-
Hi, Thanks for this repo :-)
I am try to train the network (and do an overfitt) using solov2 config
When I start the training process I see the image but the masks are wrong (the image flipped, but …
-
### Current Behavior:
According to the Keras documentation, the `__call__` and `predict` should yield identical results -- at least for small batches. That isn't the case for RetinaNet, where we get …
Dapid updated
8 months ago
-
### :question: Question
By default nnDetection use a RetinaNet as a backbone. I am wondering if one can replace the backbone with another network, e.g., Swin Transformer?
-
### Prerequisite
- [X] I have searched [Issues](https://github.com/open-mmlab/mmrotate/issues) and [Discussions](https://github.com/open-mmlab/mmrotate/discussions) but cannot get the expected help.
…
-
When we try to reproduce the instructions given [here](https://github.com/mlcommons/inference_results_v3.1/tree/main/closed/Qualcomm/code/retinanet/retinanet_kilt_loadgen_qaic#setup-docker-images) we …
-
This issue is created to keep track of the CK2/CM based workflows for all the MLPerf inference workflows being prepared by the [MLCommons Taskforce on Automation and Reproducibility](https://github.co…
-
### 🐛 Describe the bug
I encountered this problem when export compiled retinanet.
with code:
```python
import torch
import torchvision
model = torchvision.models.detection.retinanet_resnet5…
-
## Instructions To Reproduce the 🐛 Bug:
1. what changes you made (`git diff`) or what code you wrote
```
No change to code
```
2. what exact command you run:
I performed 2 machines 4 gpus (…
-
## Keyword: detection
### Video Anomaly Detection by Estimating Likelihood of Representations
- **Authors:** Yuqi Ouyang, Victor Sanchez
- **Subjects:** Computer Vision and Pattern Recognition (cs.C…
-
### Description of the bug | 错误描述
Exception ignored on calling ctypes callback function:
Traceback (most recent call last):
File "/opt/conda/lib/python3.10/site-packages/threadpoolctl.py", line …