I have a small doubt, in the novel fine-tuning stage in run_voc.sh script, could you kindly tell why are you removing the model_final.pth file. If you are removed this, you would lose the trained weights. But, still in evaluation you get nice result, how is this possible? I am really confused. Any help from anyone would be really appreciated.
Hello @er-muyue,
I have a small doubt, in the novel fine-tuning stage in run_voc.sh script, could you kindly tell why are you removing the model_final.pth file. If you are removed this, you would lose the trained weights. But, still in evaluation you get nice result, how is this possible? I am really confused. Any help from anyone would be really appreciated.
DeFRCN/run_voc.sh Line number: 38