Code Size Too Large [Solved]

letmeetran 2019-06-25 21:59:34 UTC #1

Hello, I am working on a project using Wavenet where I am currently attempting to train the machine with a 3-second clip of Michael Jackson's vocals. Initially, I was able to successfully train and generate an output, however when I tried to repeat the training process with new parameters, I was hit with a "Code size too large, please limit size to 150 mb" error. I have uploaded a database but I am unclear as to how I am supposed to access it when I run the code.

This is the picture of what I am talking about ^^. --data_dir=mj is telling the computer that the Michael Jackson voice sample is under the "mj" folder of my home location. Can someone please guide me to the right direction? Thank you!

redeipirati 2019-06-26 09:32:03 UTC #2

Hi @letmeetran,

Please check this help article and let me know if you have any questions or need more help.

letmeetran 2019-06-26 19:16:08 UTC #3

Hello @redeipirati, I have uploaded the database but I am confused as to how I can use it locally since my code requires the data to be in the same root directory/folder as my train.py file. How do I move the database that has been uploaded separately from the code into the same folder as my train.py file?

redeipirati 2019-06-27 08:14:05 UTC #4

How do I move the database that has been uploaded separately from the code into the same folder as my train.py file?

You can mount the dataset inside your Workspace or Job, then you can use --data_dir argument of train.py to retrieve your data. You can also find this FAQ useful to retrieve the path to your data file.

Hope that helps.