Swin Transformer for Object Detection - Demo Test

on Projects

Swin Transformer for Object Detection - Demo Test

1. Checkpoint

README.md
image
Download configs and .pth file

2. In the Docker

@ /mmdetection/checkpoints
image

@ /mmdetection/configs/swin
image

3. Tutorial

3.1. mask_rcnn_swin-s-p4-w7_fpn_fp16_ms-crop-3x_coco

image

from mmdet.apis import init_detector, inference_detector
import mmcv

# Specify the path to model config and checkpoint file
config_file = 'configs/swin/mask_rcnn_swin-s-p4-w7_fpn_fp16_ms-crop-3x_coco.py'
checkpoint_file = 'checkpoints/mask_rcnn_swin-s-p4-w7_fpn_fp16_ms-crop-3x_coco_20210903_104808-b92c91f1.pth'

# build the model from a config file and a checkpoint file
model = init_detector(config_file, checkpoint_file, device='cuda:0')

# test a single image and show the results
img = 'demo/demo.jpg'  # or img = mmcv.imread(img), which will only load it once
result = inference_detector(model, img)
# visualize the results in a new window
model.show_result(img, result)
# or save the visualization results to image files
model.show_result(img, result, out_file='demo/result_mask_rcnn_swin-s-p4-w7_fpn_fp16_ms-crop-3x_coco.jpg')

# test a video and show the results
video = mmcv.VideoReader('demo/demo.mp4')
for frame in video:
    result = inference_detector(model, frame)
    model.show_result(frame, result, wait_time=1)

3.2. mask_rcnn_swin-t-p4-w7_fpn_1x_coco

image

from mmdet.apis import init_detector, inference_detector
import mmcv

# Specify the path to model config and checkpoint file
config_file = 'configs/swin/mask_rcnn_swin-t-p4-w7_fpn_1x_coco.py'
checkpoint_file = 'checkpoints/mmask_rcnn_swin-t-p4-w7_fpn_1x_coco_20210902_120937-9d6b7cfa.pth'

# build the model from a config file and a checkpoint file
model = init_detector(config_file, checkpoint_file, device='cuda:0')

# test a single image and show the results
img = 'demo/demo.jpg'  # or img = mmcv.imread(img), which will only load it once
result = inference_detector(model, img)
# visualize the results in a new window
model.show_result(img, result)
# or save the visualization results to image files
model.show_result(img, result, out_file='demo/result_mask_rcnn_swin-t-p4-w7_fpn_1x_coco.jpg')

# test a video and show the results
video = mmcv.VideoReader('demo/demo.mp4')
for frame in video:
    result = inference_detector(model, frame)
    model.show_result(frame, result, wait_time=1)

3.3. mask_rcnn_swin-t-p4-w7_fpn_fp16_ms-crop-3x_coco

image

from mmdet.apis import init_detector, inference_detector
import mmcv

# Specify the path to model config and checkpoint file
config_file = 'configs/swin/mask_rcnn_swin-t-p4-w7_fpn_fp16_ms-crop-3x_coco.py'
checkpoint_file = 'checkpoints/mask_rcnn_swin-t-p4-w7_fpn_fp16_ms-crop-3x_coco_20210908_165006-90a4008c.pth'

# build the model from a config file and a checkpoint file
model = init_detector(config_file, checkpoint_file, device='cuda:0')

# test a single image and show the results
img = 'demo/demo.jpg'  # or img = mmcv.imread(img), which will only load it once
result = inference_detector(model, img)
# visualize the results in a new window
model.show_result(img, result)
# or save the visualization results to image files
model.show_result(img, result, out_file='demo/result_mask_rcnn_swin-t-p4-w7_fpn_fp16_ms-crop-3x_coco.jpg')

# test a video and show the results
video = mmcv.VideoReader('demo/demo.mp4')
for frame in video:
    result = inference_detector(model, frame)
    model.show_result(frame, result, wait_time=1)

3.4. mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco

image

from mmdet.apis import init_detector, inference_detector
import mmcv

# Specify the path to model config and checkpoint file
config_file = 'configs/swin/mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco.py'
checkpoint_file = 'checkpoints/mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco_20210906_131725-bacf6f7b.pth'

# build the model from a config file and a checkpoint file
model = init_detector(config_file, checkpoint_file, device='cuda:0')

# test a single image and show the results
img = 'demo/demo.jpg'  # or img = mmcv.imread(img), which will only load it once
result = inference_detector(model, img)
# visualize the results in a new window
model.show_result(img, result)
# or save the visualization results to image files
model.show_result(img, result, out_file='demo/result_mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco.jpg')

# test a video and show the results
video = mmcv.VideoReader('demo/demo.mp4')
for frame in video:
    result = inference_detector(model, frame)
    model.show_result(frame, result, wait_time=1)

3.5. tools/test.py

python tools/test.py \
    configs/swin/mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco.py \
    checkpoints/mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_coco_20210906_131725-bacf6f7b.pth \
    --show-dir mask_rcnn_swin-t-p4-w7_fpn_ms-crop-3x_results