Sha-Lab FEAT issues - Githubissues

Sha-Lab / FEAT

The code repository for "Few-Shot Learning via Embedding Adaptation with Set-to-Set Functions"

MIT License

418 stars 84 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

cannot reproduce the performance of FEAT

#82 Duuhiy opened 1 year ago
0
About the accuracy of Protonet.

#81 Glutaredoxin closed 2 years ago
4
question about pre training　and feat model

#80 dota2mhxy closed 2 years ago
2
fewshot model

#79 Catofblue closed 2 years ago
2
Can you upload data/tieredimagenet/split please?

#78 CandiceHan123 closed 2 years ago
1
Accuracy of pre-trained model for Tiered Imagenet Dataset

#77 mayug closed 2 years ago
1
Pretraining hyperparameters for CUB dataset

#76 mayug closed 2 years ago
1
Some question about CTM

#75 WoshiBoluo closed 2 years ago
1
the pre-training hyperparameters for ResNet12 on CUB

#74 DongYY127 closed 2 years ago
1
We can reproduce all numbers except for FEAT

#73 hieulem closed 2 years ago
1
DDP available?

#72 brando90 closed 2 years ago
1
Question about proto

#71 LoveMiki closed 2 years ago
1
Download the pretrained weights

#70 OverVan closed 2 years ago
1
hyperparameters that you used for ResNet12 pre-training on CUB

#69 gyr0104 closed 2 years ago
0
有关transductive FSL 的设置

#68 DongYY127 closed 2 years ago
1
Question about lr and gamma.

#67 zhanyuanyang closed 3 years ago
4
Key and value intuition of transformer in FEAT

#66 kumam92 closed 3 years ago
3
关于位置编码positional encoding

#65 YiX98 closed 2 years ago
1
reason behind good performance of Convnet over ResNet for noisy image

#64 kumam92 closed 3 years ago
1
Confusion about the contrastive learning.

#63 z1358 closed 3 years ago
1
Question about the cosine distance.

#62 z1358 closed 3 years ago
3
有关Pre-Avg和Post_Avg

#61 sherrijj closed 3 years ago
1
Curious about classification results without pre-training

#60 LoveMiki closed 3 years ago
4
dataset label in batches are not the actual labels

#59 kumam92 closed 3 years ago
5
Confusion matrix plot for testing episodes

#58 kumarmanas closed 3 years ago
0
请问关于提高准确率的建议

#57 YiX98 closed 3 years ago
3
用自己的数据集训练FEAT报错

#56 YiX98 closed 3 years ago
0
Could you give the hyperparameters for the CUB ?

#55 xue19890510 closed 3 years ago
1
Hyperparameter training command for protonet with WRN backbone on miniImagenet dataset

#54 techNN closed 3 years ago
2
Pretraining hyperparameters for Convnet and Mini Imagenet

#53 mayug closed 3 years ago
3
About the results of MAML on mini-imageNet with ConvNet

#52 JimZAI closed 3 years ago
0
关于feat.py模型内部关于输入数据shape的问题

#51 shuowang666 closed 3 years ago
2
About the 5-shot results on miniimagenet

#50 zhanyuanyang closed 3 years ago
4
请问可否提供WRN作backbone时fine-tune阶段训练的相关参数

#49 SnowyJune973 closed 3 years ago
1
预训练的结果没有达到您公布模型的准确率

#48 YuanMLer closed 3 years ago
1
pretraining hyper-parameters for tiered imagenet

#47 Tsingularity closed 3 years ago
1
label problem for meta-training stage

#46 freeman-1995 closed 3 years ago
0
The ProtoNet achieves the best Acc just after 1 epoch based on the given pre-trained ResNet12 weights.

#45 JimZAI closed 3 years ago
2
no avg_pooling in resnet-12?

#44 Tsingularity closed 3 years ago
2
hyper-parameters to reproduce protonet results on resnet-12?

#43 Tsingularity closed 3 years ago
5
what is the difference between the model and para_model in prepare_model?

#42 seyeeet closed 4 years ago
2
should I use the cache?

#41 seyeeet closed 4 years ago
1
understanding the basics of dataloader and procedure

#40 seyeeet closed 4 years ago
4
Question the weight of pre-training

#39 blue-blue272 closed 3 years ago
1
how can I test the model using pretrained models?

#38 seyeeet closed 4 years ago
0
Can anyone reproduce the results for tiered-imagenet?

#37 smallbox120 closed 3 years ago
3
a question about trainning error

#36 wongsihan closed 4 years ago
2
some confusion about how to get "aux_center" in feat.py

#35 tian0810 closed 4 years ago
1
a question about resnet-18/12 kernel size

#34 Sword-keeper closed 4 years ago
1
About the results you reported in the paper

#33 molybdenumyz closed 3 years ago
5