Skip to content

Commit

Permalink
Add val results.
Browse files Browse the repository at this point in the history 
Also fix a bug in model_inspect and show examples of non-square images.
  • Loading branch information
@mingxingtan
mingxingtan committed Apr 14, 2020
1 parent c51876e commit e579bab
Show file tree
Hide file tree
Showing 2 changed files with 16 additions and 16 deletions.
30 changes: 15 additions & 15 deletions efficientdet/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -47,18 +47,18 @@ Our model family starts from EfficientDet-D0, which has comparable accuracy as [

We have provided a list of EfficientDet checkpoints and results as follows:

| Model | AP | AP<sub>50</sub> | AP<sub>75</sub> |AP<sub>S</sub> | AP<sub>M</sub> | AP<sub>L</sub> | | #params | #FLOPs |
|---------- |------ |------ | -------- | ------| ------| ------ |------ |------ | :------: |
| EfficientDet-D0 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d0.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d0_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d0_coco_test-dev2017.txt)) | 33.8 | 52.2 | 35.8 | 12.0 | 38.3 | 51.2 | | 3.9M | 2.54B |
| EfficientDet-D1 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d1.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d1_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d1_coco_test-dev2017.txt)) | 39.6 | 58.6 | 42.3 | 17.9 | 44.3 | 56.0 | | 6.6M | 6.10B |
| EfficientDet-D2 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d2.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d2_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d2_coco_test-dev2017.txt)) | 43.0 | 62.3 | 46.2 | 22.5 | 47.0 | 58.4 | | 8.1M | 11.0B |
| EfficientDet-D3 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d3.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d3_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d3_coco_test-dev2017.txt)) | 45.8 | 65.0 | 49.3 | 26.6 | 49.4 | 59.8 | | 12.0M | 24.9B |
| EfficientDet-D4 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d4.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d4_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d4_coco_test-dev2017.txt)) | 49.4 | 69.0 | 53.4 | 30.3 | 53.2 | 63.2 | | 20.7M | 55.2B |
| EfficientDet-D5 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d5.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d5_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d5_coco_test-dev2017.txt)) | 50.7 | 70.2 | 54.7 | 33.2 | 53.9 | 63.2 | | 33.7M | 135.4B |
| EfficientDet-D6 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d6.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d6_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d6_coco_test-dev2017.txt)) | 51.7 | 71.2 | 56.0 | 34.1 | 55.2 | 64.1 | | 51.9M | 225.6B |
| EfficientDet-D7 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d7.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d7_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d7_coco_test-dev2017.txt)) | 52.2 | 71.4 | 56.3 | 34.8 | 55.5 | 64.6 | | 51.9M | 324.8B |

** <em>val</em> denotes validation results, <em>test-dev</em> denotes test-dev2017 results. AP results in the table are for COCO test-dev2017. All checkpoints are trained with baseline preprocessing (no autoaugmentation).
| Model | AP<sup>val</sup> | AP<sup>test</sup> | AP<sub>50</sub> | AP<sub>75</sub> |AP<sub>S</sub> | AP<sub>M</sub> | AP<sub>L</sub> | | #params | #FLOPs |
|---------- |------ |------ |------ | -------- | ------| ------| ------ |------ |------ | :------: |
| EfficientDet-D0 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d0.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d0_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d0_coco_test-dev2017.txt)) | 33.5 | 33.8 | 52.2 | 35.8 | 12.0 | 38.3 | 51.2 | | 3.9M | 2.54B |
| EfficientDet-D1 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d1.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d1_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d1_coco_test-dev2017.txt)) | 39.1 | 39.6 | 58.6 | 42.3 | 17.9 | 44.3 | 56.0 | | 6.6M | 6.10B |
| EfficientDet-D2 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d2.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d2_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d2_coco_test-dev2017.txt)) | 42.5 | 43.0 | 62.3 | 46.2 | 22.5 | 47.0 | 58.4 | | 8.1M | 11.0B |
| EfficientDet-D3 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d3.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d3_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d3_coco_test-dev2017.txt)) | 45.9 | 45.8 | 65.0 | 49.3 | 26.6 | 49.4 | 59.8 | | 12.0M | 24.9B |
| EfficientDet-D4 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d4.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d4_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d4_coco_test-dev2017.txt)) | 49.0 | 49.4 | 69.0 | 53.4 | 30.3 | 53.2 | 63.2 | | 20.7M | 55.2B |
| EfficientDet-D5 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d5.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d5_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d5_coco_test-dev2017.txt)) | 50.5 | 50.7 | 70.2 | 54.7 | 33.2 | 53.9 | 63.2 | | 33.7M | 135.4B |
| EfficientDet-D6 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d6.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d6_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d6_coco_test-dev2017.txt)) | 51.3 | 51.7 | 71.2 | 56.0 | 34.1 | 55.2 | 64.1 | | 51.9M | 225.6B |
| EfficientDet-D7 ([ckpt](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/efficientdet-d7.tar.gz), [val](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/val/d7_coco_val.txt), [test-dev](https://storage.googleapis.com/cloud-tpu-checkpoints/efficientdet/coco/testdev/d7_coco_test-dev2017.txt)) | 51.8 | 52.2 | 71.4 | 56.3 | 34.8 | 55.5 | 64.6 | | 51.9M | 324.8B |

** <em>val</em> denotes validation results, <em>test-dev</em> denotes test-dev2017 results. AP<sup>val</sup> is for validation accuracy, all other AP results in the table are for COCO test-dev2017. All accuracy numbers are for single-model single-scale without ensemble or test-time augmentation. All checkpoints are trained with baseline preprocessing (no autoaugmentation).

## 3. Run inference.

Expand All @@ -71,7 +71,7 @@ We have provided a list of EfficientDet checkpoints and results as follows:

# Run inference.
!python model_inspect.py --runmode=infer --model_name=$MODEL \
--input_image_size=1920 --max_boxes_to_draw=100 --min_score_thresh=0.2 \
--input_image_size=1920x1280 --max_boxes_to_draw=100 --min_score_thresh=0.2 \
--ckpt_path=$CKPT_PATH --input_image=img.png --output_image_dir=/tmp
# you can visualize the output /tmp/0.jpg

Expand All @@ -88,13 +88,13 @@ You can also export a saved model, and use it to serve image inference.
# Step 1: export model
!python model_inspect.py --runmode=saved_model \
--model_name=efficientdet-d0 --ckpt_path=efficientdet-d0 \
--input_image_size=1920 --max_boxes_to_draw=100 --min_score_thresh=0.2 \
--input_image_size=1920x1280 --max_boxes_to_draw=100 --min_score_thresh=0.2 \
--saved_model_dir=/tmp/saved_model

# Step 2: do inference with saved model.
!python model_inspect.py --runmode=saved_model_infer \
--model_name=efficientdet-d0 --ckpt_path=efficientdet-d0 \
--input_image_size=1920 \
--input_image_size=1920x1280 \
--max_boxes_to_draw=100 --min_score_thresh=0.2 \
--line_thickness=4 --saved_model_dir=/tmp/saved_model \
--input_image=img.png --output_image_dir=/tmp/
Expand Down
2 changes: 1 addition & 1 deletion efficientdet/model_inspect.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -100,7 +100,7 @@ def __init__(self,
elif 'x' in image_size:
# image_size is in format of WIDTHxHEIGHT
width, height = image_size.split('x')
image_size = (int(width), int(height))
image_size = (int(height), int(width))
else:
# image_size is integer, witht the same width and height.
image_size = (int(image_size), int(image_size))
Expand Down

0 comments on commit e579bab

Please sign in to comment.