fixmatch

fixmatch copied to clipboard

Hi All, Thanks for the nice work. I have a question regarding the depiction in Figure 1. Why do compute the consistency loss after sharpening the predictions? Why not minimize...

MLDeS

Hi I would like to ask a clarification question regarding experiment setting, i'm curious did u tune the hyperparameter for FixMatch or the other compared baselines?

hzhz2020

commands to reproduce results in the paper

1

Hi can you provide the commands (like you did in MixMatch repo, the runs/ folder) to reproduce the results in the paper? Thanks a lot! and thanks for the great...

hzhz2020

Hi guys, I have read your paper and really like the approach of Fixmatch. I am trying to implement this method for a different domain than image classification. But I...

imagio87

Hi It would be great if you can provide code for table 11 in your "FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence". I found the ablation interesting, and would...

hzhz2020

Really performs worse when number of classes are High

7

This algorithm really performs worse when number of classes are high like around 100, which is most of the time the case. At that time, we need more images per...

jaytimbadia

Question about RandAugment Implementation

16

In the paper you mention that flips are applied with 50 percent probability. Is it also the case that each randaugment sample is also applied with 50 percent probability or...

psellcam

Data augmentation for SVHN dataset

4

Many thanks for your clean codes. (1) It seems there is no code of dataloader for SVHN dataset. Would you like to introduce what kinds of data pre-processing and augmenation...

QiushiYang

``` CUDA_VISIBLE_DEVICES=6 python -B supervised.py --imagenet_data='/data/imagenet/tf_records/train/' --model_dir='/home/fixmatch-master/models/supervised' --steps_per_run=100 --dataset="imagenet" --hparams="\"bfloat16\": false, \"num_epochs\": 200, \"augment\": { \"type\": \"cta\" }" ``` The error： ``` 2021-07-04 21:12:57.888177: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic...

kiwen-lee

Hi, do you have results of FixMatch on ImageNet using ResNet18? For example 100 epochs or 200/300 epochs? Thanks in advance.

Ir1d