Comments (4)
Yes we have visualized the reconstruction results several times and found the reconstructed results very similar to those obtained by ViT MAE (or Swin SimMIM), even before adding the GRN layer (ie. using the v1 model). In other words, it seems one cannot really judge the representation quality from the pixel space reconstruction quality: even we get “perfect” reconstructions, the finetuning results can still have a large gap.
from convnext-v2.
@s9xie Does ConvNeXt-v1/v2 model equipped without sparse conv also work well in image reconstruction?
from convnext-v2.
Yes we have visualized the reconstruction results several times and found the reconstructed results very similar to those obtained by ViT MAE (or Swin SimMIM), even before adding the GRN layer (ie. using the v1 model). In other words, it seems one cannot really judge the representation quality from the pixel space reconstruction quality: even we get “perfect” reconstructions, the finetuning results can still have a large gap.
Since the gap between pretraining and finetuning in the self-supervised paradigm, could we introcude the MAE-based ConvNeXt Encoder in a Semi-supvised framework such as FixMatch?
from convnext-v2.
@s9xie @shwoo93 Did you make a comparison among the supervised ConvNeXt-v2 models across different model sizes?
As model size shrinks, does the superiority of MAE-pretrained vanish compared with the supervised ones?
from convnext-v2.
Related Issues (20)
- unable to install apex HOT 2
- On masking input images HOT 1
- Make image and patch sizes dynamic
- Adapting ConvNetV2 for Time Series: Inconsistencies between Pre-training Visuals and Fine-Tuning Performance HOT 1
- could you share the dense masked conv based sparse encoding ? HOT 2
- ImageNet-1K pre-trained weights ConvNeXt V2 supervised HOT 1
- GRN can be used on any convent with FCMAE ?anyone tried this ?
- Cannot install MinkowskiEngine with provided instructions HOT 2
- Weights trained using pre-training script produce NaN values when running fine-tuning script
- Pre-trained weights incompatible with backbone HOT 2
- deployment issue in trt fp16
- Onnx Export HOT 1
- [Question] Cosine Similarity
- Finetuning with limited Labels ?
- ImageNet 22k(21k) Traning loss at the end of training
- Why not provide 22k-supervised finetuning model??? I am really shocked by that every available ConvNeXt-V2 pre-training weights has been finetuned on imagenet-1k. Please make 22k-supervised ConvNeXt-V2 open just like ConvNeXt-V1 !!!!!! 🙏🙏🙏 HOT 1
- can grn add to resnet
- downsample_layer model order
- Colab Implementation
- GRN in paper uses x_i / sum x_j whereas code uses x_i / mean x_j
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 convnext-v2.