search

multimodallearning / pytorch-mask-rcnn

Other
2.01k stars 557 forks source link

UnpicklingError: invalid load key, 'H' #79

Open HsiaX opened 5 years ago

HsiaX commented 5 years ago

I met an error information like this:

File "/usr/local/lib/python3.5/dist-packages/torch/serialization.py", line 399, in _load magic_number = pickle_module.load(f)

UnpicklingError: invalid load key, 'H'

Is there anybody can tell me how to solve it??? Thanks very muck!!!

HsiaX commented 5 years ago

Aha I've solved it. I download a ".h5" pre-trained model which is not match the pytorch, then I download the ".pth" file and it can work now! Anyway thanks all of you :-D

zhangbin01 commented 5 years ago

hello, i have the same problem as you. And thank you to solve the problem, but i can not find the '.pth' folder, can you give me one?

HsiaX commented 5 years ago

hello, i have the same problem as you. And thank you to solve the problem, but i can not find the '.pth' folder, can you give me one?

I download it from Google, to prevent that you may unable to download, I deposited it on Baidu Netdisk， here is the link: https://pan.baidu.com/s/1A2hz5FWemsZWp-qi7U_abw, and password: 340c.

zhangbin01 commented 5 years ago

wow~ ⊙o⊙，thank you very much!

zhangbin01 commented 5 years ago

hello, the files you shared with me is 'mask_rcnn_coco.h5', not 'mask_rcnn_coco.pth'. but i find the file ('mask_rcnn_coco.pth'). Do you need it? 你好， 您分享的是 mask_rcnn_coco.h5 ，不是mask_rcnn_coco.pth文件，但是我已经找到了pth文件，请问您还需要吗？我可以分享您一份。

HsiaX commented 5 years ago

hello, the files you shared with me is 'mask_rcnn_coco.h5', not 'mask_rcnn_coco.pth'. but i find the file ('mask_rcnn_coco.pth'). Do you need it? 你好， 您分享的是 mask_rcnn_coco.h5 ，不是mask_rcnn_coco.pth文件，但是我已经找到了pth文件，请问您还需要吗？我可以分享您一份。

really??? i'm sure the file i deposited is '.pth', it's so strange. no need thanks, i already have it (´･ᴗ･`)

dppselena commented 3 years ago

@zhangbin01 您好，您现在还保存有pth的预训练模型，可以分享一份吗？感谢您，盼复！

jf0391 commented 2 years ago

@HsiaX 你好，还有.pth的预训练模型么？希望可以百度网盘分享以下，谢谢，期待你的回复