flying-pigeon
flying-pigeon
 When I'm running torch_client.py, there exists a question related to boto3 and S3. How to solve it?
训练效果问题
目前在DuIE上用代码跑了,跑了28轮,到后面验证集上的准确率和f1值反而会下降,整体看来验证集上的效果波动比较大,请问这是正常现象吗
In file fl_dataloader.py, the definiton of local_dataloader has changed to one argument "args". However, in the file app.py (related to RaspberryPi) the usage of local_dataloader is still: train_data_num, test_data_num, train_data_global,...