-
### 💡 Your Question
When running a training with YOLONAS, I get an error towards the end of training (or so I presume):
This is the line where the error occurs (line 156 in the code):
```python…
-
### Describe the bug
### Environment:
- Cuda=11.8, Torch=1.13
- Super Gradients …
-
## ❔Question
## Additional context
-
# Description
I am using a modified version of the YOLO NAS fine [tuning notebook](https://colab.research.google.com/drive/1q0RmeVRzLwRXW-h9dPFSOchwJkThUy6d?usp=sharing) which can be found in https:/…
-
> 🔨 Dead Line 🔨: 2023년 3월 20일
Training Data 및 Test Data 디렉토리 구축하고
모델 학습까지 해오기 !
데이터 디렉토리 구축 및 모델 학습은 [이거](https://github.com/Team-BoonMoSa/PANPP/blob/main/main.ipynb) 참고 ~
-
DO I have to export the models in modelzoo for deployment purpose or use it as it is
-
-
ppyolo使用paddle-trt在jetson平台上cpp推理出现问题，没有框，全是标签。fluid模式检测正常。因为后面打算做一下量化，必须使用trt才能加速。
-
🔥 PP-TinyPose是针对移动端设备研发的轻量级实时关键点检测模型，具有以下特色：
- 精度高：场景适应能力更强，在COCO上AP可达50.3。
- 速度快：在SD865上，单人配置可达122FPS，多人配置可达30FPS以上。
- 部署便捷：支持基于Paddle Inference/Paddle Lite的部署方案，我们同时提供了C++、Python及Paddle Lite的部署…
-
- 版本、环境信息：
1）PaddlePaddle版本：1.8.4
2）CPU：i7
3）GPU：N/A
4）Mac OS 10.15, Python 3.7
- 预测信息
1）参考 [PaddleDetection/deploy/python/infer.py](https://github.com/PaddlePaddle/PaddleDet…