Comments (2)
Hello, batched inference is not implemented and is on the TODO list.
It should be straightforward using torch checkpoints but might require some more engineering using tensorrt and onnx.
The postprocessing to extract the coordinates from heatmaps should work right away, you can find it here.
I would be interested to merge a batched inference version if the performance gain is noticeable.
If you want to work on this I can provide you feedback and help. I would start by adapting the onnx models because the trt inference is more complex. Check the notebook used to export the models link and try to run inference on a batched input, it should have dynamic batch size already enabled.
Once it works check the other notebook to_trt.ipynb
that exports the trt models from the onnx ones. The export function and inference engine must be adapted to accept arbitrary batch sizes.
If you manage to run the models on batches we can later adapt the code to also run batched pre/post processing and bbox extraction.
from easy_vitpose.
Closing for inactivity, feel free to reopen
from easy_vitpose.
Related Issues (20)
- easy-ViTPose Import error HOT 2
- Normalize image before inference HOT 3
- Finetuning on custom dataset HOT 4
- No module named 'configs' HOT 2
- could not build wheels for pycuda and nvidia tensorrt version not found HOT 11
- error when i use .engine files to use tensorrt HOT 4
- Error with tensorrt model HOT 16
- Shoulder joint HOT 6
- Running inference on video with both humans and horses ? HOT 12
- Origins of coco_25 ckpt HOT 3
- First frame takes long, others are faster HOT 3
- ModuleNotFoundError: No module named 'easy_ViTPose.vit_models.backbone' HOT 1
- Slow loss convergence HOT 3
- failed fp16 inference. HOT 4
- fintuning HOT 1
- How to use official pretrained model in your project? HOT 1
- Evaluation on coco dataset HOT 5
- Why the offical model can not match? HOT 4
- When I use my own model the points are error. HOT 5
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 easy_vitpose.