double_blocks.15.img_attn_k.weight', 'double_blocks.15.img_attn_q.bias', 'double_blocks.15.img_attn_q.weight', 'double_blocks.15.img_attn_v.bias', 'double_blocks.15.img_attn_v.weight', 'double_blocks.15.txt_attn_k.bias', 'double_blocks.15.txt_attn_k.weight', 'double_blocks.15.txt_attn_q.bias', 'double_blocks.15.txt_attn_q.weight', 'double_blocks.15.txt_attn_v.bias', 'double_blocks.15.txt_attn_v.weight', 'double_blocks.16.img_attn_k.bias', 'double_blocks.16.img_attn_k.weight', 'double_blocks.16.img_attn_q.bias', 'double_blocks.16.img_attn_q.weight', 'double_blocks.16.img_attn_v.bias', 'double_blocks.16.img_attn_v.weight', 'double_blocks.16.txt_attn_k.bias', 'double_blocks.16.txt_attn_k.weight', 'double_blocks.16.txt_attn_q.bias', 'double_blocks.16.txt_attn_q.weight', 'double_blocks.16.txt_attn_v.bias', 'double_blocks.16.txt_attn_v.weight', 'double_blocks.17.img_attn_k.bias', 'double_blocks.17.img_attn_k.weight', 'double_blocks.17.img_attn_q.bias', 'double_blocks.17.img_attn_q.weight', 'double_blocks.17.img_attn_v.bias', 'double_blocks.17.img_attn_v.weight', 'double_blocks.17.txt_attn_k.bias', 'double_blocks.17.txt_attn_k.weight', 'double_blocks.17.txt_attn_q.bias', 'double_blocks.17.txt_attn_q.weight', 'double_blocks.17.txt_attn_v.bias', 'double_blocks.17.txt_attn_v.weight', 'double_blocks.18.img_attn_k.bias', 'double_blocks.18.img_attn_k.weight', 'double_blocks.18.img_attn_q.bias', 'double_blocks.18.img_attn_q.weight', 'double_blocks.18.img_attn_v.bias', 'double_blocks.18.img_attn_v.weight', 'double_blocks.18.txt_attn_k.bias', 'double_blocks.18.txt_attn_k.weight', 'double_blocks.18.txt_attn_q.bias', 'double_blocks.18.txt_attn_q.weight', 'double_blocks.18.txt_attn_v.bias', 'double_blocks.18.txt_attn_v.weight', 'double_blocks.19.img_attn_k.bias', 'double_blocks.19.img_attn_k.weight', 'double_blocks.19.img_attn_q.bias', 'double_blocks.19.img_attn_q.weight', 'double_blocks.19.img_attn_v.bias', 'double_blocks.19.img_attn_v.weight', 'double_blocks.19.txt_attn_k.bias', 'double_blocks.19.txt_attn_k.weight', 'double_blocks.19.txt_attn_q.bias', 'double_blocks.19.txt_attn_q.weight', 'double_blocks.19.txt_attn_v.bias', 'double_blocks.19.txt_attn_v.weight', 'single_blocks.0.linear1_k.bias', 'single_blocks.0.linear1_k.weight', 'single_blocks.0.linear1_mlp.bias', 'single_blocks.0.linear1_mlp.weight', 'single_blocks.0.linear1_q.bias', 'single_blocks.0.linear1_q.weight', 'single_blocks.0.linear1_v.bias', 'single_blocks.0.linear1_v.weight', 'single_blocks.0.linear2.fc.bias', 'single_blocks.0.linear2.fc.weight', 'single_blocks.1.linear1_k.bias', 'single_blocks.1.linear1_k.weight', 'single_blocks.1.linear1_mlp.bias', 'single_blocks.1.linear1_mlp.weight', 'single_blocks.1.linear1_q.bias', 'single_blocks.1.linear1_q.weight', 'single_blocks.1.linear1_v.bias', 'single_blocks.1.linear1_v.weight', 'single_blocks.1.linear2.fc.bias', 'single_blocks.1.linear2.fc.weight', 'single_blocks.2.linear1_k.bias', 'single_blocks.2.linear1_k.weight', 'single_blocks.2.linear1_mlp.bias', 'single_blocks.2.linear1_mlp.weight', 'single_blocks.2.linear1_q.bias', 'single_blocks.2.linear1_q.weight', 'single_blocks.2.linear1_v.bias', 'single_blocks.2.linear1_v.weight', 'single_blocks.2.linear2.fc.bias', 'single_blocks.2.linear2.fc.weight', 'single_blocks.3.linear1_k.bias', 'single_blocks.3.linear1_k.weight', 'single_blocks.3.linear1_mlp.bias', 'single_blocks.3.linear1_mlp.weight', 'single_blocks.3.linear1_q.bias', 'single_blocks.3.linear1_q.weight', 'single_blocks.3.linear1_v.bias', 'single_blocks.3.linear1_v.weight', 'single_blocks.3.linear2.fc.bias', 'single_blocks.3.linear2.fc.weight', 'single_blocks.4.linear1_k.bias', 'single_blocks.4.linear1_k.weight', 'single_blocks.4.linear1_mlp.bias', 'single_blocks.4.linear1_mlp.weight', 'single_blocks.4.linear1_q.bias', 'single_blocks.4.linear1_q.weight', 'single_blocks.4.linear1_v.bias', 'single_blocks.4.linear1_v.weight', 'single_blocks.4.linear2.fc.bias', 'single_blocks.4.linear2.fc.weight', 'single_blocks.5.linear1_k.bias', 'single_blocks.5.linear1_k.weight', 'single_blocks.5.linear1_mlp.bias', 'single_blocks.5.linear1_mlp.weight', 'single_blocks.5.linear1_q.bias', 'single_blocks.5.linear1_q.weight', 'single_blocks.5.linear1_v.bias', 'single_blocks.5.linear1_v.weight', 'single_blocks.5.linear2.fc.bias', 'single_blocks.5.linear2.fc.weight', 'single_blocks.6.linear1_k.bias', 'single_blocks.6.linear1_k.weight', 'single_blocks.6.linear1_mlp.bias', 'single_blocks.6.linear1_mlp.weight', 'single_blocks.6.linear1_q.bias', 'single_blocks.6.linear1_q.weight', 'single_blocks.6.linear1_v.bias', 'single_blocks.6.linear1_v.weight', 'single_blocks.6.linear2.fc.bias', 'single_blocks.6.linear2.fc.weight', 'single_blocks.7.linear1_k.bias', 'single_blocks.7.linear1_k.weight', 'single_blocks.7.linear1_mlp.bias', 'single_blocks.7.linear1_mlp.weight', 'single_blocks.7.linear1_q.bias', 'single_blocks.7.linear1_q.weight', 'single_blocks.7.linear1_v.bias', 'single_blocks.7.linear1_v.weight', 'single_blocks.7.linear2.fc.bias', 'single_blocks.7.linear2.fc.weight', 'single_blocks.8.linear1_k.bias', 'single_blocks.8.linear1_k.weight', 'single_blocks.8.linear1_mlp.bias', 'single_blocks.8.linear1_mlp.weight', 'single_blocks.8.linear1_q.bias', 'single_blocks.8.linear1_q.weight', 'single_blocks.8.linear1_v.bias', 'single_blocks.8.linear1_v.weight', 'single_blocks.8.linear2.fc.bias', 'single_blocks.8.linear2.fc.weight', 'single_blocks.9.linear1_k.bias', 'single_blocks.9.linear1_k.weight', 'single_blocks.9.linear1_mlp.bias', 'single_blocks.9.linear1_mlp.weight', 'single_blocks.9.linear1_q.bias', 'single_blocks.9.linear1_q.weight', 'single_blocks.9.linear1_v.bias', 'single_blocks.9.linear1_v.weight', 'single_blocks.9.linear2.fc.bias', 'single_blocks.9.linear2.fc.weight', 'single_blocks.10.linear1_k.bias', 'single_blocks.10.linear1_k.weight', 'single_blocks.10.linear1_mlp.bias', 'single_blocks.10.linear1_mlp.weight', 'single_blocks.10.linear1_q.bias', 'single_blocks.10.linear1_q.weight', 'single_blocks.10.linear1_v.bias', 'single_blocks.10.linear1_v.weight', 'single_blocks.10.linear2.fc.bias', 'single_blocks.10.linear2.fc.weight', 'single_blocks.11.linear1_k.bias', 'single_blocks.11.linear1_k.weight', 'single_blocks.11.linear1_mlp.bias', 'single_blocks.11.linear1_mlp.weight', 'single_blocks.11.linear1_q.bias', 'single_blocks.11.linear1_q.weight', 'single_blocks.11.linear1_v.bias', 'single_blocks.11.linear1_v.weight', 'single_blocks.11.linear2.fc.bias', 'single_blocks.11.linear2.fc.weight', 'single_blocks.12.linear1_k.bias', 'single_blocks.12.linear1_k.weight', 'single_blocks.12.linear1_mlp.bias', 'single_blocks.12.linear1_mlp.weight', 'single_blocks.12.linear1_q.bias', 'single_blocks.12.linear1_q.weight', 'single_blocks.12.linear1_v.bias', 'single_blocks.12.linear1_v.weight', 'single_blocks.12.linear2.fc.bias', 'single_blocks.12.linear2.fc.weight', 'single_blocks.13.linear1_k.bias', 'single_blocks.13.linear1_k.weight', 'single_blocks.13.linear1_mlp.bias', 'single_blocks.13.linear1_mlp.weight', 'single_blocks.13.linear1_q.bias', 'single_blocks.13.linear1_q.weight', 'single_blocks.13.linear1_v.bias', 'single_blocks.13.linear1_v.weight', 'single_blocks.13.linear2.fc.bias', 'single_blocks.13.linear2.fc.weight', 'single_blocks.14.linear1_k.bias', 'single_blocks.14.linear1_k.weight', 'single_blocks.14.linear1_mlp.bias', 'single_blocks.14.linear1_mlp.weight', 'single_blocks.14.linear1_q.bias', 'single_blocks.14.linear1_q.weight', 'single_blocks.14.linear1_v.bias', 'single_blocks.14.linear1_v.weight', 'single_blocks.14.linear2.fc.bias', 'single_blocks.14.linear2.fc.weight', 'single_blocks.15.linear1_k.bias', 'single_blocks.15.linear1_k.weight', 'single_blocks.15.linear1_mlp.bias', 'single_blocks.15.linear1_mlp.weight', 'single_blocks.15.linear1_q.bias', 'single_blocks.15.linear1_q.weight', 'single_blocks.15.linear1_v.bias', 'single_blocks.15.linear1_v.weight', 'single_blocks.15.linear2.fc.bias', 'single_blocks.15.linear2.fc.weight', 'single_blocks.16.linear1_k.bias', 'single_blocks.16.linear1_k.weight', 'single_blocks.16.linear1_mlp.bias', 'single_blocks.16.linear1_mlp.weight', 'single_blocks.16.linear1_q.bias', 'single_blocks.16.linear1_q.weight', 'single_blocks.16.linear1_v.bias', 'single_blocks.16.linear1_v.weight', 'single_blocks.16.linear2.fc.bias', 'single_blocks.16.linear2.fc.weight', 'single_blocks.17.linear1_k.bias', 'single_blocks.17.linear1_k.weight', 'single_blocks.17.linear1_mlp.bias', 'single_blocks.17.linear1_mlp.weight', 'single_blocks.17.linear1_q.bias', 'single_blocks.17.linear1_q.weight', 'single_blocks.17.linear1_v.bias', 'single_blocks.17.linear1_v.weight', 'single_blocks.17.linear2.fc.bias', 'single_blocks.17.linear2.fc.weight', 'single_blocks.18.linear1_k.bias', 'single_blocks.18.linear1_k.weight', 'single_blocks.18.linear1_mlp.bias', 'single_blocks.18.linear1_mlp.weight', 'single_blocks.18.linear1_q.bias', 'single_blocks.18.linear1_q.weight', 'single_blocks.18.linear1_v.bias', 'single_blocks.18.linear1_v.weight', 'single_blocks.18.linear2.fc.bias', 'single_blocks.18.linear2.fc.weight', 'single_blocks.19.linear1_k.bias', 'single_blocks.19.linear1_k.weight', 'single_blocks.19.linear1_mlp.bias', 'single_blocks.19.linear1_mlp.weight', 'single_blocks.19.linear1_q.bias', 'single_blocks.19.linear1_q.weight', 'single_blocks.19.linear1_v.bias', 'single_blocks.19.linear1_v.weight', 'single_blocks.19.linear2.fc.bias', 'single_blocks.19.linear2.fc.weight', 'single_blocks.20.linear1_k.bias', 'single_blocks.20.linear1_k.weight', 'single_blocks.20.linear1_mlp.bias', 'single_blocks.20.linear1_mlp.weight', 'single_blocks.20.linear1_q.bias', 'single_blocks.20.linear1_q.weight', 'single_blocks.20.linear1_v.bias', 'single_blocks.20.linear1_v.weight', 'single_blocks.20.linear2.fc.bias', 'single_blocks.20.linear2.fc.weight', 'single_blocks.21.linear1_k.bias', 'single_blocks.21.linear1_k.weight', 'single_blocks.21.linear1_mlp.bias', 'single_blocks.21.linear1_mlp.weight', 'single_blocks.21.linear1_q.bias', 'single_blocks.21.linear1_q.weight', 'single_blocks.21.linear1_v.bias', 'single_blocks.21.linear1_v.weight', 'single_blocks.21.linear2.fc.bias', 'single_blocks.21.linear2.fc.weight', 'single_blocks.22.linear1_k.bias', 'single_blocks.22.linear1_k.weight', 'single_blocks.22.linear1_mlp.bias', 'single_blocks.22.linear1_mlp.weight', 'single_blocks.22.linear1_q.bias', 'single_blocks.22.linear1_q.weight', 'single_blocks.22.linear1_v.bias', 'single_blocks.22.linear1_v.weight', 'single_blocks.22.linear2.fc.bias', 'single_blocks.22.linear2.fc.weight', 'single_blocks.23.linear1_k.bias', 'single_blocks.23.linear1_k.weight', 'single_blocks.23.linear1_mlp.bias', 'single_blocks.23.linear1_mlp.weight', 'single_blocks.23.linear1_q.bias', 'single_blocks.23.linear1_q.weight', 'single_blocks.23.linear1_v.bias', 'single_blocks.23.linear1_v.weight', 'single_blocks.23.linear2.fc.bias', 'single_blocks.23.linear2.fc.weight', 'single_blocks.24.linear1_k.bias', 'single_blocks.24.linear1_k.weight', 'single_blocks.24.linear1_mlp.bias', 'single_blocks.24.linear1_mlp.weight', 'single_blocks.24.linear1_q.bias', 'single_blocks.24.linear1_q.weight', 'single_blocks.24.linear1_v.bias', 'single_blocks.24.linear1_v.weight', 'single_blocks.24.linear2.fc.bias', 'single_blocks.24.linear2.fc.weight', 'single_blocks.25.linear1_k.bias', 'single_blocks.25.linear1_k.weight', 'single_blocks.25.linear1_mlp.bias', 'single_blocks.25.linear1_mlp.weight', 'single_blocks.25.linear1_q.bias', 'single_blocks.25.linear1_q.weight', 'single_blocks.25.linear1_v.bias', 'single_blocks.25.linear1_v.weight', 'single_blocks.25.linear2.fc.bias', 'single_blocks.25.linear2.fc.weight', 'single_blocks.26.linear1_k.bias', 'single_blocks.26.linear1_k.weight', 'single_blocks.26.linear1_mlp.bias', 'single_blocks.26.linear1_mlp.weight', 'single_blocks.26.linear1_q.bias', 'single_blocks.26.linear1_q.weight', 'single_blocks.26.linear1_v.bias', 'single_blocks.26.linear1_v.weight', 'single_blocks.26.linear2.fc.bias', 'single_blocks.26.linear2.fc.weight', 'single_blocks.27.linear1_k.bias', 'single_blocks.27.linear1_k.weight', 'single_blocks.27.linear1_mlp.bias', 'single_blocks.27.linear1_mlp.weight', 'single_blocks.27.linear1_q.bias', 'single_blocks.27.linear1_q.weight', 'single_blocks.27.linear1_v.bias', 'single_blocks.27.linear1_v.weight', 'single_blocks.27.linear2.fc.bias', 'single_blocks.27.linear2.fc.weight', 'single_blocks.28.linear1_k.bias', 'single_blocks.28.linear1_k.weight', 'single_blocks.28.linear1_mlp.bias', 'single_blocks.28.linear1_mlp.weight', 'single_blocks.28.linear1_q.bias', 'single_blocks.28.linear1_q.weight', 'single_blocks.28.linear1_v.bias', 'single_blocks.28.linear1_v.weight', 'single_blocks.28.linear2.fc.bias', 'single_blocks.28.linear2.fc.weight', 'single_blocks.29.linear1_k.bias', 'single_blocks.29.linear1_k.weight', 'single_blocks.29.linear1_mlp.bias', 'single_blocks.29.linear1_mlp.weight', 'single_blocks.29.linear1_q.bias', 'single_blocks.29.linear1_q.weight', 'single_blocks.29.linear1_v.bias', 'single_blocks.29.linear1_v.weight', 'single_blocks.29.linear2.fc.bias', 'single_blocks.29.linear2.fc.weight', 'single_blocks.30.linear1_k.bias', 'single_blocks.30.linear1_k.weight', 'single_blocks.30.linear1_mlp.bias', 'single_blocks.30.linear1_mlp.weight', 'single_blocks.30.linear1_q.bias', 'single_blocks.30.linear1_q.weight', 'single_blocks.30.linear1_v.bias', 'single_blocks.30.linear1_v.weight', 'single_blocks.30.linear2.fc.bias', 'single_blocks.30.linear2.fc.weight', 'single_blocks.31.linear1_k.bias', 'single_blocks.31.linear1_k.weight', 'single_blocks.31.linear1_mlp.bias', 'single_blocks.31.linear1_mlp.weight', 'single_blocks.31.linear1_q.bias', 'single_blocks.31.linear1_q.weight', 'single_blocks.31.linear1_v.bias', 'single_blocks.31.linear1_v.weight', 'single_blocks.31.linear2.fc.bias', 'single_blocks.31.linear2.fc.weight', 'single_blocks.32.linear1_k.bias', 'single_blocks.32.linear1_k.weight', 'single_blocks.32.linear1_mlp.bias', 'single_blocks.32.linear1_mlp.weight', 'single_blocks.32.linear1_q.bias', 'single_blocks.32.linear1_q.weight', 'single_blocks.32.linear1_v.bias', 'single_blocks.32.linear1_v.weight', 'single_blocks.32.linear2.fc.bias', 'single_blocks.32.linear2.fc.weight', 'single_blocks.33.linear1_k.bias', 'single_blocks.33.linear1_k.weight', 'single_blocks.33.linear1_mlp.bias', 'single_blocks.33.linear1_mlp.weight', 'single_blocks.33.linear1_q.bias', 'single_blocks.33.linear1_q.weight', 'single_blocks.33.linear1_v.bias', 'single_blocks.33.linear1_v.weight', 'single_blocks.33.linear2.fc.bias', 'single_blocks.33.linear2.fc.weight', 'single_blocks.34.linear1_k.bias', 'single_blocks.34.linear1_k.weight', 'single_blocks.34.linear1_mlp.bias', 'single_blocks.34.linear1_mlp.weight', 'single_blocks.34.linear1_q.bias', 'single_blocks.34.linear1_q.weight', 'single_blocks.34.linear1_v.bias', 'single_blocks.34.linear1_v.weight', 'single_blocks.34.linear2.fc.bias', 'single_blocks.34.linear2.fc.weight', 'single_blocks.35.linear1_k.bias', 'single_blocks.35.linear1_k.weight', 'single_blocks.35.linear1_mlp.bias', 'single_blocks.35.linear1_mlp.weight', 'single_blocks.35.linear1_q.bias', 'single_blocks.35.linear1_q.weight', 'single_blocks.35.linear1_v.bias', 'single_blocks.35.linear1_v.weight', 'single_blocks.35.linear2.fc.bias', 'single_blocks.35.linear2.fc.weight', 'single_blocks.36.linear1_k.bias', 'single_blocks.36.linear1_k.weight', 'single_blocks.36.linear1_mlp.bias', 'single_blocks.36.linear1_mlp.weight', 'single_blocks.36.linear1_q.bias', 'single_blocks.36.linear1_q.weight', 'single_blocks.36.linear1_v.bias', 'single_blocks.36.linear1_v.weight', 'single_blocks.36.linear2.fc.bias', 'single_blocks.36.linear2.fc.weight', 'single_blocks.37.linear1_k.bias', 'single_blocks.37.linear1_k.weight', 'single_blocks.37.linear1_mlp.bias', 'single_blocks.37.linear1_mlp.weight', 'single_blocks.37.linear1_q.bias', 'single_blocks.37.linear1_q.weight', 'single_blocks.37.linear1_v.bias', 'single_blocks.37.linear1_v.weight', 'single_blocks.37.linear2.fc.bias', 'single_blocks.37.linear2.fc.weight', 'single_blocks.38.linear1_k.bias', 'single_blocks.38.linear1_k.weight', 'single_blocks.38.linear1_mlp.bias', 'single_blocks.38.linear1_mlp.weight', 'single_blocks.38.linear1_q.bias', 'single_blocks.38.linear1_q.weight', 'single_blocks.38.linear1_v.bias', 'single_blocks.38.linear1_v.weight', 'single_blocks.38.linear2.fc.bias', 'single_blocks.38.linear2.fc.weight', 'single_blocks.39.linear1_k.bias', 'single_blocks.39.linear1_k.weight', 'single_blocks.39.linear1_mlp.bias', 'single_blocks.39.linear1_mlp.weight', 'single_blocks.39.linear1_q.bias', 'single_blocks.39.linear1_q.weight', 'single_blocks.39.linear1_v.bias', 'single_blocks.39.linear1_v.weight', 'single_blocks.39.linear2.fc.bias', 'single_blocks.39.linear2.fc.weight']
Requested to load HunyuanImage21Refiner
Exception Code: 0xC0000005
                          0x00007FF87E4C65DA, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x865DA byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0x7B4A byte(s)
0x00007FF87E4C6D58, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x86D58 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0x82C8 byte(s)
0x00007FF87E4C98CB, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x898CB byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0xAE3B byte(s)
0x00007FF87E4C9597, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x89597 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0xAB07 byte(s)
0x00007FF87E4C8C30, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x88C30 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0xA1A0 byte(s)
0x00007FF87E4C3CC2, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x83CC2 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0x5232 byte(s)
0x00007FF87E4C3B59, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x83B59 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0x50C9 byte(s)
0x00007FF87E4C3DA6, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x83DA6 byte(s), ?typeMetaDatas@TypeMeta@caffe2@@CAPEAUTypeMetaData@detail@2@XZ() + 0x5316 byte(s)
0x00007FF87E493FB8, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x53FB8 byte(s), ?alloc_cpu@c10@@YAPEAX_K@Z() + 0x58 byte(s)
0x00007FF87E449BEC, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x9BEC byte(s), ?allocate@DefaultCPUAllocator@c10@@UEAA?AVDataPtr@2@_K@Z() + 0x3C byte(s)
0x00007FF87E463DA3, D:\ComfyUI\venv\Lib\site-packages\torch\lib\c10.dll(0x00007FF87E440000) + 0x23DA3 byte(s), ??0StorageImpl@c10@@QEAA@Uuse_byte_size_t@01@AEBVSymInt@1@PEAUAllocator@1@_N@Z() + 0x63 byte(s)
0x00007FFF90E038A9, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x5DC38A9 byte(s), ?sym@DynamicLibrary@at@@QEAAPEAXPEBD@Z() + 0xD39 byte(s)
0x00007FFF90E05AC7, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x5DC5AC7 byte(s), ?empty_strided_cpu@detail@at@@YA?AVTensorBase@2@V?$ArrayRef@_J@c10@@0V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@7@V?$optional@UDevice@c10@@@7@V?$optional@_N@7@@Z() + 0x97 byte(s)
0x00007FFF9138B2D9, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x634B2D9 byte(s), ?empty_strided_cpu@native@at@@YA?AVTensor@2@V?$ArrayRef@_J@c10@@0V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@7@V?$optional@UDevice@c10@@@7@V?$optional@_N@7@@Z() + 0x79 byte(s)
0x00007FFF920F2910, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x70B2910 byte(s), ?where_outf@cpu@at@@YAAEAVTensor@2@AEBV32@00AEAV32@@Z() + 0x9C0 byte(s)
0x00007FFF920EA198, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x70AA198 byte(s), ?bucketize_outf@cpu@at@@YAAEAVTensor@2@AEBV32@0_N1AEAV32@@Z() + 0x10008 byte(s)
0x00007FFF91A403DB, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6A003DB byte(s), ?redispatch@xlogy__Tensor@_ops@at@@SAAEAVTensor@3@VDispatchKeySet@c10@@AEAV43@AEBV43@@Z() + 0x7DDAB byte(s)
0x00007FFF91B1546E, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6AD546E byte(s), ?redispatch@empty_strided@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@V?$ArrayRef@VSymInt@c10@@@6@1V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@9@V?$optional@UDevice@c10@@@9@V?$optional@_N@9@@Z() + 0xEE byte(s)
0x00007FFF91D7893F, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D3893F byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x23F5F byte(s)
0x00007FFF91D75748, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D35748 byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x20D68 byte(s)
0x00007FFF91A8A97F, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6A4A97F byte(s), ?call@empty_strided@_ops@at@@SA?AVTensor@3@V?$ArrayRef@VSymInt@c10@@@c10@@0V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@8@V?$optional@UDevice@c10@@@8@V?$optional@_N@8@@Z() + 0x23F byte(s)
0x00007FFF90E6D58D, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x5E2D58D byte(s), ?declare_static_shape@TensorIteratorConfig@at@@QEAAAEAV12@V?$ArrayRef@_J@c10@@@Z() + 0x3AD byte(s)
0x00007FFF9137721A, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x633721A byte(s), ?_to_copy@native@at@@YA?AVTensor@2@AEBV32@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@5@V?$optional@UDevice@c10@@@5@V?$optional@_N@5@_NV?$optional@W4MemoryFormat@c10@@@5@@Z() + 0x89A byte(s)
0x00007FFF91E948BC, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6E548BC byte(s), ?view_copy_symint_outf@compositeexplicitautograd@at@@YAAEAVTensor@2@AEBV32@V?$ArrayRef@VSymInt@c10@@@c10@@AEAV32@@Z() + 0x306C byte(s)
0x00007FFF91E643D2, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6E243D2 byte(s), ?bucketize_outf@compositeexplicitautograd@at@@YAAEAVTensor@2@AEBVScalar@c10@@AEBV32@_N2AEAV32@@Z() + 0x48002 byte(s)
0x00007FFF91777A24, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6737A24 byte(s), ?make_tensor@TensorMaker@at@@QEAA?AVTensor@2@XZ() + 0xADB34 byte(s)
0x00007FFF9183D4EC, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x67FD4EC byte(s), ?redispatch@_to_copy@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@8@V?$optional@UDevice@c10@@@8@V?$optional@_N@8@_NV?$optional@W4MemoryFormat@c10@@@8@@Z() + 0xDC byte(s)
0x00007FFF91D6FF42, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D2FF42 byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x1B562 byte(s)
0x00007FFF91D74FE2, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D34FE2 byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x20602 byte(s)
0x00007FFF917B3913, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6773913 byte(s), ?call@_to_copy@_ops@at@@SA?AVTensor@3@AEBV43@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@6@V?$optional@UDevice@c10@@@6@V?$optional@_N@6@_NV?$optional@W4MemoryFormat@c10@@@6@@Z() + 0x203 byte(s)
0x00007FFF9137FC4E, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x633FC4E byte(s), ?to_dense_backward@native@at@@YA?AVTensor@2@AEBV32@0V?$optional@_N@std@@@Z() + 0x54E byte(s)
0x00007FFF9137F4D7, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x633F4D7 byte(s), ?to@native@at@@YA?AVTensor@2@AEBV32@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@5@V?$optional@UDevice@c10@@@5@V?$optional@_N@5@_N5V?$optional@W4MemoryFormat@c10@@@5@@Z() + 0x107 byte(s)
0x00007FFF9201DBC8, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6FDDBC8 byte(s), ?where@compositeimplicitautograd@at@@YA?AVTensor@2@AEBV32@AEBVScalar@c10@@1@Z() + 0x7238 byte(s)
0x00007FFF9200086E, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6FC086E byte(s), ?broadcast_to_symint@compositeimplicitautograd@at@@YA?AVTensor@2@AEBV32@V?$ArrayRef@VSymInt@c10@@@c10@@@Z() + 0x3A8BE byte(s)
0x00007FFF918E5140, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x68A5140 byte(s), ?redispatch@zeros_out@_ops@at@@SAAEAVTensor@3@VDispatchKeySet@c10@@V?$ArrayRef@VSymInt@c10@@@6@AEAV43@@Z() + 0x7BA10 byte(s)
0x00007FFF919C0878, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6980878 byte(s), ?redispatch@to_dtype_layout@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@8@V?$optional@UDevice@c10@@@8@V?$optional@_N@8@_N6V?$optional@W4MemoryFormat@c10@@@8@@Z() + 0xE8 byte(s)
0x00007FFF91D7C79E, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D3C79E byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x27DBE byte(s)
0x00007FFF91D74F5E, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x6D34F5E byte(s), ?redispatch@view_as_real@_ops@at@@SA?AVTensor@3@VDispatchKeySet@c10@@AEBV43@@Z() + 0x2057E byte(s)
0x00007FFF9196554D, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x692554D byte(s), ?call@to_dtype_layout@_ops@at@@SA?AVTensor@3@AEBV43@V?$optional@W4ScalarType@c10@@@std@@V?$optional@W4Layout@c10@@@6@V?$optional@UDevice@c10@@@6@V?$optional@_N@6@_N5V?$optional@W4MemoryFormat@c10@@@6@@Z() + 0x21D byte(s)
0x00007FFF90DAC91D, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_cpu.dll(0x00007FFF8B040000) + 0x5D6C91D byte(s), ?to@Tensor@at@@QEBA?AV12@UTensorOptions@c10@@_N1V?$optional@W4MemoryFormat@c10@@@std@@@Z() + 0x1CD byte(s)
0x00007FF86111A2C0, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x12A2C0 byte(s), ?dimnamelist@PythonArgs@torch@@QEAA?AV?$vector@UDimname@at@@V?$allocator@UDimname@at@@@std@@@std@@H@Z() + 0x390 byte(s)
0x00007FF8611074C4, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x1174C4 byte(s), ?get@?$THPPointer@UTHPStorage@@@@QEBAPEBUTHPStorage@@XZ() + 0xCE054 byte(s)
0x00007FF87ED630E3, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x630E3 byte(s), PyErr_SetNone() + 0xDF byte(s)
0x00007FF87EDB2815, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB2815 byte(s), PyObject_Call() + 0x125 byte(s)
0x00007FF87EDB275F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB275F byte(s), PyObject_Call() + 0x6F byte(s)
0x00007FF87ED866EE, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x866EE byte(s), _PyEval_EvalFrameDefault() + 0x40DE byte(s)
0x00007FF87ED8103C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x8103C byte(s), _PyFunction_Vectorcall() + 0x17C byte(s)
0x00007FF87ED7BD8E, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7BD8E byte(s), _PyArg_CheckPositional() + 0x3D2 byte(s)
0x00007FF87ED81A5F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x81A5F byte(s), PyObject_Vectorcall() + 0xCF byte(s)
0x00007FF87ED07E26, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7E26 byte(s), PyObject_CallMethodObjArgs() + 0x14E byte(s)
0x00007FF87ED07CD2, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7CD2 byte(s), PyObject_CallFunctionObjArgs() + 0x42 byte(s)
0x00007FF8618FAB75, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x90AB75 byte(s), ?check_deprecated@PythonArgParser@torch@@AEAAXAEBUFunctionSignature@2@@Z() + 0x13E5 byte(s)
0x00007FF8618FC032, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x90C032 byte(s), ?handle_torch_function_no_python_arg_parser@torch@@YAPEAU_object@@V?$ArrayRef@PEAU_object@@@c10@@PEAU2@1PEBD12W4TorchFunctionName@1@@Z() + 0x3D2 byte(s)
0x00007FF8618FB34B, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x90B34B byte(s), ?get_signatures@PythonArgParser@torch@@QEBA?AV?$vector@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@V?$allocator@V?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@2@@std@@XZ() + 0x2AB byte(s)
0x00007FF86110763B, D:\ComfyUI\venv\Lib\site-packages\torch\lib\torch_python.dll(0x00007FF860FF0000) + 0x11763B byte(s), ?get@?$THPPointer@UTHPStorage@@@@QEBAPEBUTHPStorage@@XZ() + 0xCE1CB byte(s)
0x00007FF87ED379E4, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x379E4 byte(s), PyThread_acquire_lock_timed() + 0x5F0 byte(s)
0x00007FF87EDB27A6, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB27A6 byte(s), PyObject_Call() + 0xB6 byte(s)
0x00007FF87ED866EE, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x866EE byte(s), _PyEval_EvalFrameDefault() + 0x40DE byte(s)
0x00007FF87ED8103C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x8103C byte(s), _PyFunction_Vectorcall() + 0x17C byte(s)
0x00007FF87ED7BD8E, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7BD8E byte(s), _PyArg_CheckPositional() + 0x3D2 byte(s)
0x00007FF87EDB2815, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB2815 byte(s), PyObject_Call() + 0x125 byte(s)
0x00007FF87EDB275F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB275F byte(s), PyObject_Call() + 0x6F byte(s)
0x00007FF87ED866EE, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x866EE byte(s), _PyEval_EvalFrameDefault() + 0x40DE byte(s)
0x00007FF87ED8103C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x8103C byte(s), _PyFunction_Vectorcall() + 0x17C byte(s)
0x00007FF87ED7BE18, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7BE18 byte(s), _PyArg_CheckPositional() + 0x45C byte(s)
0x00007FF87EDB286B, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB286B byte(s), PyObject_Call() + 0x17B byte(s)
0x00007FF87EDB275F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB275F byte(s), PyObject_Call() + 0x6F byte(s)
0x00007FF87ED866EE, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x866EE byte(s), _PyEval_EvalFrameDefault() + 0x40DE byte(s)
0x00007FF87ED4B201, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x4B201 byte(s), _PyGen_Finalize() + 0x389 byte(s)
0x00007FF87EF44F58, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x244F58 byte(s), PyGen_NewWithQualName() + 0x2C byte(s)
0x00007FF87EF3B57D, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x23B57D byte(s), PyIter_Send() + 0x35 byte(s)
0x00007FF8BD2B7334, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\DLLs\_asyncio.pyd(0x00007FF8BD2B0000) + 0x7334 byte(s), PyInit__asyncio() + 0x5AE4 byte(s)
0x00007FF8BD2B6C7E, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\DLLs\_asyncio.pyd(0x00007FF8BD2B0000) + 0x6C7E byte(s), PyInit__asyncio() + 0x542E byte(s)
0x00007FF87ED6560E, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x6560E byte(s), _PyObject_MakeTpCall() + 0xA2 byte(s)
0x00007FF87EF696E0, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x2696E0 byte(s), _PyContext_NewHamtForTests() + 0x80 byte(s)
0x00007FF87EF699C0, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x2699C0 byte(s), _PyContext_NewHamtForTests() + 0x360 byte(s)
0x00007FF87ED68376, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x68376 byte(s), PySet_Add() + 0xAC2 byte(s)
0x00007FF87EDB2815, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB2815 byte(s), PyObject_Call() + 0x125 byte(s)
0x00007FF87EDB275F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB275F byte(s), PyObject_Call() + 0x6F byte(s)
0x00007FF87ED866EE, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x866EE byte(s), _PyEval_EvalFrameDefault() + 0x40DE byte(s)
0x00007FF87ED8103C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x8103C byte(s), _PyFunction_Vectorcall() + 0x17C byte(s)
0x00007FF87ED7BE6C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x7BE6C byte(s), _PyArg_CheckPositional() + 0x4B0 byte(s)
0x00007FF87EDB2815, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB2815 byte(s), PyObject_Call() + 0x125 byte(s)
0x00007FF87EDB275F, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0xB275F byte(s), PyObject_Call() + 0x6F byte(s)
0x00007FF87ED2304C, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x2304C byte(s), PyInterpreterState_Delete() + 0x500 byte(s)
0x00007FF87ED22FF2, C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\python312.dll(0x00007FF87ED00000) + 0x22FF2 byte(s), PyInterpreterState_Delete() + 0x4A6 byte(s)
0x00007FF94E5837B0, C:\WINDOWS\System32\ucrtbase.dll(0x00007FF94E580000) + 0x37B0 byte(s), wcsrchr() + 0x150 byte(s)
0x00007FF94F53E8D7, C:\WINDOWS\System32\KERNEL32.DLL(0x00007FF94F510000) + 0x2E8D7 byte(s), BaseThreadInitThunk() + 0x17 byte(s)
0x00007FF950C48D9C, C:\WINDOWS\SYSTEM32\ntdll.dll(0x00007FF950C40000) + 0x8D9C byte(s), RtlUserThreadStart() + 0x2C byte(s)

--------
[Program Crash, Exit Code 3221225477 (0xC0000005)]

Physical Memory: 10,298 MB available / 16,384 MB total
Virtual Memory: 44,138 MB available / 48,223 MB total (initial 0 MB, maximum 0 MB)

Note: The application may have crashed due to running out of memory.
Your available memory might be insufficient for the current operation.
Try adding more RAM or increasing the size of the available virtual memory to keep at least 16 GB free.

Below is an analysis of the exit code. This may be inaccurate—use it for reference only!
System Exit Code Name: ACCESS_VIOLATION
System Exit Code Description: The instruction at 0x%p referenced memory at 0x%p. The memory could not be %s.

Crash Stack Trace:
Windows fatal exception: access violation

Stack (most recent call first):
File "D:\ComfyUI\venv\Lib\site-packages\torch_tensor.py", line 1669 in torch_function
File "D:\ComfyUI\custom_nodes\ComfyUI-GGUF\ops.py", line 58 in to
File "D:\ComfyUI\venv\Lib\site-packages\torch\nn\modules\module.py", line 1355 in convert
File "D:\ComfyUI\venv\Lib\site-packages\torch\nn\modules\module.py", line 955 in _apply
...
File "D:\ComfyUI\main.py", line 193 in prompt_worker
File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\Lib\threading.py", line 1012 in run
File "<enhanced_experience vendors.sentry_sdk.integrations.threading>", line 92 in _run_old_run_func
File "<enhanced_experience vendors.sentry_sdk.integrations.threading>", line 99 in run
File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\Lib\threading.py", line 1075 in _bootstrap_inner
File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.12_3.12.2800.0_x64__qbz5n2kfra8p0\Lib\threading.py", line 1032 in _bootstrap

Please note: the exit code and its description alone are usually insufficient to diagnose the problem!
Always upload the full log rather than just the exit code.
makisekurisu-jp changed discussion status to closed
makisekurisu-jp changed discussion status to open
QuantStack org

ACCESS_VIOLATION could be an anticheat or antivirus preventing access to the memory

QuantStack org

That shouldnt be a model issue, try disabling av for a test and if that was the issue, maybe do an exception for comfyui, though be careful what nodes you install (;

That shouldnt be a model issue, try disabling av for a test and if that was the issue, maybe do an exception for comfyui, though be careful what nodes you install (;

I’ve tried every possible fix, but nothing worked. I think the issue is that the refiner model is processing 2K-size images, which means it needs at least 32 GB of system memory or 16 GB of VRAM.

QuantStack org

that could be a reason too

I have a 5090 and can run this technically, it almost hits the VRAM limit (using the Q8 quants). However, something's wrong with the refiner it seems, I used the official workflow and only replaced the loader nodes with Unet loader (GGUF). The main model works fine and I get the image. When refiner loads there's a long error, presumably the keys are incorrectly named or converted:

2025-09-20_12-25-55.png
After that I get an exception: !!! Exception during processing !!! mat1 and mat2 shapes cannot be multiplied (16384x3328 and 9984x3328)

It ends with:

  File "/app/comfy/ldm/hunyuan_video/model.py", line 448, in _forward
    out = self.forward_orig(x, img_ids, context, txt_ids, attention_mask, timestep, y, txt_byt5, guidance, guiding_frame_index, ref_latent, disable_time_r=disable_time_r, control=con
trol, transformer_options=transformer_options)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/comfy/ldm/hunyuan_video/model.py", line 363, in forward_orig
    img, txt = block(img=img, txt=txt, vec=vec, pe=pe, attn_mask=attn_mask, modulation_dims_img=modulation_dims, modulation_dims_txt=modulation_dims_txt, transformer_options=transfor
mer_options)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/comfy/ldm/flux/layers.py", line 169, in forward
    img_qkv = self.img_attn.qkv(img_modulated)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1773, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1784, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/comfy/ops.py", line 113, in forward
    return self.forward_comfy_cast_weights(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/custom_nodes/ComfyUI-GGUF/ops.py", line 217, in forward_comfy_cast_weights
    out = super().forward_comfy_cast_weights(input, *args, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/comfy/ops.py", line 109, in forward_comfy_cast_weights
    return torch.nn.functional.linear(input, weight, bias)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: mat1 and mat2 shapes cannot be multiplied (16384x3328 and 9984x3328)

But I speculate that it's caused by the first error with key names.

QuantStack org

Thanks for the log, this actually looks like some issue, ill investigate (;

QuantStack org

Okay i might have found the issue, it seems that the model used in comfyui renamed some stuff (probably to distinguish), we didnt which causes some issue. Ill convert theirs which should then work fine (;

QuantStack org

Ill upload q2_k first (takes around 1h) so if you would test it then and tell me if it works that would be amazing (:

QuantStack org

The new q2k is online, should work now

Works now, no errors! Thank you.

when i first tried hunyuan image 2.1 the results were terrible. great complex prompt following, but resolution issues of base SDXL model level. With a sometimes working refiner not much better. so i gave up on it.

2 weeks later i just tried it again, same model, only some comfyui updates in betweeen... it is actually very good now! I don't know what happened there, but the results are on level or better with Qwen image. The prompt following is still great but people have actual eyes now.
As for the refiner when it doesn't cause oom error even on a 5090, it's just making things blurry. It is in my expereince useless. But now unnecessary

sorry to heat this topic up (i looked into it back then because the refiner was my last hope) , but something has happened to hunyuan image 2.1 and comfy to make it good.
Since 3.0 isn't available yet...

I've got a similar error with the distilled Q4 K_M: like that: "!!! Exception during processing !!! mat1 and mat2 shapes cannot be multiplied (1944x3584 and 10752x3584)"
No matter which input size I choose, it looks always the same.
The not distilled version "HunyuanImage2.1-Q4_K_M.gguf" works.

Sign up or log in to comment