- 8 NVIDIA Tesla V100 GPUs
- Intel Xeon 4114 CPU @ 2.20GHz
- Python 3.6 / 3.7
- PyTorch Nightly
- CUDA 9.0.176
- CUDNN 7.0.4
- NCCL 2.1.15
We use AWS as the main site to host our model zoo, and maintain a mirror on aliyun.
You can replace https://s3.ap-northeast-2.amazonaws.com/open-mmlab
with https://open-mmlab.oss-cn-beijing.aliyuncs.com
in model urls.
- All baselines were trained using 8 GPU with a batch size of 16 (2 images per GPU).
- All models were trained on
coco_2017_train
, and tested on thecoco_2017_val
. - We use distributed training and BN layer stats are fixed.
- We adopt the same training schedules as Detectron. 1x indicates 12 epochs and 2x indicates 24 epochs, which corresponds to slightly less iterations than Detectron and the difference can be ignored.
- All pytorch-style pretrained backbones on ImageNet are from PyTorch model zoo.
- For fair comparison with other codebases, we report the GPU memory as the maximum value of
torch.cuda.max_memory_allocated()
for all 8 GPUs. Note that this value is usually less than whatnvidia-smi
shows. - We report the inference time as the overall time including data loading, network forwarding and post processing.
More models with different backbones will be added to the model zoo.
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | AR1000 | Download |
---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 3.3 | 0.253 | 16.9 | 58.2 | - |
R-50-FPN | pytorch | 1x | 3.5 | 0.276 | 17.7 | 57.1 | model |
R-50-FPN | pytorch | 2x | - | - | - | 57.6 | model |
R-101-FPN | caffe | 1x | 5.2 | 0.379 | 13.9 | 59.4 | - |
R-101-FPN | pytorch | 1x | 5.4 | 0.396 | 14.4 | 58.6 | model |
R-101-FPN | pytorch | 2x | - | - | - | 59.1 | model |
X-101-32x4d-FPN | pytorch | 1x | 6.6 | 0.589 | 11.8 | 59.4 | model |
X-101-32x4d-FPN | pytorch | 2x | - | - | - | 59.9 | model |
X-101-64x4d-FPN | pytorch | 1x | 9.5 | 0.955 | 8.3 | 59.8 | model |
X-101-64x4d-FPN | pytorch | 2x | - | - | - | 60.0 | model |
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 3.6 | 0.333 | 12.9 | 36.7 | - |
R-50-FPN | pytorch | 1x | 3.8 | 0.353 | 12.5 | 36.4 | model |
R-50-FPN | pytorch | 2x | - | - | - | 37.7 | model |
R-101-FPN | caffe | 1x | 5.5 | 0.465 | 10.7 | 38.8 | - |
R-101-FPN | pytorch | 1x | 5.7 | 0.474 | 10.8 | 38.6 | model |
R-101-FPN | pytorch | 2x | - | - | - | 39.4 | model |
X-101-32x4d-FPN | pytorch | 1x | 6.9 | 0.672 | 9.3 | 40.2 | model |
X-101-32x4d-FPN | pytorch | 2x | - | - | - | 40.5 | model |
X-101-64x4d-FPN | pytorch | 1x | 9.8 | 1.040 | 7.1 | 41.3 | model |
X-101-64x4d-FPN | pytorch | 2x | - | - | - | 40.7 | model |
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | mask AP | Download |
---|---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 3.8 | 0.430 | 9.9 | 37.5 | 34.4 | - |
R-50-FPN | pytorch | 1x | 3.9 | 0.453 | 9.6 | 37.3 | 34.2 | model |
R-50-FPN | pytorch | 2x | - | - | - | 38.6 | 35.1 | model |
R-101-FPN | caffe | 1x | 5.7 | 0.534 | 8.8 | 39.9 | 36.1 | - |
R-101-FPN | pytorch | 1x | 5.8 | 0.571 | 8.9 | 39.4 | 35.9 | model |
R-101-FPN | pytorch | 2x | - | - | - | 40.4 | 36.6 | model |
X-101-32x4d-FPN | pytorch | 1x | 7.1 | 0.759 | 7.9 | 41.2 | 37.2 | model |
X-101-32x4d-FPN | pytorch | 2x | - | - | - | 41.4 | 37.1 | model |
X-101-64x4d-FPN | pytorch | 1x | 10.0 | 1.102 | 5.8 | 42.2 | 38.1 | model |
X-101-64x4d-FPN | pytorch | 2x | - | - | - | 42.0 | 37.8 | model |
Backbone | Style | Type | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | mask AP | Download |
---|---|---|---|---|---|---|---|---|---|
R-50-FPN | caffe | Faster | 1x | 3.3 | 0.242 | 18.4 | 36.6 | - | - |
R-50-FPN | pytorch | Faster | 1x | 3.5 | 0.250 | 16.5 | 35.8 | - | model |
R-50-FPN | pytorch | Faster | 2x | - | - | - | 37.1 | - | model |
R-101-FPN | caffe | Faster | 1x | 5.2 | 0.355 | 14.4 | 38.6 | - | - |
R-101-FPN | pytorch | Faster | 1x | 5.4 | 0.388 | 13.2 | 38.1 | - | model |
R-101-FPN | pytorch | Faster | 2x | - | - | - | 38.8 | - | model |
R-50-FPN | caffe | Mask | 1x | 3.4 | 0.328 | 12.8 | 37.3 | 34.5 | - |
R-50-FPN | pytorch | Mask | 1x | 3.5 | 0.346 | 12.7 | 36.8 | 34.1 | model |
R-50-FPN | pytorch | Mask | 2x | - | - | - | 37.9 | 34.8 | model |
R-101-FPN | caffe | Mask | 1x | 5.2 | 0.429 | 11.2 | 39.4 | 36.1 | - |
R-101-FPN | pytorch | Mask | 1x | 5.4 | 0.462 | 10.9 | 38.9 | 35.8 | model |
R-101-FPN | pytorch | Mask | 2x | - | - | - | 39.9 | 36.4 | model |
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 3.4 | 0.285 | 12.5 | 35.8 | - |
R-50-FPN | pytorch | 1x | 3.6 | 0.308 | 12.1 | 35.6 | model |
R-50-FPN | pytorch | 2x | - | - | - | 36.5 | model |
R-101-FPN | caffe | 1x | 5.3 | 0.410 | 10.4 | 37.8 | - |
R-101-FPN | pytorch | 1x | 5.5 | 0.429 | 10.9 | 37.7 | model |
R-101-FPN | pytorch | 2x | - | - | - | 38.1 | model |
X-101-32x4d-FPN | pytorch | 1x | 6.7 | 0.632 | 9.3 | 39.0 | model |
X-101-32x4d-FPN | pytorch | 2x | - | - | - | 39.3 | model |
X-101-64x4d-FPN | pytorch | 1x | 9.6 | 0.993 | 7.0 | 40.0 | model |
X-101-64x4d-FPN | pytorch | 2x | - | - | - | 39.6 | model |
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 3.9 | 0.464 | 9.7 | 40.6 | - |
R-50-FPN | pytorch | 1x | 4.1 | 0.455 | 10.1 | 40.5 | model |
R-50-FPN | pytorch | 20e | - | - | - | 41.1 | model |
R-101-FPN | caffe | 1x | 5.8 | 0.569 | 8.7 | 42.5 | - |
R-101-FPN | pytorch | 1x | 6.0 | 0.584 | 8.7 | 42.1 | model |
R-101-FPN | pytorch | 20e | - | - | - | 42.6 | model |
X-101-32x4d-FPN | pytorch | 1x | 7.2 | 0.770 | 7.8 | 43.7 | model |
X-101-32x4d-FPN | pytorch | 20e | - | - | - | 44.1 | model |
X-101-64x4d-FPN | pytorch | 1x | 10.0 | 1.133 | 6.1 | 44.6 | model |
X-101-64x4d-FPN | pytorch | 20e | - | - | - | 44.8 | model |
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | mask AP | Download |
---|---|---|---|---|---|---|---|---|
R-50-FPN | caffe | 1x | 5.1 | 0.692 | 6.7 | 41.0 | 35.6 | - |
R-50-FPN | pytorch | 1x | 5.3 | 0.683 | 6.5 | 41.3 | 35.7 | model |
R-50-FPN | pytorch | 20e | - | - | - | 42.4 | 36.6 | model |
R-101-FPN | caffe | 1x | 7.0 | 0.803 | 6.3 | 43.1 | 37.3 | - |
R-101-FPN | pytorch | 1x | 7.2 | 0.807 | 6.1 | 42.7 | 37.1 | model |
R-101-FPN | pytorch | 20e | - | - | - | 43.4 | 37.6 | model |
X-101-32x4d-FPN | pytorch | 1x | 8.4 | 0.976 | 5.7 | 44.4 | 38.3 | model |
X-101-32x4d-FPN | pytorch | 20e | - | - | - | 44.9 | 38.7 | model |
X-101-64x4d-FPN | pytorch | 1x | 11.4 | 1.33 | 4.7 | 45.3 | 39.1 | model |
X-101-64x4d-FPN | pytorch | 20e | - | - | - | 45.8 | 39.5 | model |
Notes:
- The
20e
schedule in Cascade (Mask) R-CNN indicates decreasing the lr at 16 and 19 epochs, with a total of 20 epochs.
Backbone | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | mask AP | Download |
---|---|---|---|---|---|---|---|---|
R-50-FPN | pytorch | 1x | 7.4 | 0.936 | 3.5 | 42.2 | 37.3 | model |
R-50-FPN | pytorch | 20e | - | - | - | 43.2 | 38.0 | model |
R-101-FPN | pytorch | 20e | 9.3 | 1.051 | 3.4 | 44.9 | 39.4 | model |
X-101-32x4d-FPN | pytorch | 20e | 5.8 | 0.769 | 3.3 | 46.1 | 40.3 | model |
X-101-64x4d-FPN | pytorch | 20e | 7.5 | 1.120 | 3.0 | 47.0 | 40.9 | model |
Notes:
- Please refer to Hybrid Task Cascade for details and more a powerful model (50.7/43.9).
Backbone | Size | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|---|
VGG16 | 300 | caffe | 120e | 3.5 | 0.256 | 25.9 / 34.6 | 25.7 | model |
VGG16 | 512 | caffe | 120e | 7.6 | 0.412 | 20.7 / 25.4 | 29.3 | model |
Backbone | Size | Style | Lr schd | Mem (GB) | Train time (s/iter) | Inf time (fps) | box AP | Download |
---|---|---|---|---|---|---|---|---|
VGG16 | 300 | caffe | 240e | 2.5 | 0.159 | 35.7 / 53.6 | 77.5 | model |
VGG16 | 512 | caffe | 240e | 4.3 | 0.214 | 27.5 / 35.9 | 80.0 | model |
Notes:
cudnn.benchmark
is set asTrue
for SSD training and testing.- Inference time is reported for batch size = 1 and batch size = 8.
- The speed difference between VOC and COCO is caused by model parameters and nms.
Please refer to Group Normalization for details.
Please refer to Weight Standardization for details.
Please refer to Deformable Convolutional Networks for details.
We compare mmdetection with Detectron and maskrcnn-benchmark. The backbone used is R-50-FPN.
In general, mmdetection has 3 advantages over Detectron.
- Higher performance (especially in terms of mask AP)
- Faster training speed
- Memory efficient
Detectron and maskrcnn-benchmark use caffe-style ResNet as the backbone. We report results using both caffe-style (weights converted from here) and pytorch-style (weights from the official model zoo) ResNet backbone, indicated as pytorch-style results / caffe-style results.
We find that pytorch-style ResNet usually converges slower than caffe-style ResNet, thus leading to slightly lower results in 1x schedule, but the final results of 2x schedule is higher.
Type | Lr schd | Detectron | maskrcnn-benchmark | mmdetection |
---|---|---|---|---|
RPN | 1x | 57.2 | - | 57.1 / 58.2 |
2x | - | - | 57.6 / - | |
Faster R-CNN | 1x | 36.7 | 36.8 | 36.4 / 36.7 |
2x | 37.9 | - | 37.7 / - | |
Mask R-CNN | 1x | 37.7 & 33.9 | 37.8 & 34.2 | 37.3 & 34.2 / 37.5 & 34.4 |
2x | 38.6 & 34.5 | - | 38.6 & 35.1 / - | |
Fast R-CNN | 1x | 36.4 | - | 35.8 / 36.6 |
2x | 36.8 | - | 37.1 / - | |
Fast R-CNN (w/mask) | 1x | 37.3 & 33.7 | - | 36.8 & 34.1 / 37.3 & 34.5 |
2x | 37.7 & 34.0 | - | 37.9 & 34.8 / - |
The training speed is measure with s/iter. The lower, the better.
Type | Detectron (P1001) | maskrcnn-benchmark (V100) | mmdetection (V1002) |
---|---|---|---|
RPN | 0.416 | - | 0.253 |
Faster R-CNN | 0.544 | 0.353 | 0.333 |
Mask R-CNN | 0.889 | 0.454 | 0.430 |
Fast R-CNN | 0.285 | - | 0.242 |
Fast R-CNN (w/mask) | 0.377 | - | 0.328 |
*1. Facebook's Big Basin servers (P100/V100) is slightly faster than the servers we use. mmdetection can also run slightly faster on FB's servers.
*2. For fair comparison, we list the caffe-style results here.
The inference speed is measured with fps (img/s) on a single GPU. The higher, the better.
Type | Detectron (P100) | maskrcnn-benchmark (V100) | mmdetection (V100) |
---|---|---|---|
RPN | 12.5 | - | 16.9 |
Faster R-CNN | 10.3 | 7.9 | 12.9 |
Mask R-CNN | 8.5 | 7.7 | 9.9 |
Fast R-CNN | 12.5 | - | 18.4 |
Fast R-CNN (w/mask) | 9.9 | - | 12.8 |
Type | Detectron | maskrcnn-benchmark | mmdetection |
---|---|---|---|
RPN | 6.4 | - | 3.3 |
Faster R-CNN | 7.2 | 4.4 | 3.6 |
Mask R-CNN | 8.6 | 5.2 | 3.8 |
Fast R-CNN | 6.0 | - | 3.3 |
Fast R-CNN (w/mask) | 7.9 | - | 3.4 |
There is no doubt that maskrcnn-benchmark and mmdetection is more memory efficient than Detectron, and the main advantage is PyTorch itself. We also perform some memory optimizations to push it forward.
Note that Caffe2 and PyTorch have different apis to obtain memory usage with different implementations.
For all codebases, nvidia-smi
shows a larger memory usage than the reported number in the above table.