Issues
- 1
More detail about T-SNE visual
#88 opened by yuheyuan - 0
bad results
#91 opened by SSSilvester - 2
MiT-B3 log
#66 opened by Jjomaeng - 11
Training weights cannot be downloaded
#74 opened by jiuyuzhang - 2
Question about default .py for config
#78 opened by ShenZheng2000 - 1
Unable to reproduce results
#69 opened by IzumiKDl - 1
Data Preprocessing
#90 opened by ZhenboZhao77 - 1
test.sh: 4: Bad substitution
#83 opened by yuheyuan - 1
synthia dataset get unexpected json
#82 opened by ShenZheng2000 - 0
cv::OutOfMemoryError
#89 opened by ZhenboZhao77 - 1
- 1
resizing to 1024,512 cityscape
#72 opened by elhamAm - 2
visuial of t-sne
#56 opened by yuheyuan - 7
- 0
How to use a new backbone?
#87 opened by wxpppppp0621 - 0
How to obtain parameters in your code.
#86 opened by yuheyuan - 0
- 3
MIT weight无法获得
#84 opened by Ronever - 0
decode.loss_seg and mix.decode.loss_seg
#81 opened by jiuyuzhang - 0
- 0
Classes for Palette for DarkZurich
#79 opened by ShenZheng2000 - 1
Google Drive Download Seems down
#76 opened by steven12138 - 1
About the threshold
#70 opened by JinYoon002 - 0
- 0
- 1
Model for validation
#68 opened by Rui-Zhou-2 - 1
Test method
#62 opened by HKQX - 1
Reproduction
#64 opened by Wenlve-Zhou - 2
T-SNE code
#67 opened by endaoguansanlu - 2
The running order of code.
#63 opened by QChhh123 - 1
Welcome update to OpenMMLab 2.0
#61 opened by vansin - 4
- 1
What's the "IMTRv21_5"?
#58 opened by nightrain-vampire - 3
Questions about training on custom dataset
#57 opened by qubewx - 2
- 2
Rare class sampling magic number min_pixels=3000
#60 opened by msseibel - 2
T-sne visualization
#51 opened by YYDSDD - 1
Different batch sizes for source and target data
#54 opened by kaigelee - 2
MiT-B3 is much better than MiT-B4
#47 opened by 5as4as - 4
code about "Rare Class Sampling (RCS)"
#53 opened by yuheyuan - 1
Compatible with earlier MMSEG versions
#49 opened by LuPaoPao - 3
- 3
Losses are backpropagated separately
#48 opened by HKQX - 4
DarkZurich
#52 opened by mkbkxdd - 4
- 1
- 2
About 'sample_class_stats.json'
#50 opened by liwei1101 - 0
- 1
- 0
How to tune hyper-parameters?
#41 opened by super233