R3DCNN

R3DCNN copied to clipboard

Error when extract dataset

1

Hello Sir, I downloaded the dataset from your Google Drive link, I used 7-zip to unzip but it showed "Can not open file as [7z] archive", and for winRAR it...

ghost

no_action label in test mode

8

Hi, in ActionRecognition.py, I tested some videos from nvgesture dataset(untrimmed video). And it outputs class no. in 0~24 in each frames. i.e. it didn't outputs any `blank` or `no_action` label....

asdfqwer2015

No matter how to adjust the superparameters, the loss is always between 2~3. The value of loss decays very quickly. Has anyone experienced this.

buaa-luzhi

@breadbread1984 Thanks for your great work! I wonder why the "class_num" is set to 25 instead of 26, namely class_num = predefined_classes + blank. Thanks

buaa-luzhi

@breadbread1984 Thanks for your great work! I wonder why the "class_num" is set to 25 instead of 26, namely class_num = predefined_classes + blank. Thanks

buaa-luzhi

thanks for your codes, i meet a error,when i run create_dataset.py, the error is keyerror: 'duo_left' in readdata.py ,do you know how to solve it ? thanks

world2025

Hello I just wanted to know if the pretrained model given in the other repository is trained on the NVDIA dataset ? To test should I just download it, put...

Omenranr

overfitting very fast?

2

Hi, sorry to bother you once again. But I'm very interested in your model. I found the model may overfitting in training very fast. describe: I tested a trained model,...

asdfqwer2015

evaluation accuracy

2

hi, the eval result is mean_edit_distance, could I use tf.metrics.accuracy to evaluate the accuracy? Could you please tell me how should I modify the code? eval_metric_ops = {"accuracy": tf.metrics.accuracy(labels,predictions)} it...

sunnymoon155

dataset

1

Sorry,I can't download the dataset,do you have any other download channels?thank you very much

shajie17