Knowledge-Distillation
Knowledge-Distillation copied to clipboard
ujjwal-9
Blog https://medium.com/neuralmachine/knowledge-distillation-dc241d7c2322
Results
2
Knowledge-Distillation issues
Sort by
recently updated
recently updated
newest added
The official implementation uses KL-Div loss while your implementation seems to use categorical cross entropy loss of Keras. Using the latter would completely invalidate the use of soft predictions in...
It seems that the data is pulled from /datatrain and /dataval, how should the data be organized when placed in this folder for the data to be properly loaded and...
