Unable to execute /bin/train_all.sh for Supervised embeddings
ahtsham58 opened this issue · 3 comments
I am in the process of reproducing the system. I am done with all the installations and now trying to execute supervised-embedding/bin/train_all.sh. I executed build_experiments.py and now trying to train the model but getting this error. I cannot see any .tsv files in my data. How to create these files ..?
I suspect there needs to run "build_vocabulary.py" file before training. I also tried to run this script but the console stuck there without execution till 30 mints.
Please let me know what I am missing here or is there any issue with the code?
looking forward to your positive response.
Hi @ahtsham58, thanks for the interest. Unfortunately, its been so long since I worked on this project that my memory is a bit hazy. Are you using the correct version of TensorFlow--0.12.1? (For each of the models, refer to README inside their directory for exact requirements, usage, etc.)
Here is the repository from which I adapted the supervised embeddings approach: https://github.com/sld/supervised-embedding-model
Refer also to other closed issues such as #3
Hope it helps!
Hi, I worked on the model, and there ware many path issues that I figured out. Now the model is getting trained for all the five tasks.
I hope it works.
Great!