Mohammed Ayub comments

Results 48 comments of


                                            Mohammed Ayub

可以正常训练，但是测试模型出现UnknownError (see above for traceback): Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.

@argman I get the same error. It started to train fine on CPU but since it was very slow, trying this on one GPU fails with the same stack trace....

可以正常训练，但是测试模型出现UnknownError (see above for traceback): Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.

Looks like this was a CuDNN issue which was popping up in the log ``` 2021-04-15 08:49:03.630044: E tensorflow/stream_executor/cuda/cuda_dnn.cc:319] Loaded runtime CuDNN library: 7.5.1 but source was compiled with: 7.6.0....

checkpoint has good accuracy, however frozen one is not very poor.

@kurikabocya I ran into the same issue. Apart from the C# code I'm just using python to load the frozen graph and test. Any update is you solved this. Thanks...

checkpoint has good accuracy, however frozen one is not very poor.

@kurikabocya That's interesting. Any reason you had to do resize to that specific dimensions (160x224) ?

checkpoint has good accuracy, however frozen one is not very poor.

Is this true for the pre-trained model from the authors as well or just the model you trained ?

Can I detect the rotation of the text? If so, how do I do that?

@joaossmacedo does the EAST model out of the box detect text region of any orientation or you had to make some changes to do it ? Currently, I have trained...

Can I detect the rotation of the text? If so, how do I do that?

@joaossmacedo Thanks. That's makes sense. I'll poke around a bit more.

Can I detect the rotation of the text? If so, how do I do that?

Here are couple of results using `eval.py` script on my trained model. It's very basic one using all defaults. (no changes made). The bounding boxes looks like they are not...

Does anyone know the output_node_names?

@argman I froze the graph this way and tried to load and test it. The frozen graph is giving really bad predictions. Is there a better way to freeze the...

[FR] Add image artifact to "Comparing 2 Runs" UI

Really interested in this feature too. Our project has a lot of object detection models that run against multiple test sets/edge cases (lets say each test set has 25 images)....