Mask_rcnn: OSError: Unable to open file (truncated file: eof = 74436151, sblock->base_addr = 0, stored_eof = 257557808)

[farzadzare]

I want to use Mask-RCNN demo.py in my jupyter notebook but when i ran the the demo code , i faced such a problem
OSError Traceback (most recent call last)
in ()
3
4 # Load weights trained on MS-COCO
----> 5 model.load_weights(COCO_MODEL_PATH, by_name=True)

~/Mask_RCNN/mrcnn/model.py in load_weights(self, filepath, by_name, exclude)
2102 if h5py is None:
2103 raise ImportError('load_weights requires h5py.')
-> 2104 f = h5py.File(filepath, mode='r')
2105 if 'layer_names' not in f.attrs and 'model_weights' in f:
2106 f = f['model_weights']

~/anaconda3/envs/MaskRCNN/lib/python3.6/site-packages/h5py/_hl/files.py in __init__(self, name, mode, driver, libver, userblock_size, swmr, *kwds)
310 with phil:
311 fapl = make_fapl(driver, libver, *kwds)
--> 312 fid = make_fid(name, mode, userblock_size, fapl, swmr=swmr)
313
314 if swmr_support:

~/anaconda3/envs/MaskRCNN/lib/python3.6/site-packages/h5py/_hl/files.py in make_fid(name, mode, userblock_size, fapl, fcpl, swmr)
140 if swmr and swmr_support:
141 flags |= h5f.ACC_SWMR_READ
--> 142 fid = h5f.open(name, flags, fapl=fapl)
143 elif mode == 'r+':
144 fid = h5f.open(name, h5f.ACC_RDWR, fapl=fapl)

h5py/_objects.pyx in h5py._objects.with_phil.wrapper()

h5py/_objects.pyx in h5py._objects.with_phil.wrapper()

h5py/h5f.pyx in h5py.h5f.open()

OSError: Unable to open file (truncated file: eof = 74436151, sblock->base_addr = 0, stored_eof = 257557808)

i can not solve my problem
can any body help me??
thank's

[WindowsDriver]

you are sopposed to download the .h5 file.

[farzadzare]

if you mean download the mask_rcnn_coco.h5 file
i had downloaded it but i faced this error
AttributeError Traceback (most recent call last)
in ()
3
4 # Load weights trained on MS-COCO
----> 5 model.load_weights(COCO_MODEL_PATH, by_name=True)

~/Downloads/Mask_RCNN/mrcnn/model.py in load_weights(self, filepath, by_name, exclude)
2117
2118 if by_name:
-> 2119 topology.load_weights_from_hdf5_group_by_name(f, layers)
2120 else:
2121 topology.load_weights_from_hdf5_group(f, layers)

AttributeError: module 'keras.engine.topology' has no attribute 'load_weights_from_hdf5_group_by_name'

[freedom521jin]

I have the same error,do you solve the problem?

[farzadzare]

@freedom521jin,,,,, I coudn't solve it but i deleted all of anaconda and reinstall all of them and it solved

[freedom521jin]

@farzadzare ....oh!I have made it.I delete the old model ' **.h5 ' , and re-download the pre-train model ,and it works.

[CrookedNoob]

@freedom521jin It worked

[vajpaye]

@freedom521jin can u please tell me where is .h5 file is located

[DerOzean]

I had the same Problem.
I coud fix it by changing the ROOT_DIR (the path to root) in balloon.py.
In the Example it is given by ROOT_DIR = '../../'
In my case it is ROOT_DIR='../../../'
So I just had to go on step futher.

[vajpaye]

TQ @DerOzean

COCO_MODEL_PATH = os.path.join(ROOT_DIR, "mask_rcnn_balloon.h5")
print(COCO_MODEL_PATH)

Download COCO trained weights from Releases if needed

if not os.path.exists(COCO_MODEL_PATH):
utils.download_trained_weights(COCO_MODEL_PATH)-> this is the responsible for downloading mask_rcnn_ballon.h5 or mask_rcnn_coco.h5

[yiakwy]

I have the same problem each time I load coco pretrained model after I clean the data. (rm mask_rcnn_coco.h5). Is the pretrained model corrupted ?

[yiakwy]

@farzadzare Have you resolve the OSError problem?

[AniPrat]

I have the same issue of OSError in h5py.h5f.open(). I am using windows python 3.6 and latest h5py version.
Any Solution is helpful

[AniPrat]

I found the solution. The .h5 file was not downloaded completely so it was corrupt. I downloaded it manually by going to the model release page.

[pissw2016]

I have the same problem, and inspired by you guy. Mine is because that I am training with coco.h5 and I copy this weight file to another dir sothat my demo.ipynb can read it. After I load the original version on demo.ipynb it worked perfect.

[sliawatimena]

OSError: Unable to open file (truncated file: eof = 218,644,480, sblock->base_addr = 0, stored_eof = 257,557,808)

the solution is:
wget -c https://github.com/matterport/Mask_RCNN/releases/download/v2.0/mask_rcnn_coco.h5

[farabee22]

OSError: Unable to open file (truncated file: eof = 286720, sblock->base_addr = 0, stored_eof = 17225924)

[YeShangguan]

I've got the same error. And I checked the file name. I've downloaded the wrong file mask_rcnn_ballon.h5. Then I downloaded the right file mask_rcnn_coco.h5, it worked smoothly.

[yiakwy]

We solved the problem by downloading the file from microsoft coco website
manually. Actually you can use wget or curl to download it directly.

YeShangguan [e-mail] 于2019年11月26日周二 下午4:05写道：

I've got the same error. And I checked the file name. I've downloaded the
wrong file mask_rcnn_ballon.h5. Then I downloaded the right file
mask_rcnn_coco.h5, it worked smoothly.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/matterport/Mask_RCNN/issues/653?email_source=notifications&email_token=ACA526AVXD7Y3GZPE2KI3MLQVTKEVA5CNFSM4FDYAE22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFFDALY#issuecomment-558510127,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ACA526ATA4ZQT4W4WKXU6JTQVTKEVANCNFSM4FDYAE2Q
.

[yiakwy]

The thread should be marked as closed because this was resolved many many
months ago.

Yiak Wang Yi [e-mail] 于2019年12月3日周二 下午11:13写道：

We solved the problem by downloading the file from microsoft coco website
manually. Actually you can use wget or curl to download it directly.

YeShangguan [e-mail] 于2019年11月26日周二 下午4:05写道：

I've got the same error. And I checked the file name. I've downloaded the
wrong file mask_rcnn_ballon.h5. Then I downloaded the right file
mask_rcnn_coco.h5, it worked smoothly.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/matterport/Mask_RCNN/issues/653?email_source=notifications&email_token=ACA526AVXD7Y3GZPE2KI3MLQVTKEVA5CNFSM4FDYAE22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFFDALY#issuecomment-558510127,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/ACA526ATA4ZQT4W4WKXU6JTQVTKEVANCNFSM4FDYAE2Q
.

[Omar-Belghaouti]

It is solved by deleting the old h5 and redownload it again

[shghahremani]

Solution!
You need to delet the corrupted h5 file first. If you want to see where it is located, jut print the path directly as follow;

VGG_Weights_path = keras.utils.get_file(
pretrained_url.split("/")[-1], pretrained_url)
print(VGG_Weights_path)