mirrors
/
mmocr
mirror of https://github.com/open-mmlab/mmocr.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Add metafiles (#179) 

* add metafiles

* add metafiles

* add metafiles

* fix metafile errors

* fix metafiles

* Update metafile.yml

Co-authored-by: Hongbin Sun <hongbin306@gmail.com>
pull/180/head
quincylin1 2021-05-13 10:46:26 +08:00 committed by GitHub
parent ee700e94cd
commit cc27d8254b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
13 changed files with 591 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
configs/kie/sdmgr/metafile.yml 100644
View File

@ -0,0 +1,36 @@
Collections:
- Name: SDMGR
Metadata:
Training Data: KIEDataset
Training Techniques:
- Adam
Training Resources: 1x NVIDIA 1080Ti GPU
Architecture:
- UNet
Paper: https://arxiv.org/abs/2103.14470.pdf
README: configs/kie/sdmgr/README.md
Models:
- Name: sdmgr_unet16_60e_wildreceipt
In Collection: SDMGR
Config: configs/kie/sdmgr/sdmgr_unet16_60e_wildreceipt.py
Metadata:
Training Data: wildreceipt
Results:
- Task: Key Information Extraction
Dataset: wildreceipt
Metrics:
macro_f1: 0.876
Weights: https://download.openmmlab.com/mmocr/kie/sdmgr/sdmgr_unet16_60e_wildreceipt_20210405-16a47642.pth
- Name: sdmgr_novisual_60e_wildreceipt
In Collection: SDMGR
Config: configs/kie/sdmgr/sdmgr_novisual_60e_wildreceipt.py
Metadata:
Training Data: wildreceipt
Results:
- Task: Key Information Extraction
Dataset: wildreceipt
Metrics:
macro_f1: 0.864
Weights: https://download.openmmlab.com/mmocr/kie/sdmgr/sdmgr_novisual_60e_wildreceipt_20210405-07bc26ad.pth

37
configs/textdet/dbnet/metafile.yml 100644
View File

@ -0,0 +1,37 @@
Collections:
- Name: DBNet
Metadata:
Training Data: ICDAR2015
Training Techniques:
- SGD with Momentum
Training Resources: 8x NVIDIA V100 GPUs
Architecture:
- ResNet
- FPNC
Paper: https://arxiv.org/pdf/1911.08947.pdf
README: configs/textdet/dbnet/README.md
Models:
- Name: dbnet_r18_fpnc_1200e_icdar2015
In Collection: DBNet
Config: configs/textdet/dbnet/dbnet_r18_fpnc_1200e_icdar2015.py
Metadata:
Training Data: ICDAR2015
Results:
- Task: Instance Segmentation
Dataset: ICDAR2015
Metrics:
hmean-iou: 0.795
Weights: https://download.openmmlab.com/mmocr/textdet/dbnet/dbnet_r18_fpnc_sbn_1200e_icdar2015_20210329-ba3ab597.pth
- Name: dbnet_r50dcnv2_fpnc_1200e_icdar2015
In Collection: DBNet
Config: configs/textdet/dbnet/dbnet_r50dcnv2_fpnc_1200e_icdar2015.py
Metadata:
Training Data: ICDAR2015
Results:
- Task: Instance Segmentation
Dataset: ICDAR2015
Metrics:
hmean-iou: 0.830
Weights: https://download.openmmlab.com/mmocr/textdet/dbnet/dbnet_r50dcnv2_fpnc_sbn_1200e_icdar2015_20210325-91cef9af.pth

49
configs/textdet/maskrcnn/metafile.yml 100644
View File

@ -0,0 +1,49 @@
Collections:
- Name: Mask R-CNN
Metadata:
Training Data: ICDARDataset
Training Techniques:
- SGD with Momentum
Training Resources: 8x NVIDIA 1080Ti GPUs
Architecture:
- RoI Align
- RPN
Paper: https://arxiv.org/pdf/1703.06870.pdf
README: configs/textdet/maskrcnn/README.md
Models:
- Name: mask_rcnn_r50_fpn_160e_ctw1500
In Collection: Mask R-CNN
Config: configs/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_ctw1500.py
Metadata:
Training Data: CTW1500
Results:
- Task: Instance Segmentation
Dataset: CTW1500
Metrics:
hmean: 0.732
Weights: https://download.openmmlab.com/mmocr/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_ctw1500_20210219-96497a76.pth
- Name: mask_rcnn_r50_fpn_160e_icdar2015
In Collection: Mask R-CNN
Config: configs/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_icdar2015.py
Metadata:
Training Data: ICDAR2015
Results:
- Task: Instance Segmentation
Dataset: ICDAR2015
Metrics:
hmean: 0.825
Weights: https://download.openmmlab.com/mmocr/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_icdar2015_20210219-8eb340a3.pth
- Name: mask_rcnn_r50_fpn_160e_icdar2017
In Collection: Mask R-CNN
Config: configs/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_icdar2017.py
Metadata:
Training Data: ICDAR2017
Results:
- Task: Instance Segmentation
Dataset: ICDAR2017
Metrics:
hmean: 0.789
Weights: https://download.openmmlab.com/mmocr/textdet/maskrcnn/mask_rcnn_r50_fpn_160e_icdar2017_20210218-c6ec3ebb.pth

49
configs/textdet/panet/metafile.yml 100644
View File

@ -0,0 +1,49 @@
Collections:
- Name: PANet
Metadata:
Training Data: ICDARDataset
Training Techniques:
- Adam
Training Resources: 8x NVIDIA V100 GPUs
Architecture:
- ResNet
- FPEM_FFM
Paper: https://arxiv.org/pdf/1803.01534.pdf
README: configs/textdet/panet/README.md
Models:
- Name: panet_r18_fpem_ffm_600e_ctw1500
In Collection: PANet
Config: configs/textdet/panet/panet_r18_fpem_ffm_600e_ctw1500.py
Metadata:
Training Data: CTW1500
Results:
- Task: Instance Segmentation
Dataset: CTW1500
Metrics:
hmean-iou: 0.806
Weights: https://download.openmmlab.com/mmocr/textdet/panet/panet_r18_fpem_ffm_sbn_600e_ctw1500_20210219-3b3a9aa3.pth
- Name: panet_r18_fpem_ffm_600e_icdar2015
In Collection: PANet
Config: configs/textdet/panet/panet_r18_fpem_ffm_600e_icdar2015.py
Metadata:
Training Data: ICDAR2015
Results:
- Task: Instance Segmentation
Dataset: ICDAR2015
Metrics:
hmean-iou: 0.791
Weights: https://download.openmmlab.com/mmocr/textdet/panet/panet_r18_fpem_ffm_sbn_600e_icdar2015_20210219-42dbe46a.pth
- Name: panet_r50_fpem_ffm_600e_icdar2017
In Collection: PANet
Config: configs/textdet/panet/panet_r50_fpem_ffm_600e_icdar2017.py
Metadata:
Training Data: ICDAR2017
Results:
- Task: Instance Segmentation
Dataset: ICDAR2017
Metrics:
hmean-iou:
Weights:

49
configs/textdet/psenet/metafile.yml 100644
View File

@ -0,0 +1,49 @@
Collections:
- Name: PSENet
Metadata:
Training Data: ICDARDataset
Training Techniques:
- Adam
Training Resources: 8x NVIDIA 1080Ti GPUs
Architecture:
- ResNet
- FPNF
Paper: https://arxiv.org/abs/1806.02559.pdf
README: configs/textdet/psenet/README.md
Models:
- Name: psenet_r50_fpnf_600e_ctw1500
In Collection: PSENet
Config: configs/textdet/psenet/psenet_r50_fpnf_600e_ctw1500.py
Metadata:
Training Data: CTW1500
Results:
- Task: Instance Segmentation
Dataset: CTW1500
Metrics:
hmean-iou: 0.784
Weights: https://download.openmmlab.com/mmocr/textdet/psenet/psenet_r50_fpnf_600e_ctw1500_20210401-216fed50.pth
- Name: psenet_r50_fpnf_600e_icdar2015
In Collection: PSENet
Config: configs/textdet/psenet/psenet_r50_fpnf_600e_icdar2015.py
Metadata:
Training Data: ICDAR2015
Results:
- Task: Instance Segmentation
Dataset: ICDAR2015
Metrics:
hmean-iou: 0.807
Weights: https://download.openmmlab.com/mmocr/textdet/psenet/psenet_r50_fpnf_600e_icdar2015-c6131f0d.pth
- Name: psenet_r50_fpnf_600e_icdar2015_with_pretrain
In Collection: PSENet
Config: configs/textdet/psenet/psenet_r50_fpnf_600e_icdar2015.py
Metadata:
Training Data: ICDAR2017
Results:
- Task: Instance Segmentation
Dataset: ICDAR2017
Metrics:
hmean-iou: 0.847
Weights: https://download.openmmlab.com/mmocr/textdet/psenet/psenet_r50_fpnf_600e_icdar2015_pretrain-eefd8fe6.pth

25
configs/textdet/textsnake/metafile.yml 100644
View File

@ -0,0 +1,25 @@
Collections:
- Name: TextSnake
Metadata:
Training Data: ICDARDataset
Training Techniques:
- SGD with Momentum
Training Resources: 8x NVIDIA V100 GPUs
Architecture:
- ResNet
- FPN_UNET
Paper: https://arxiv.org/abs/1807.01544.pdf
README: configs/textdet/textsnake/README.md
Models:
- Name: textsnake_r50_fpn_unet_1200e_ctw1500
In Collection: TextSnake
Config: configs/textdet/textsnake/textsnake_r50_fpn_unet_1200e_ctw1500.py
Metadata:
Training Data: CTW1500
Results:
- Task: Instance Segmentation
Dataset: CTW1500
Metrics:
hmean-iou: 0.817
Weights: https://download.openmmlab.com/mmocr/textdet/textsnake/textsnake_r50_fpn_unet_1200e_ctw1500-27f65b64.pth

31
configs/textrecog/crnn/metafile.yml 100644
View File

@ -0,0 +1,31 @@
Collections:
- Name: CRNN
Metadata:
Training Data: OCRDataset
Training Techniques:
- Adadelta
Training Resources: 4x NVIDIA 1080Ti GPUs
Architecture:
- VeryDeepVgg
- CRNNDecoder
Paper: https://arxiv.org/pdf/1507.05717.pdf
README: configs/textrecog/crnn/README.md
Models:
- Name: crnn_academic_dataset
In Collection: CRNN
Config: configs/textrecog/crnn/crnn_academic_dataset.py
Metadata:
Training Data: Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 80.5
- 81.5
- 86.5
Weights: https://download.openmmlab.com/mmocr/textrecog/crnn/crnn_academic-a723a1c5.pth

74
configs/textrecog/nrtr/metafile.yml 100644
View File

@ -0,0 +1,74 @@
Collections:
- Name: NRTR
Metadata:
Training Data: OCRDataset
Training Techniques:
- Adam
Training Resources: 64x NVIDIA 1080Ti GPUs
Architecture:
- ResNet31OCR
- TFEncoder
- TFDecoder
Paper: https://arxiv.org/pdf/1806.00926.pdf
README: configs/textrecog/nrtr/README.md
Models:
- Name: nrtr_r31_1by16_1by8_academic
In Collection: NRTR
Config: configs/textrecog/nrtr/nrtr_r31_1by16_1by8_academic.py
Metadata:
Training Data:
- SynthText
- Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 93.9
- 90.0
- 93.5
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 74.5
- 78.5
- 86.5
Weights: https://download.openmmlab.com/mmocr/textrecog/nrtr/nrtr_r31_academic_20210406-954db95e.pth
- Name: nrtr_r31_1by8_1by4_academic
In Collection: NRTR
Config: configs/textrecog/nrtr/nrtr_r31_1by8_1by4_academic.py
Metadata:
Training Data:
- SynthText
- Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 94.7
- 87.5
- 93.3
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 75.1
- 78.9
- 87.9
Weights: https://download.openmmlab.com/mmocr/textrecog/nrtr/nrtr_r31_1by8_1by4_academic_20210406-ce16e7cc.pth

50
configs/textrecog/robust_scanner/metafile.yml 100644
View File

@ -0,0 +1,50 @@
Collections:
- Name: RobustScanner
Metadata:
Training Data: OCRDataset
Training Techniques:
- Adam
Training Resources: 16x NVIDIA 1080Ti GPUs
Architecture:
- ResNet31OCR
- ChannelReductionEncoder
- RobustScannerDecoder
Paper: https://arxiv.org/pdf/2007.07542.pdf
README: configs/textrecog/robust_scanner/README.md
Models:
- Name: robustscanner_r31_academic
In Collection: RobustScanner
Config: configs/textrecog/robust_scanner/robustscanner_r31_academic.py
Metadata:
Training Data:
- ICDAR2011
- ICDAR2013
- ICDAR2015
- COCO text
- IIIT5K
- SynthText
- SynthAdd
- Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 95.1
- 89.2
- 93.1
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 77.8
- 80.3
- 90.3
Weights: https://download.openmmlab.com/mmocr/textrecog/robustscanner/robustscanner_r31_academic-5f05874f.pth

98
configs/textrecog/sar/metafile.yml 100644
View File

@ -0,0 +1,98 @@
Collections:
- Name: SAR
Metadata:
Training Data: OCRDataset
Training Techniques:
- Adam
Training Resources: 48x NVIDIA 1080Ti GPUs
Architecture:
- ResNet31OCR
- SAREncoder
- ParallelSARDecoder
Paper: https://arxiv.org/pdf/1811.00751.pdf
README: configs/textrecog/sar/README.md
Models:
- Name: sar_r31_parallel_decoder_academic
In Collection: SAR
Config: configs/textrecog/sar/sar_r31_parallel_decoder_academic.py
Metadata:
Training Data:
- ICDAR2011
- ICDAR2013
- ICDAR2015
- COCO text
- IIIT5K
- SynthText
- SynthAdd
- Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 95.0
- 89.6
- 93.7
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 79.0
- 82.2
- 88.9
Weights: https://download.openmmlab.com/mmocr/textrecog/crnn/crnn_academic-a723a1c5.pth
- Name: sar_r31_sequential_decoder_academic
In Collection: SAR
Config: configs/textrecog/sar/sar_r31_sequential_decoder_academic.py
Metadata:
Training Data:
- ICDAR2011
- ICDAR2013
- ICDAR2015
- COCO text
- IIIT5K
- SynthText
- SynthAdd
- Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 95.2
- 88.7
- 92.4
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 78.2
- 81.9
- 89.6
Weights: https://download.openmmlab.com/mmocr/textrecog/sar/sar_r31_sequential_decoder_academic-d06c9a8e.pth
- Name: sar_r31_parallel_decoder_chinese
In Collection: SAR
Config: configs/textrecog/sar/sar_r31_parallel_decoder_chinese.py
Metadata:
Training Data:
Results:
- Task:
Dataset:
Metrics:
acc:
Weights: https://download.openmmlab.com/mmocr/textrecog/sar/sar_r31_parallel_decoder_chineseocr_20210507-b4be8214.pth

35
configs/textrecog/seg/metafile.yml 100644
View File

@ -0,0 +1,35 @@
Collections:
- Name: SEG
Metadata:
Training Data: mixture
Training Techniques:
- Adam
Training Resources: 4x NVIDIA 1080Ti GPUs
Architecture:
- ResNet31OCR
- FPNOCR
Paper:
README: configs/textrecog/seg/README.md
Models:
- Name: seg_r31_1by16_fpnocr_academic
In Collection: SEG
Config: configs/textrecog/seg/seg_r31_1by16_fpnocr_academic.py
Metadata:
Training Data: SynthText
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 90.9
- 81.8
- 90.7
- Task: Irregular Text Recognition
Dataset: CT80
Metrics:
acc: 80.9
Weights: https://download.openmmlab.com/mmocr/textrecog/seg/seg_r31_1by16_fpnocr_academic-72235b11.pth

45
configs/textrecog/tps/metafile.yml 100644
View File

@ -0,0 +1,45 @@
Collections:
- Name: TPS
Metadata:
Training Data: OCRDataset
Training Techniques:
- Adadelta
Training Resources: 4x NVIDIA 1080Ti GPUs
Architecture:
- TPSPreprocessor
- VeryDeepVgg
- CRNNDecoder
- CTCLoss
Paper:
- https://arxiv.org/pdf/1507.05717.pdf
- https://arxiv.org/pdf/1603.03915.pdf
README: configs/textrecog/tps/README.md
Models:
- Name: crnn_tps_academic_dataset
In Collection: TPS-CRNN
Config: configs/textrecog/tps/crnn_tps_academic_dataset.py
Metadata:
Training Data: Syn90k
Results:
- Task: Regular Text Recognition
Dataset:
- IIIT5K
- SVT
- ICDAR2013
Metrics:
acc:
- 80.8
- 81.3
- 85.0
- Task: Irregular Text Recognition
Dataset:
- ICDAR2015
- SVTP
- CT80
Metrics:
acc:
- 59.6
- 68.1
- 53.8
Weights: https://download.openmmlab.com/mmocr/textrecog/tps/crnn_tps_academic_dataset_20210510-d221a905.pth

13
model_zoo.yml 100644
View File

@ -0,0 +1,13 @@
Import:
- configs/textdet/dbnet/metafile.yml
- configs/textdet/maskrcnn/metafile.yml
- configs/textdet/panet/metafile.yml
- configs/textdet/psenet/metafile.yml
- configs/textdet/textsnake/metafile.yml
- configs/textrecog/crnn/metafile.yml
- configs/textrecog/nrtr/metafile.yml
- configs/textrecog/robust_scanner/metafile.yml
- configs/textrecog/sar/metafile.yml
- configs/textrecog/seg/metafile.yml
- configs/textrecog/tps/metafile.yml
- configs/kie/sdmgr/metafile.yml