Closed YufanPaPa closed 4 years ago
This is a batch of data, not the whole data. You can print the batch data to see if there is some errors in data_load.py
I have the same question...
In data_load.py, the code likes what @YufanPaPa said. Can you explain about this? Thank you.
This is a batch of data, not the whole data. You can print the batch data to see if there is some errors in data_load.py
But in the code, you have use load_train_data(), which will return the whole processed dataset.
We conduct the Ubuntu experiment follow your pipeline, but we get "ValueError: Cannot create a tensor proto whose content is larger than 2GB". We found that you convert the whole training data to tensor in load_data.py:
We are sure that the Ubuntu dataset is mush larger than 2GB, so we are confused how did you do the Ubuntu experiment?