feat(rapidocr): add ppocr v5 torch version (#511)

* feat: support torch ppocrv5 det

* feat: support torch ppocrv5 rec

* chore: update files

* chore: update files

* test: add ppocrv5 torch unit testings

* fix: fix typos

Author: SWHL

python/rapidocr/default_models.yaml

@@ -385,6 +385,27 @@ torch:
       ch_ptocr_mobile_v2.0_cls_infer.pth:
         model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/v3.1.0/torch/PP-OCRv4/cls/ch_ptocr_mobile_v2.0_cls_infer.pth
         SHA256: bfe13860824b3365c0c7f7ccfcddc8ff11645c60051739ff18bc9913f60c98e1
+  PP-OCRv5:
+    det:
+      ch_PP-OCRv5_det_mobile_infer.pth:
+        model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/torch/PP-OCRv5/det/ch_PP-OCRv5_det_mobile_infer.pth
+        SHA256: df848ed5060bac4d0f6e58572aea97d92e909a8a87cf292849237b0e84f6ffdb
+      ch_PP-OCRv5_det_server_infer.pth:
+        model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/torch/PP-OCRv5/det/ch_PP-OCRv5_det_server_infer.pth
+        SHA256: 607403c8233149f61c39f8e3302270625fab9fa2852c4329fc3c8f49cdb41e70
+    rec:
+      ch_PP-OCRv5_rec_mobile_infer.pth:
+        model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/torch/PP-OCRv5/rec/ch_PP-OCRv5_rec_mobile_infer.pth
+        dict_url: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/v3.1.0/paddle/PP-OCRv5/rec/ch_PP-OCRv5_rec_mobile_infer/ppocrv5_dict.txt
+        SHA256: d20ee8dac2ca63e2d1989b02ecc42595c71d61bf8dd8c8ddc5ad2ee68e7b5be2
+      ch_PP-OCRv5_rec_server_infer.pth:
+        model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/torch/PP-OCRv5/rec/ch_PP-OCRv5_rec_server_infer.pth
+        dict_url: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/v3.1.0/paddle/PP-OCRv5/rec/ch_PP-OCRv5_rec_server_infer/ppocrv5_dict.txt
+        SHA256: 4767ddc90c1532ec01d881a980dae0a0b92679f4f82f88c4e9f92563de69e740
+    cls:
+      ch_ptocr_mobile_v2.0_cls_infer.pth:
+        model_dir: https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/v3.1.0/torch/PP-OCRv4/cls/ch_ptocr_mobile_v2.0_cls_infer.pth
+        SHA256: bfe13860824b3365c0c7f7ccfcddc8ff11645c60051739ff18bc9913f60c98e1
 
 fonts:
   ch:

python/rapidocr/networks/arch_config.yaml

@@ -1,3 +1,88 @@
+ch_PP-OCRv5_rec_mobile_infer:
+  model_type: rec
+  algorithm: SVTR_LCNet
+  Transform:
+  Backbone:
+    name: PPLCNetV3
+    scale: 0.95
+  Head:
+    name: MultiHead
+    out_channels_list:
+      CTCLabelDecode: 18385 #'blank' + ...(18384) + ' '
+    head_list:
+      - CTCHead:
+          Neck:
+            name: svtr
+            dims: 120
+            depth: 2
+            hidden_dims: 120
+            kernel_size: [1, 3]
+            use_guide: True
+          Head:
+            fc_decay: 0.00001
+      - NRTRHead:
+          nrtr_dim: 384
+          max_text_length: 25
+
+ch_PP-OCRv5_rec_server_infer:
+  model_type: rec
+  algorithm: SVTR_HGNet
+  Transform:
+  Backbone:
+    name: PPHGNetV2_B4
+    text_rec: True
+  Head:
+    name: MultiHead
+    out_channels_list:
+      CTCLabelDecode: 18385 #'blank' + ...(18384) + ' '
+    head_list:
+      - CTCHead:
+          Neck:
+            name: svtr
+            dims: 120
+            depth: 2
+            hidden_dims: 120
+            kernel_size: [1, 3]
+            use_guide: True
+          Head:
+            fc_decay: 0.00001
+      - NRTRHead:
+          nrtr_dim: 384
+          max_text_length: 25
+
+ch_PP-OCRv5_det_mobile_infer:
+  model_type: det
+  algorithm: DB
+  Transform: null
+  Backbone:
+    name: PPLCNetV3
+    scale: 0.75
+    det: True
+  Neck:
+    name: RSEFPN
+    out_channels: 96
+    shortcut: True
+  Head:
+    name: DBHead
+    k: 50
+    fix_nan: True
+
+ch_PP-OCRv5_det_server_infer:
+  model_type: det
+  algorithm: DB
+  Transform: null
+  Backbone:
+    name: PPHGNetV2_B4
+    det: True
+  Neck:
+    name: LKPAN
+    out_channels: 256
+    intracl: true
+  Head:
+    name: PFHeadLocal
+    k: 50
+    mode: "large"
+
 ch_ptocr_mobile_v2.0_cls_infer:
   model_type: cls
   algorithm: CLS

python/rapidocr/networks/backbones/__init__.py

@@ -20,6 +20,7 @@ def build_backbone(config, model_type):
         from .det_mobilenet_v3 import MobileNetV3
         from .rec_hgnet import PPHGNet_small
         from .rec_lcnetv3 import PPLCNetV3
+        from .rec_pphgnetv2 import PPHGNetV2_B4
 
         support_dict = [
             "MobileNetV3",
@@ -28,13 +29,15 @@ def build_backbone(config, model_type):
             "ResNet_SAST",
             "PPLCNetV3",
             "PPHGNet_small",
+            "PPHGNetV2_B4",
         ]
     elif model_type == "rec" or model_type == "cls":
         from .rec_hgnet import PPHGNet_small
         from .rec_lcnetv3 import PPLCNetV3
         from .rec_mobilenet_v3 import MobileNetV3
-        from .rec_svtrnet import SVTRNet
         from .rec_mv1_enhance import MobileNetV1Enhance
+        from .rec_pphgnetv2 import PPHGNetV2_B4
+        from .rec_svtrnet import SVTRNet
 
         support_dict = [
             "MobileNetV1Enhance",
@@ -48,6 +51,7 @@ def build_backbone(config, model_type):
             "DenseNet",
             "PPLCNetV3",
             "PPHGNet_small",
+            "PPHGNetV2_B4",
         ]
     else:
         raise NotImplementedError