car-detection-bayes/README.md

192 lines
9.8 KiB
Markdown
Raw Normal View History

2019-02-13 21:15:30 +00:00
<table style="width:100%">
<tr>
2019-04-15 12:45:43 +00:00
<td>
2019-07-21 12:28:02 +00:00
<img src="https://user-images.githubusercontent.com/26833433/61591130-f7beea00-abc2-11e9-9dc0-d6abcf41d713.jpg">
2019-04-15 12:45:43 +00:00
</td>
<td align="center">
<a href="https://www.ultralytics.com" target="_blank">
2019-07-28 13:57:01 +00:00
<img src="https://storage.googleapis.com/ultralytics/logo/logoname1000.png" width="160"></a>
2019-07-21 12:28:02 +00:00
<img src="https://user-images.githubusercontent.com/26833433/61591093-2b4d4480-abc2-11e9-8b46-d88eb1dabba1.jpg">
<a href="https://itunes.apple.com/app/id1452689527" target="_blank">
2019-04-15 12:45:43 +00:00
<img src="https://user-images.githubusercontent.com/26833433/50044365-9b22ac00-0082-11e9-862f-e77aee7aa7b0.png" width="180"></a>
</td>
<td>
2019-07-21 12:28:02 +00:00
<img src="https://user-images.githubusercontent.com/26833433/61591100-55066b80-abc2-11e9-9647-52c0e045b288.jpg">
2019-04-15 12:45:43 +00:00
</td>
2019-02-13 21:15:30 +00:00
</tr>
</table>
2019-02-13 21:01:58 +00:00
2019-04-26 10:01:43 +00:00
# Introduction
This directory contains PyTorch YOLOv3 software and an iOS App developed by Ultralytics LLC, and **is freely available for redistribution under the GPL-3.0 license**. For more information please visit https://www.ultralytics.com.
2018-08-26 08:51:39 +00:00
# Description
2019-03-28 12:46:23 +00:00
The https://github.com/ultralytics/yolov3 repo contains inference and training code for YOLOv3 in PyTorch. The code works on Linux, MacOS and Windows. Training is done on the COCO dataset by default: https://cocodataset.org/#home. **Credit to Joseph Redmon for YOLO:** https://pjreddie.com/darknet/yolo/.
2018-08-26 08:51:39 +00:00
# Requirements
2018-11-27 17:14:48 +00:00
Python 3.7 or later with the following `pip3 install -U -r requirements.txt` packages:
2018-08-26 08:51:39 +00:00
- `numpy`
2019-05-29 00:02:41 +00:00
- `torch >= 1.1.0`
2018-08-26 08:51:39 +00:00
- `opencv-python`
- `tqdm`
2018-08-26 08:51:39 +00:00
2019-03-05 15:23:33 +00:00
# Tutorials
* [GCP Quickstart](https://github.com/ultralytics/yolov3/wiki/GCP-Quickstart)
2019-03-05 15:23:33 +00:00
* [Transfer Learning](https://github.com/ultralytics/yolov3/wiki/Example:-Transfer-Learning)
* [Train Single Image](https://github.com/ultralytics/yolov3/wiki/Example:-Train-Single-Image)
* [Train Single Class](https://github.com/ultralytics/yolov3/wiki/Example:-Train-Single-Class)
* [Train Custom Data](https://github.com/ultralytics/yolov3/wiki/Train-Custom-Data)
2019-05-28 14:14:37 +00:00
# Jupyter Notebook
2019-07-26 17:13:40 +00:00
Our Jupyter [notebook](https://colab.research.google.com/github/ultralytics/yolov3/blob/master/examples.ipynb) provides quick training, inference and testing examples.
2019-05-28 14:14:37 +00:00
2018-08-26 09:05:13 +00:00
# Training
2018-08-26 08:51:39 +00:00
2019-08-01 20:40:18 +00:00
**Start Training:** `python3 train.py` to begin training after downloading COCO data with `data/get_coco_dataset.sh`. Each epoch trains on 117,263 images from the train and validate COCO sets, and tests on 5000 images from the COCO validate set.
2018-09-01 16:35:28 +00:00
2019-07-15 15:54:31 +00:00
**Resume Training:** `python3 train.py --resume` to resume training from `weights/last.pt`.
2018-09-01 16:35:28 +00:00
2019-08-01 20:40:18 +00:00
**Plot Training:** `from utils import utils; utils.plot_results()` plots training results from `coco_16img.data`, `coco_64img.data`, 2 example datasets available in the `data/` folder, which train and test on the first 16 and 64 images of the COCO2014-trainval dataset.
2019-08-12 12:39:26 +00:00
![image](https://user-images.githubusercontent.com/26833433/62865295-5ed94580-bd0e-11e9-9803-e07571e2ea23.png)
2018-09-01 11:34:05 +00:00
2018-09-01 11:43:07 +00:00
## Image Augmentation
2018-09-01 11:34:05 +00:00
2018-09-01 16:57:18 +00:00
`datasets.py` applies random OpenCV-powered (https://opencv.org/) augmentation to the input images in accordance with the following specifications. Augmentation is applied **only** during training, not during inference. Bounding boxes are automatically tracked and updated with the images. 416 x 416 examples pictured below.
2018-09-01 11:41:34 +00:00
Augmentation | Description
--- | ---
2018-12-28 18:23:35 +00:00
Translation | +/- 10% (vertical and horizontal)
2018-09-01 12:04:42 +00:00
Rotation | +/- 5 degrees
2018-12-28 18:23:35 +00:00
Shear | +/- 2 degrees (vertical and horizontal)
Scale | +/- 10%
2018-09-01 12:10:06 +00:00
Reflection | 50% probability (horizontal-only)
2018-09-01 12:04:42 +00:00
H**S**V Saturation | +/- 50%
HS**V** Intensity | +/- 50%
2018-09-01 11:34:05 +00:00
2019-07-20 15:31:21 +00:00
<img src="https://user-images.githubusercontent.com/26833433/61579359-507b7d80-ab04-11e9-8a2a-bd6f59bbdfb4.jpg">
2018-08-26 08:51:39 +00:00
2019-03-20 11:35:39 +00:00
## Speed
https://cloud.google.com/deep-learning-vm/
**Machine type:** n1-standard-8 (8 vCPUs, 30 GB memory)
2019-03-20 11:35:39 +00:00
**CPU platform:** Intel Skylake
2019-08-05 00:55:03 +00:00
**GPUs:** K80 ($0.20/hr), T4 ($0.35/hr), V100 ($0.83/hr) CUDA with [Nvidia Apex](https://github.com/NVIDIA/apex) FP16/32
2019-03-20 11:35:39 +00:00
**HDD:** 100 GB SSD
2019-07-20 17:27:27 +00:00
**Dataset:** COCO train 2014 (117,263 images)
2019-08-05 00:55:03 +00:00
GPUs | `batch_size` | images/sec | epoch time | epoch cost
--- |---| --- | --- | ---
2019-08-05 00:55:03 +00:00
K80 | 64 (32x2) | 11 | 175 min | $0.58
T4 | 64 (32x2) | 40 | 49 min | $0.29
T4 x2 | 64 (64x1) | 61 | 32 min | $0.36
V100 | 64 (32x2) | 115 | 17 min | $0.24
2019-08-05 16:06:01 +00:00
V100 x2 | 64 (64x1) | 150 | 13 min | $0.36
2019-08-05 00:55:03 +00:00
2080Ti | 64 (32x2) | 69 | 28 min | -
2019-07-30 13:23:31 +00:00
2019-03-20 11:35:39 +00:00
2018-08-26 09:05:13 +00:00
# Inference
2019-02-11 13:13:27 +00:00
2019-07-20 15:31:21 +00:00
`detect.py` runs inference on all images **and videos** in the `data/samples` folder:
2018-12-22 12:05:52 +00:00
2019-03-21 11:01:07 +00:00
**YOLOv3:** `python3 detect.py --cfg cfg/yolov3.cfg --weights weights/yolov3.weights`
2019-03-21 11:00:24 +00:00
<img src="https://user-images.githubusercontent.com/26833433/50524393-b0adc200-0ad5-11e9-9335-4774a1e52374.jpg" width="600">
2018-12-22 12:05:52 +00:00
2019-03-21 11:01:07 +00:00
**YOLOv3-tiny:** `python3 detect.py --cfg cfg/yolov3-tiny.cfg --weights weights/yolov3-tiny.weights`
2019-03-21 11:00:24 +00:00
<img src="https://user-images.githubusercontent.com/26833433/50374155-21427380-05ea-11e9-8d24-f1a4b2bac1ad.jpg" width="600">
2019-03-21 11:01:07 +00:00
**YOLOv3-SPP:** `python3 detect.py --cfg cfg/yolov3-spp.cfg --weights weights/yolov3-spp.weights`
2019-03-21 11:00:24 +00:00
<img src="https://user-images.githubusercontent.com/26833433/54747926-e051ff00-4bd8-11e9-8b5d-93a41d871ec7.jpg" width="600">
2018-09-04 12:36:51 +00:00
2019-02-11 13:11:24 +00:00
## Webcam
2019-08-01 16:20:47 +00:00
`python3 detect.py --webcam` shows a live webcam feed.
2019-02-11 13:11:24 +00:00
2018-12-22 11:49:55 +00:00
# Pretrained Weights
2018-12-22 11:58:59 +00:00
2019-03-20 12:08:24 +00:00
- Darknet `*.weights` format: https://pjreddie.com/media/files/yolov3.weights
- PyTorch `*.pt` format: https://drive.google.com/drive/folders/1uxgUBemJVw9wZsdpboYbzUN4bcRhsuAI
2018-09-01 16:48:53 +00:00
2019-04-23 15:05:42 +00:00
## Darknet Conversion
```bash
git clone https://github.com/ultralytics/yolov3 && cd yolov3
# convert darknet cfg/weights to pytorch model
python3 -c "from models import *; convert('cfg/yolov3-spp.cfg', 'weights/yolov3-spp.weights')"
Success: converted 'weights/yolov3-spp.weights' to 'converted.pt'
# convert cfg/pytorch model to darknet weights
python3 -c "from models import *; convert('cfg/yolov3-spp.cfg', 'weights/yolov3-spp.pt')"
Success: converted 'weights/yolov3-spp.pt' to 'converted.weights'
```
2019-03-20 11:26:46 +00:00
# mAP
2019-07-20 15:31:21 +00:00
- `test.py --weights weights/yolov3.weights` tests official YOLOv3 weights.
- `test.py --weights weights/last.pt` tests most recent checkpoint.
- `test.py --weights weights/best.pt` tests best checkpoint.
- Compare to darknet published results https://arxiv.org/abs/1804.02767.
2019-08-14 13:54:40 +00:00
[ultralytics/yolov3](https://github.com/ultralytics/yolov3) mAP@0.5 ([darknet](https://arxiv.org/abs/1804.02767)-reported mAP@0.5)
2019-08-14 13:08:44 +00:00
<i></i> | 320 | 416 | 608
--- | --- | --- | ---
`YOLOv3` | 51.8 (51.5) | 55.4 (55.3) | 58.2 (57.9)
`YOLOv3-SPP` | 52.4 | 56.5 | 60.7 (60.6)
`YOLOv3-tiny` | 29.0 | 32.9 (33.1) | 35.5
2019-03-20 11:26:46 +00:00
``` bash
2019-07-20 15:31:21 +00:00
# install pycocotools
2019-04-12 12:24:51 +00:00
git clone https://github.com/cocodataset/cocoapi && cd cocoapi/PythonAPI && make && cd ../.. && cp -r cocoapi/PythonAPI/pycocotools yolov3
2019-03-20 12:08:24 +00:00
cd yolov3
2019-07-20 15:31:21 +00:00
python3 test.py --save-json --img-size 608
Namespace(batch_size=16, cfg='cfg/yolov3-spp.cfg', conf_thres=0.001, data='data/coco.data', img_size=608, iou_thres=0.5, nms_thres=0.5, save_json=True, weights='weights/yolov3-spp.weights')
Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)
Class Images Targets P R mAP F1: 100% 313/313 [07:40<00:00, 2.34s/it]
all 5e+03 3.58e+04 0.117 0.788 0.595 0.199
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.367
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.607 <--
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.387
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.208
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.392
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.487
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.297
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.465
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.495
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.332
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.518
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.621
2019-04-12 12:24:51 +00:00
python3 test.py --save-json --img-size 416
2019-07-20 15:31:21 +00:00
Namespace(batch_size=16, cfg='cfg/yolov3-spp.cfg', conf_thres=0.001, data='data/coco.data', img_size=416, iou_thres=0.5, nms_thres=0.5, save_json=True, weights='weights/yolov3-spp.weights')
Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)
Class Images Targets P R mAP F1: 100% 313/313 [07:01<00:00, 1.41s/it]
all 5e+03 3.58e+04 0.105 0.746 0.554 0.18
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.336
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.565 <--
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.350
2019-04-12 12:24:51 +00:00
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.151
2019-07-20 15:31:21 +00:00
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.361
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.494
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.281
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.433
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.459
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.256
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.495
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.622
2019-03-20 11:26:46 +00:00
```
2018-08-26 08:51:39 +00:00
2019-04-03 10:42:40 +00:00
# Citation
[![DOI](https://zenodo.org/badge/146165888.svg)](https://zenodo.org/badge/latestdoi/146165888)
2018-08-26 08:51:39 +00:00
# Contact
2019-04-04 15:34:11 +00:00
Issues should be raised directly in the repository. For additional questions or comments please email Glenn Jocher at glenn.jocher@ultralytics.com or visit us at https://contact.ultralytics.com.