Skip to content

Commit f76c1d4

Browse files
zylo117zylo117
committed
update readme and benchmark
1 parent 114465d commit f76c1d4

File tree

2 files changed

+33
-15
lines changed

2 files changed

+33
-15
lines changed

‎benchmark/coco_eval_result‎

Lines changed: 25 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -97,16 +97,29 @@ efficientdet-d6
9797
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.760
9898

9999
efficientdet-d7
100-
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.512
101-
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.703
102-
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.552
103-
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.361
104-
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.563
105-
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.650
106-
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.372
107-
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.598
108-
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.635
109-
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.481
110-
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.683
111-
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.761
100+
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.527
101+
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.719
102+
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.567
103+
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.362
104+
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.569
105+
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.662
106+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.385
107+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.616
108+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.656
109+
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.513
110+
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.696
111+
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.773
112112

113+
efficientdet-d7x
114+
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.539
115+
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.731
116+
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.580
117+
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.398
118+
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.575
119+
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.671
120+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.390
121+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.629
122+
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.670
123+
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.538
124+
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.704
125+
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.783

‎readme.md‎

Lines changed: 8 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -10,7 +10,7 @@ The performance is very close to the paper's, it is still SOTA.
1010

1111
The speed/FPS test includes the time of post-processing with no jit/data precision trick.
1212

13-
| coefficient | pth_download | GPU Mem(MB) | FPS | Extreme FPS (Batchsize 32) | mAP 0.5:0.95(this repo) | mAP 0.5:0.95(paper) |
13+
| coefficient | pth_download | GPU Mem(MB) | FPS | Extreme FPS (Batchsize 32) | mAP 0.5:0.95(this repo) | mAP 0.5:0.95(official) |
1414
| :-----: | :-----: | :------: | :------: | :------: | :-----: | :-----: |
1515
| D0 | [efficientdet-d0.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d0.pth) | 1049 | 36.20 | 163.14 | 33.1 | 33.8
1616
| D1 | [efficientdet-d1.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d1.pth) | 1159 | 29.69 | 63.08 | 38.8 | 39.6
@@ -19,7 +19,8 @@ The speed/FPS test includes the time of post-processing with no jit/data precisi
1919
| D4 | [efficientdet-d4.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d4.pth) | 1903 | 14.75 | - | 48.8 | 49.4
2020
| D5 | [efficientdet-d5.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d5.pth) | 2255 | 7.11 | - | 50.2 | 50.7
2121
| D6 | [efficientdet-d6.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d6.pth) | 2985 | 5.30 | - | 50.7 | 51.7
22-
| D7 | [efficientdet-d7.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.0/efficientdet-d7.pth) | 3819 | 3.73 | - | 51.2 | 52.2
22+
| D7 | [efficientdet-d7.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.2/efficientdet-d7.pth) | 3819 | 3.73 | - | 52.7 | 53.7
23+
| D7X | [efficientdet-d7x.pth](https://github.com/zylo117/Yet-Another-Efficient-Pytorch/releases/download/1.2/efficientdet-d7x.pth) | 3983 | 2.39 | - | 53.9 | 55.1
2324

2425
## Speed Test
2526

@@ -52,7 +53,11 @@ ___
5253

5354
## Update Log
5455

55-
[2020-05-11] add boolean string convertion to make sure head_only works
56+
[2020-07-23] supports efficientdet-d7x, mAP 53.9, using efficientnet-b7 as its backbone and an extra deeper pyramid level of BiFPN.
57+
58+
[2020-07-15] update efficientdet-d7 weights, mAP 52.7
59+
60+
[2020-05-11] add boolean string conversion to make sure head_only works
5661

5762
[2020-05-10] replace nms with batched_nms to further improve mAP by 0.5~0.7, thanks [Laughing-q](https://github.com/Laughing-q).
5863

0 commit comments

Comments
 (0)