Comments (4)
We use 40G A100 GPUs in object detection & instance segmentation experiment. If you use 24G cards, please try smaller input resolution such as 768x768 or 1024x1024. Notice that the performance will also degenerate with smaller inputs.
I am trying to evaluate the EVA model https://huggingface.co/BAAI/EVA/blob/main/eva_coco_seg.pth
in COCO instance segementation. I have only GTX 1080 and it always is out of memory even after I have set image_size
to 192
in /path/to/EVA/det/projects/ViTDet/configs/common/coco_loader_lsj_1536.py
. If I did not mistake the config, it can't be helped.
from eva.
We use 40G A100 GPUs in object detection & instance segmentation experiment. If you use 24G cards, please try smaller input resolution such as 768x768 or 1024x1024. Notice that the performance will also degenerate with smaller inputs.
from eva.
If you are interested in training with a smaller backbone, we also provide EVA-L (vanilla ViT-L w/ 303M parameters) with up to 89.2% top-1 acc on IN-1K: https://github.com/baaivision/EVA/tree/master/eva#eva-l-learning-better-mim-representations-from-eva-clip.
To our knowledge, EVA-L is the best open-sourced large-sized vision encoder to date. Try it out if you are interested.
from eva.
I believe the issue at hand was addressed, as such I'm closing this. Feel free to ask if you have further questions.
from eva.
Related Issues (20)
- Deep speed not supported in Windows.
- eva_clip folder does not have create_model_and_transforms.py file
- deepspeed can not work
- Error when loading&training EVAVisionTransformer with patch_dropout=0.5
- EVA-01 and EVA-02
- About export model HOT 8
- the .sh script of Evaluate the fine-tuned EVA (336px, patch_size=14) on ImageNet-1K val with a single node (click to expand) can not execute. HOT 1
- The import accimage in the dateset folder cannot resolve the load.
- How to download the init_weight for seg model, the link is faild? HOT 2
- Issue with Detectron2 and CUDA
- cannot run the detection training with mutiple gpu on a single node HOT 3
- Cannot find Evaluation Metrics on LVIS HOT 2
- About post normalization
- Training loss is unstable HOT 5
- How to run Inference and Fine-Tuning on Instance Segmentation EVA-02?
- Ripple-like Artifacts on Mask Surfaces with EVA Model
- How to convert class id to class name HOT 1
- Pooling strategy
- use torch.jit.trace export pytorch 2 torchscript fail. HOT 8
- Can eva_video_k722 be used to extract video clip features?
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 eva.