Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Lower validation accuracy

See original GitHub issue

First of all, thanks for providing your code. When I use your uploaded VOC12_scenes_20000.pth model to evaluate the single scale model on 1449 validation data, I get 71.8% which is much lower than reported 74.0%. Moreover, when I try to train the single scale model by running python train.py --random-mirror --random-scale --gpu 0, and then evaluate the trained model, I get 73.1%, which is still lower than 74%. Do you know where this discrepancy might come from? Everything is on default settings and I also use MS_DeepLab_resnet_pretrained_COCO_init.pth as the pre-trained model.

Issue Analytics

State:
Created 6 years ago
Comments:5 (2 by maintainers)

Top GitHub Comments

1reaction

siinemcommented, Nov 19, 2018

Could anybody share VOC12_scenes_20000.pth on googledrive? I couldnot install it from the link given here: https://pan.baidu.com/s/1bP52R8

in chinese and it requires me to install a exe file… baudi…something …

0reactions

monaj07commented, Nov 7, 2017

I see, thanks for reporting back. So what do you think the problem might be? In evaluate.py, I simply just modified lines 22 and 27 for the path to the data as well as the path to VOC12_scenes_20000.pth. Everything else is the same and I am a bit confused.

Top Results From Across the Web

High model accuracy vs very low validation accuarcy

Reduce the number of learning parameters in your model. These are the 3 solutions that are most likely to improve the validation accuracy...

Improving Your Model's Accuracy - Apple Developer

If your model's accuracy on your testing data is lower than your training or validation accuracy, it usually indicates that there are meaningful...

Your validation loss is lower than your training loss? This is why!

Symptoms: validation loss is consistently lower than the training loss, the gap between them remains more or less the same size and training...

tensorflow - Ways to decrease validation loss % and increase ...

So I want to check if there is something I could do to lower validation loss %, or more importantly, increase validation accuracy...

Can validation accuracy be higher than training ... - Quora

Validation accuracy will be usually less than training accuracy because training data is something with which the model is already familiar with and ......