# Huggingface Pytorch\_model.bin download corrupted **URL:** \ **Category:** 🤗Hub\ **Created:** [October 25, 2022, 7:50pm UTC](https://discuss.huggingface.co/t/huggingface-pytorch-model-bin-download-corrupted/24971 "2022-10-25T19:50:08Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![rexwang8](https://avatars.discourse-cdn.com/v4/letter/r/bcef8e/32.png) [@rexwang8](https://discuss.huggingface.co/u/rexwang8)\ **Post date:** [October 25, 2022, 7:50pm UTC](https://discuss.huggingface.co/t/huggingface-pytorch-model-bin-download-corrupted/24971/1 "2022-10-25T19:50:08Z")
I uploaded a model to huggingface, it has a pytorch\_model.bin that is 12.2gb large, the sha256 is the same for website and locally but when i download it again, it is 11.7gb sha256 is different and its unopenable. I can confirm this behavior happens on a direct download, in kobold-ai colab and kobold-ai local. I couldn’t find anyone else having this issue, any ideas? ---
**Author:** ![sparta351988](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/sparta351988/32/43153_2.png) [@sparta351988](https://discuss.huggingface.co/u/sparta351988)\ **Post date:** [March 13, 2025, 3:32am UTC](https://discuss.huggingface.co/t/huggingface-pytorch-model-bin-download-corrupted/24971/2 "2025-03-13T03:32:46Z")
I have also faced this issue multiple times… possible solutions: **Git LFS (Large File Storage)**: - Hugging Face uses Git LFS for large files. Ensure you have Git LFS installed and configured properly. You can check by running `git lfs install` and then cloning the repository again. - - Hugging Face provides a faster transfer tool called `hf_transfer`. You can enable it by installing the `huggingface_hub` library and setting the environment variable `HF_HUB_ENABLE_HF_TRANSFER=1`. This might help with large file downloads. - - Instead of downloading directly, try using the Hugging Face Python library: python ```auto from huggingface_hub import hf_hub_download hf_hub_download(repo_id="your_repo_id", filename="pytorch_model.bin") ``` Replace `"your_repo_id"` with the actual repository ID.