Model downloading problems: HuggingFace connection error

[xpixelgroup]

Model downloading problems: HuggingFace connection error

[ziyannchen]

Hey guys, all the related model should be downloaded automatically if you can access huggingface website.

However, if you stuck in huggingface model downloading for the server's internet aceess problem or you require to run the model offline, etc., any model-downloading problems, please refer to the following checklist for manually model downloading guidance.

Model downloading checklist:

• download open_clip_pytorch_model.bin to directory ~/.cache/huggingface/hub/ (this is consistant to the automatic script downloaing path, you can assign any other paths), then add a line of your clip model path into the config file configs/model/cldm.yaml as

84 cond_stage_config:
85    target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
86    params:
+       version: "${HOME}/.cache/huggingface/hub/open_clip_pytorch_model.bin"
87      freeze: True
88      layer: "penultimate"

• download alexnet-owt-7be5be79.pth to path ~/.cache/torch/hub/checkpoints；
• download DiffBIR face model face_full_v1.ckpt into project directory weights/；
• if you need to use background upsampler, please download DiffBIR general model general_full_v1.ckpt and general_swinir_v1.ckpt into project directory weights/; (if you wanna use RealESRGAN as background upsampler, please download RealESRGANx2/x4 to weights/realesrgan/);

（Here we provide all huggingface download links for DiffBIR models, in case of problems like internet access, you can also check the README for other servers BaiduCloud and OpenXLab for downloading as well）

Duplicate with #81 .

[bjlybjb]

@bjlybjb Hello! Could you please share which target models you have downloaded? If you can provide URLs, that would be even better. I believe most Chinese users will encounter this issue.

• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main I have downloaded open_clip_pytorch_model.bin. And then I found the file “~/DiffBIR-main/ldm/modules/encoders/modules.py” according to traceback and changed the parameter in line 144. I‘m not sure whether it is absolutely correct. But at least the error never occur again.

[nathubs]

the same to me

[ziyannchen]

This seems a HuggingFace connection error. The script downloads pre-trained models from HuggingFace automatically.
Try visiting HuggingFace and see if you can access it.
If you are working on a remote server, an alternative is to download the target model to a local device and then upload it to the server.

[bjlybjb]

This seems a HuggingFace connection error. The script downloads pre-trained models from HuggingFace automatically. Try visiting HuggingFace and see if you can access it. If you are working on a remote server, an alternative is to download the target model to a local device and then upload it to the server.

Thanks for replying to my question. I did run it on a remote server. Now I tried to download the target model manually and then successfully solved the problem. Thank you very much！

[0x3f3f3f3fun]

@bjlybjb Hello! Could you please share which target models you have downloaded? If you can provide URLs, that would be even better. I believe most Chinese users will encounter this issue.

[24wenjie-li]

This seems a HuggingFace connection error. The script downloads pre-trained models from HuggingFace automatically. Try visiting HuggingFace and see if you can access it. If you are working on a remote server, an alternative is to download the target model to a local device and then upload it to the server.

Can you tell me the location of the folder to upload the downloaded files to?

[24wenjie-li]

这似乎是一个拥抱脸连接错误。该脚本会自动从HuggingFace下载预先训练的模型。尝试访问HuggingFace，看看是否可以访问它。如果您在远程服务器上工作，另一种方法是将目标模型下载到本地设备，然后将其上传到服务器。

感谢您回复我的问题。我确实在远程服务器上运行了它。现在我尝试手动下载目标模型，然后成功解决问题。谢谢！

Can you tell me to which folder to upload the downloaded file? How do I change the parameter in line 144?

[bjlybjb]

这似乎是一个拥抱脸连接错误。该脚本会自动从HuggingFace下载预先训练的模型。尝试访问HuggingFace，看看是否可以访问它。如果您在远程服务器上工作，另一种方法是将目标模型下载到本地设备，然后将其上传到服务器。

感谢您回复我的问题。我确实在远程服务器上运行了它。现在我尝试手动下载目标模型，然后成功解决问题。谢谢！

Can you tell me to which folder to upload the downloaded file? How do I change the parameter in line 144?

Change the parameter “version” to the path where you upload. You can upload the file to everywhere you want.

[24wenjie-li]

这似乎是一个拥抱脸连接错误。该脚本会自动从HuggingFace下载预先训练的模型。尝试访问HuggingFace，看看是否可以访问它。如果您在远程服务器上工作，另一种方法是将目标模型下载到本地设备，然后将其上传到服务器。

感谢您回复我的问题。我确实在远程服务器上运行了它。现在我尝试手动下载目标模型，然后成功解决问题。谢谢！

Can you tell me to which folder to upload the downloaded file? How do I change the parameter in line 144?

Change the parameter “version” to the path where you upload. You can upload the file to everywhere you want.

Thank you！

[wuxiaoxue1006]

@bjlybjb Hello, I used your method but reported this error, may I ask you have encountered this error？
RuntimeError: PytorchStreamReader failed reading zip archive: invalid header or archive is corrupted

[iruhh]

These links may be helpful.
mlfoundations/open_clip#622
mlfoundations/open_clip#528

[wiqe1314]

这似乎是一个拥抱脸连接错误。该脚本会自动从HuggingFace下载预先训练的模型。尝试访问HuggingFace，看看是否可以访问它。如果您在远程服务器上工作，另一种方法是将目标模型下载到本地设备，然后将其上传到服务器。

感谢您回复我的问题。我确实在远程服务器上运行了它。现在我尝试手动下载目标模型，然后成功解决问题。谢谢！

Can you tell me to which folder to upload the downloaded file? How do I change the parameter in line 144?

Change the parameter “version” to the path where you upload. You can upload the file to everywhere you want.

Thank you！

我能看一下你修改后的144行代码吗？

[wiqe1314]

@bjlybjb你好!你能分享一下你下载的目标型号吗?如果你能提供URL,那就更好了。我相信大多数**用户会遇到这个问题。

• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main 我下载了开放的.然后我找到了一个文件"~/扩散主/LDM/模块/编码器/模块。我不确定它是否完全正确。但至少这个错误再也不会发生了。

@bjlybjb Hello! Could you please share which target models you have downloaded? If you can provide URLs, that would be even better. I believe most Chinese users will encounter this issue.

• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main I have downloaded open_clip_pytorch_model.bin. And then I found the file “~/DiffBIR-main/ldm/modules/encoders/modules.py” according to traceback and changed the parameter in line 144. I‘m not sure whether it is absolutely correct. But at least the error never occur again.

@bjlybjb I am a beginner in programming and tried to deploy Diffbir on my home win platform. I also encountered the problem that huggingface could not connect when starting.I tried to download the model locally,
but I don't know how to apply it and where should it be stored?
How should line 144 in the /modules.py file you mentioned be modified?
I am a newbie, can you provide a modified code of line 144 for reference?

[catsled]

@bjlybjb你好!你能分享一下你下载的目标型号吗?如果你能提供URL,那就更好了。我相信大多数**用户会遇到这个问题。

• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main 我下载了开放的.然后我找到了一个文件"~/扩散主/LDM/模块/编码器/模块。我不确定它是否完全正确。但至少这个错误再也不会发生了。

@bjlybjb Hello! Could you please share which target models you have downloaded? If you can provide URLs, that would be even better. I believe most Chinese users will encounter this issue.

• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main I have downloaded open_clip_pytorch_model.bin. And then I found the file “~/DiffBIR-main/ldm/modules/encoders/modules.py” according to traceback and changed the parameter in line 144. I‘m not sure whether it is absolutely correct. But at least the error never occur again.

@bjlybjb I am a beginner in programming and tried to deploy Diffbir on my home win platform. I also encountered the problem that huggingface could not connect when starting.I tried to download the model locally, but I don't know how to apply it and where should it be stored? How should line 144 in the /modules.py file you mentioned be modified? I am a newbie, can you provide a modified code of line 144 for reference?

[wiqe1314]

Hey guys, all the related model should be downloaded automatically if you can access huggingface website.

However, if you stuck in huggingface model downloading for the server's internet aceess problem or you require to run the model offline, etc., any model-downloading problems, please refer to the following checklist for manually model downloading guidance.

Model downloading checklist:

• download open_clip_pytorch_model.bin to directory ~/.cache/huggingface/hub/ (this is consistant to the automatic script downloaing path, you can assign any other paths), then add a line of your clip model path into the config file configs/model/cldm.yaml as

84 cond_stage_config:
85    target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
86    params:
+       version: "${HOME}/.cache/huggingface/hub/open_clip_pytorch_model.bin"
87      freeze: True
88      layer: "penultimate"

• download alexnet-owt-7be5be79.pth to path ~/.cache/torch/hub/checkpoints；
• download DiffBIR face model face_full_v1.ckpt into project directory weights/；
• if you need to use background upsampler, please download DiffBIR general model general_full_v1.ckpt and general_swinir_v1.ckpt into project directory weights/; (if you wanna use RealESRGAN as background upsampler, please download RealESRGANx2/x4 to weights/realesrgan/);

（Here we provide all huggingface download links for DiffBIR models, in case of problems like internet access, you can also check the README for other servers BaiduCloud and OpenXLab for downloading as well）

Duplicate with #81 .

thank you！

[wiqe1314]

Hey guys, all the related model should be downloaded automatically if you can access huggingface website.
However, if you stuck in huggingface model downloading for the server's internet aceess problem or you require to run the model offline, etc., any model-downloading problems, please refer to the following checklist for manually model downloading guidance.
Model downloading checklist:

• download open_clip_pytorch_model.bin to directory ~/.cache/huggingface/hub/ (this is consistant to the automatic script downloaing path, you can assign any other paths), then add a line of your clip model path into the config file configs/model/cldm.yaml as

84 cond_stage_config:
85    target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
86    params:
+       version: "${HOME}/.cache/huggingface/hub/open_clip_pytorch_model.bin"
87      freeze: True
88      layer: "penultimate"

• download alexnet-owt-7be5be79.pth to path ~/.cache/torch/hub/checkpoints；
• download DiffBIR face model face_full_v1.ckpt into project directory weights/；
• if you need to use background upsampler, please download DiffBIR general model general_full_v1.ckpt and general_swinir_v1.ckpt into project directory weights/; (if you wanna use RealESRGAN as background upsampler, please download RealESRGANx2/x4 to weights/realesrgan/);

（Here we provide all huggingface download links for DiffBIR models, in case of problems like internet access, you can also check the README for other servers BaiduCloud and OpenXLab for downloading as well）
Duplicate with #81 .

thank you！
I have successfully run

[baiguiang]

Hey guys, all the related model should be downloaded automatically if you can access huggingface website.

However, if you stuck in huggingface model downloading for the server's internet aceess problem or you require to run the model offline, etc., any model-downloading problems, please refer to the following checklist for manually model downloading guidance.

Model downloading checklist:

• download open_clip_pytorch_model.bin to directory ~/.cache/huggingface/hub/ (this is consistant to the automatic script downloaing path, you can assign any other paths), then add a line of your clip model path into the config file configs/model/cldm.yaml as

84 cond_stage_config:
85    target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
86    params:
+       version: "${HOME}/.cache/huggingface/hub/open_clip_pytorch_model.bin"
87      freeze: True
88      layer: "penultimate"

• download alexnet-owt-7be5be79.pth to path ~/.cache/torch/hub/checkpoints；
• download DiffBIR face model face_full_v1.ckpt into project directory weights/；
• if you need to use background upsampler, please download DiffBIR general model general_full_v1.ckpt and general_swinir_v1.ckpt into project directory weights/; (if you wanna use RealESRGAN as background upsampler, please download RealESRGANx2/x4 to weights/realesrgan/);

（Here we provide all huggingface download links for DiffBIR models, in case of problems like internet access, you can also check the README for other servers BaiduCloud and OpenXLab for downloading as well）

Duplicate with #81 .

hi，首先谢谢您的解答，我按照您的方式去解决这个问题，但是仍然报错
RuntimeError: Pretrained weights ($/home/hjt/下载/wy/DiffBIR-main/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k'].
您可以帮忙解答吗，谢谢。

[ziyannchen]

hi，首先谢谢您的解答，我按照您的方式去解决这个问题，但是仍然报错 RuntimeError: Pretrained weights ($/home/hjt/下载/wy/DiffBIR-main/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k']. 您可以帮忙解答吗，谢谢。

看起来是CLIP模型文件没有下载成功，可以试着重新下载一下

[baiguiang]

hi，首先谢谢您的解答，我按照您的方式去解决这个问题，但是仍然报错 RuntimeError: Pretrained weights ($/home/hjt/下载/wy/DiffBIR-main/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k']. 您可以帮忙解答吗，谢谢。

看起来是CLIP模型文件没有下载成功，可以试着重新下载一下

谢谢回复，但我尝试了重新下载这个文件，仍然是同样的错误，我是从镜像网站下载的，这会有影响吗？

[ziyannchen]

hi，首先谢谢您的解答，我按照您的方式去解决这个问题，但是仍然报错 RuntimeError: Pretrained weights ($/home/hjt/下载/wy/DiffBIR-main/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k']. 您可以帮忙解答吗，谢谢。

看起来是CLIP模型文件没有下载成功，可以试着重新下载一下

谢谢回复，但我尝试了重新下载这个文件，仍然是同样的错误，我是从镜像网站下载的，这会有影响吗？

应该没有影响。不过看起来你的路径“$/home/hjt/下载/”似乎有点问题，可以检查一下。

[baiguiang]

hi，首先谢谢您的解答，我按照您的方式去解决这个问题，但是仍然报错 RuntimeError: Pretrained weights ($/home/hjt/下载/wy/DiffBIR-main/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k']. 您可以帮忙解答吗，谢谢。

看起来是CLIP模型文件没有下载成功，可以试着重新下载一下

谢谢回复，但我尝试了重新下载这个文件，仍然是同样的错误，我是从镜像网站下载的，这会有影响吗？

应该没有影响。不过看起来你的路径“$/home/hjt/下载/”似乎有点问题，可以检查一下。

我尝试更改了模型位置，仍然有同样的错误Pretrained weights ($/home/hjt/.cache/huggingface/hub/open_clip_pytorch_model.bin) not found for model ViT-H-14. Available pretrained tags (['laion2b_s32b_b79k'].
您能在帮我看看这个路径正确吗，可以怎么修改呢？