Comments (6)
Currently, we have no plan to train the OCNet series models based on ResNet-50. If you mean the ImageNet pre-trained checkpoints, please download from here.
from openseg.pytorch.
During training both resnet50-imagenet.pth and resnet101-imagenet.pth you provide from scratch on ImageNet-1K, you make all the following modifications or not:
- the first 7x7 conv. of vanilla resnet50/101 is replaced with 3 3x3 ones ?
- dilation rates are set to 2 and 4 for last two stages ?
- nn.BatchNorm2d is replaced with InPlaceABNSync ?
from openseg.pytorch.
We should make the same modifications for both resnet50 and resnet101.
from openseg.pytorch.
@PkuRainBow
For resnet50-imagenet.pth and resnet101-imagenet.pth,you train them using all above three strategies ?
from openseg.pytorch.
We have not trained & reported any results based on resnet-50. We suggest making the modifications if you plan to use resnet-50.
from openseg.pytorch.
@PkuRainBow
The backbone model resnet101-imagenet.pth you used need to be trained from scratch on ImageNet-1K.
When train it, you adopt all above three strategies ?
from openseg.pytorch.
Related Issues (20)
- Address already in use: how to enable multiple training process on one machine?
- validation in inference process has a large difference with validation in train process HOT 1
- ERROR: Found no prediction for ground truth HOT 1
- Wrong! bash scripts/cityscapes/segfix/run_h_48_d_4_segfix.sh segfix_pred_val 1 HOT 1
- Mask loss is 0.0 and overall loss doesnt converge && what is relevance of dt_num_classes ? HOT 2
- TypeError: 'NoneType' object is not callable HOT 2
- Unexpected key(s) in state_dict: "config_dict", "state_dict". HOT 1
- FSSlowOhemCELoss is not defined in loss_helper
- why the refined result is white?
- can not reproduce eval result with segfix
- question about flops
- Result of refinement by SegFix on HRNet / HRNet-Semantic-Segmentation open source
- a pretrained SegFix model,
- preprocess scripts for LIP
- Checkpoints for SegFix models trained with Pytorch-1.7 have been released?
- Preprocess VOC dataset HOT 2
- need *.mat when I want to train segfix on my own dataset
- custom dataset
- label list
- About Train and Val HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from openseg.pytorch.