Closed MuireannMacCarthy closed 5 years ago
Problem is resolved for me, if you received similar error continue reading for solution.
g2p-seq2seq --train cmudict.dict --model_dir /c/Users/mmacc/OneDrive/Documents/g2p-seq2seq/ --p2g --cleanup
The cmudict.dict reference from the above command is stored in the same directory from which the command was executed so I did not include entire file path. However replacing cmudict.dict with its complete path worked for me. See below:
g2p-seq2seq --train /path/to/cmudict.dict --model_dir /c/Users/mmacc/OneDrive/Documents/g2p-seq2seq/ --p2g --cleanup
hi,I ran into the same issue.
However when I fed the whole path of cmudict.dict, the error still occured.
The command line I used is:
g2p-seq2seq --train /data00/home/***/anaconda3/bin/cmudict.dict --model_dir g2p-seq2seq-model-6.2-cmudict-nostress --p2g --cleanup
The error is
WARNING:tensorflow:Invalid symbol:zavada WARNING:tensorflow:Invalid symbol:A WARNING:tensorflow:Invalid symbol:H WARNING:tensorflow:Invalid symbol:z
Hi,
Running the G2P in interactive mode works perfectly but I'm having issues using --p2g to solve the problem in reverse.
I'm training the model on the CMU dict linked on this repo underneath the "Training G2P" section.
However when I run the command below:
g2p-seq2seq --train cmudict.dict --model_dir /c/Users/mmacc/OneDrive/Documents/g2p-seq2seq/ --p2g --cleanup
I get all sorts of errors, see below:
Any suggestions on where I'm going wrong would be very helpful,
Thanks