Issues
- 0
add activate graph
#97 opened by Unqua - 0
add dsn_self_training
#95 opened by Unqua - 3
fix Data_sample.set_epoch(epoch) bug
#94 opened by Unqua - 3
- 2
change apex/amp to torch.cuda.amp
#63 opened by Unqua - 4
update dsn+revnet
#90 opened by Unqua - 4
与TSTNet中复现的差别
#4 opened by Unqua - 0
add more dataset
#87 opened by Unqua - 0
add 3d cnn
#86 opened by Unqua - 0
try gate in dsn
#85 opened by Unqua - 0
save rgb image
#83 opened by Unqua - 0
- 0
- 0
- 0
add domain specific and inbvariant
#76 opened by Unqua - 0
- 1
add domain invariant and domain specific
#70 opened by Unqua - 0
fix bugs in mmoe
#71 opened by Unqua - 1
add fft preprocess
#68 opened by Unqua - 8
- 0
整理下configs
#62 opened by Unqua - 0
- 0
统一下writer内容
#57 opened by Unqua - 0
apex.amp is deprecated and will be removed by the end of February 2023. Use [PyTorch AMP] instead
#56 opened by Unqua - 1
try reinforcement learning
#55 opened by Unqua - 0
change apex.parallel.DistributedDataParallel to torch.nn.parallel.DistributedDataParallel
#51 opened by Unqua - 0
self-training类方法添加metric
#50 opened by Unqua - 0
- 1
add multi-head classifier for dst
#48 opened by Unqua - 1
训练时loss_epoch的统计有问题
#45 opened by Unqua - 1
add debias self-training
#44 opened by Unqua - 0
add select module
#42 opened by Unqua - 1
- 0
- 0
L1距离里怎么有softmax?
#36 opened by Unqua - 0
- 0
update measure files
#31 opened by Unqua - 1
add cycle experiment
#30 opened by Unqua - 0
update readme
#27 opened by Unqua - 0
- 0
add backbone
#23 opened by Unqua - 1
remove ratio
#21 opened by Unqua - 0
Shanghai-Hangzhou数据集源域和目标域反了
#19 opened by Unqua - 0
add self-training
#18 opened by Unqua - 1
- 0
- 2
把loss_composer解耦到train.py中
#10 opened by Unqua - 1
change dataset to float32 and apex to O0/O1
#7 opened by Unqua - 0
add result for Shanghai-Hangzhou dataset
#6 opened by Unqua - 0
add result for HyRANK
#3 opened by Unqua