Skip to content Skip to sidebar Skip to footer

Invalidargumenterror When Traing Attention_ocr : Assign Requires Shapes Of Both Tensors To Match

I am working on Alex Gordan's Attention_ocr project. I have follow the guide and store my data in FSNS format acoording to Alex's answer. However, when I run command: python train.

Solution 1:

The code use a pretrained model stored in my tmp path, I clean up /tmp to solve it.


Post a Comment for "Invalidargumenterror When Traing Attention_ocr : Assign Requires Shapes Of Both Tensors To Match"