Comments (2)
This is more about a performance-driven choice.
The MIM pre-trained EVA usually achieves higher accuracy w/ GAP feature in e2e fine-tuning. This is also ture for other MIM pre-trained ViTs, e.g., BEiT, MAE, etc.
For EVA-CLIP, we find using GAP feature achieves sub-optimal performance than using cls_token
.
During MIM pre-training, although there is no direct supervision on cls_token
, we believe the cls_token
indeed represents some kinds of image-level / global content. On the other hand, using GAP image feature is not an absolutely reasonable choice: mask token
causes a discrepancy between MIM pre-training and finetuning, as the artificial mask tokens never appear in the fine-tuning stage.
As another example, MAE also uses cls_token
for linear probing. We tried with GAP feature for MAE linear probing, the performance is not very satisfactory.
from eva.
OK, thanks for your patient explanation.
from eva.
Related Issues (20)
- CUDA out of memory when inference
- whether the training of EVA involves masking text (caption) token? HOT 1
- Robust accuracy on ObjectNet dataset.
- Is there any plan to use lora for finetune eva_clip? HOT 1
- How to use EVA-02 detection models for inference?
- [EVA-CLIP]Β Are you sure about your way to compute probabilities? HOT 1
- Why are the model weights saved at each epoch the same when using DeepSpeed? HOT 1
- Can EVA-CLIP-8B and EVA-CLIP-18B support quantization? HOT 6
- ```EVA-CLIP-18B``` weight is not available at HuggingFace HOT 2
- Is there any reason why you didn't use the EVA02 architecture for EVA-CLIP models larger than 4B? HOT 1
- different weight decay between code and paper for clip 18b HOT 1
- Any plans for releasing model weights of COCO-Stuff trained segmentation models with EVA-02 backbones
- EVA CLIP 8B GPU out of memory in COLAB PRO Hugging face version HOT 1
- can remove xformers dependency? HOT 1
- Issues about the performance of vanilla EVA-02-S
- Dockerfile error
- How much GPU does EVA-CLIP-8B-plus need?
- Where is EVA G? HOT 3
- Size 14x14 to 16x16 patch interpolation for smaller EVA 2 models HOT 1
- Chinese ability of EVA-CLIP-18B
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
π 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. πππ
-
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 β€οΈ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from eva.