Rohit Gupta
Rohit Gupta
Hi @tjdwo1289, may I know what version of Tensorflow and Keras you are using, so I can reproduce the error ? Also do you have a GPU ?
@tjdwo1289 This is a known issue with Tensorflow 1.3, other users have solved it by upgrading to TF 1.5 or newer. Please refer to this: https://github.com/rohit-gupta/Video2Language/issues/3#issuecomment-372096811
@roycedry How much RAM do you have ? I haven't tried it with a very low RAM machine, but it works fine if you have 64GB of RAM, usually.
@roycedry the file will be generated when you run the model on MSVD dataset, this is the results file.
What video are you running it on ? This shouldn't happen
@isbada Well, there's a copy of my Masters' Thesis. https://github.com/rohit-gupta/Video2Language/blob/master/Rohit_Gupta_Thesis__Public_Copy_.pdf
Hi @SandeepRddy, accuracy is not a good metric to judge a captioning model, please use the METEOR score.
Oh, seems like you need the vocabulary file. You can create it by using: ``` bash fetch-data.sh bash preprocess-data.sh CleanOnly ```
@vaibh1297 Give me a day or so, I will upload the `vocabulary_10.p` file that I generated along with the pre-trained model.
@siddish-reddy unfortunately I no longer have access to it, as it was in a AWS bucket that has since been deleted. You can still use the code for training, though...