Comments (2)
Hi, thanks for the question! We tried different initialization methods for prompt in Appendix B, Fig. 13. To our surprise, default random initialization works the best in general, in contrast to comparable studies in NLP.
As for why we randomly initialize prompts with xavier uniform initialization scheme, we follow the initialization of linear layers of ViT in MAE and JAX ViT. The intuition is that prompts actually act as an "additional filter" for the image patches in ViT.
from vpt.
Thanks for your prompt reply.
from vpt.
Related Issues (20)
- Compared with full fine-tuning, did the training time get reduced?
- Dataset Link Update HOT 1
- TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'
- Segmentation task
- CUB dataset HOT 2
- Wrong MoCo-v3 pretrained weights?
- How to perform segmentation task on my own segmentation dataset?
- How to train? HOT 1
- Run the final evaluation using the full training data? HOT 1
- Result summarization wrong?
- Hello, save model
- Hello, I would like to know about transfer_type
- For swin-B training in pad prompt
- what does the x mean? HOT 1
- Why is the training performance of my VPT so poor? HOT 1
- Where do we get the model files for pretrain?
- what's the difference between train.py and tune_*.py
- dataset split file link is not available
- dSprites dataset
- download the CUB dataset, but there is no train.json in it. HOT 2
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 vpt.