OSError: Error no file named diffusion_pytorch_model.bin found in directory D:\AI\ComfyUI_windows_portable\ComfyUI\models\CogVideo\CogVideoX-Fun-V1.1-5b-Control.

[logo66612]

模型自动下载了,为什么还是报错呢? models\CogVideo\CogVideoX-Fun-V1.1-5b-Control. Defaulting to unsafe serialization. Pass allow_pickle=False to raise an error instead. !!! Exception during processing !!! Error no file named diffusion_pytorch_model.bin found in directory D:\AI\ComfyUI_windows_portable\ComfyUI\models\CogVideo\CogVideoX-Fun-V1.1-5b-Control. Traceback (most recent call last): File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 323, in execute output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 198, in get_output_data return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 169, in _map_node_over_list process_inputs(input_dict, i) File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 158, in process_inputs results.append(getattr(obj, func)(inputs)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-CogVideoXWrapper\model_loading.py", line 230, in loadmodel vae = AutoencoderKLCogVideoX.from_pretrained(base_path, subfolder="vae").to(dtype).to(offload_device) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\huggingface_hub\utils_validators.py", line 114, in _inner_fn return fn(*args, *kwargs) ^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\diffusers\models\modeling_utils.py", line 804, in from_pretrained model_file = _get_model_file( ^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\huggingface_hub\utils_validators.py", line 114, in _inner_fn return fn(args, kwargs) ^^^^^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\diffusers\utils\hub_utils.py", line 309, in _get_model_file raise EnvironmentError( OSError: Error no file named diffusion_pytorch_model.bin found in directory D:\AI\ComfyUI_windows_portable\ComfyUI\models\CogVideo\CogVideoX-Fun-V1.1-5b-Control.

[KrakeyMTL]

similar errors have already been solved here in this very recent closed thread. -> https://github.com/kijai/ComfyUI-CogVideoXWrapper/issues/269

Essentially it didn't finish downloading for whatever reason. It happens randomly. Manually put the model.bin into the proper folder it is complaining about above in your error (path is there), regardless what model you are using. OR clear the workspace, reload the script AFTER updating, it will refresh the nodes and auto-download again once refreshed. Both methods work.

[GeffersAtWork]

Can I respectfully point out that the issue was the example workflow in github for 1.5 image to video has the old node in it. To solve I had to add the new load cog model node, and reconnect it. Might it be possible to update the example workflow?

[logo66612]

在最近关闭的这个线程中，类似的错误已经得到解决。-> #269

本质上，无论出于什么原因，它都没有完成下载。这是随机发生的。手动将 model.bin 放入上面错误中抱怨的正确文件夹中（路径在那里），无论您使用的是什么模型。或者清除工作区，在更新后重新加载脚本，它将刷新节点并在刷新后再次自动下载。这两种方法都有效。

昨晚跑一晚上都没跑起来,是不是20系显卡跑不起来?

[kijai]

Can I respectfully point out that the issue was the example workflow in github for 1.5 image to video has the old node in it. To solve I had to add the new load cog model node, and reconnect it. Might it be possible to update the example workflow?

It's not that the download node never works, it just sometimes fails as it relies on the Huggingface downloader. Most of the time it still works, the new node is an alternative way which I probably will make standard going forward, but for now not all the models are even converted and that's why the download node is still there by default.

Your point stands though, the example could have both options as it's not otherwise clear it even exists (if you don't read the update notes).