paddlex 3.0.0rc1__py3-none-any.whl → 3.0.2__py3-none-any.whl

paddlex/inference/utils/io/readers.py

@@ -26,8 +26,8 @@ from ....utils.deps import class_requires_deps, is_dep_available
 
 if is_dep_available("opencv-contrib-python"):
     import cv2
-if is_dep_available("PyMuPDF"):
-    import fitz
+if is_dep_available("pypdfium2"):
+    import pypdfium2 as pdfium
 if is_dep_available("soundfile"):
     import soundfile
 
@@ -96,7 +96,7 @@ class _BaseReader(object):
 class PDFReader(_BaseReader):
     """PDFReader"""
 
-    def __init__(self, backend="fitz", **bk_args):
+    def __init__(self, backend="pypdfium2", **bk_args):
         super().__init__(backend, **bk_args)
 
     def read(self, in_path):
@@ -244,19 +244,19 @@ class PILImageReaderBackend(_ImageReaderBackend):
         return ImageOps.exif_transpose(Image.open(in_path))
 
 
-@class_requires_deps("PyMuPDF", "opencv-contrib-python")
+@class_requires_deps("pypdfium2", "opencv-contrib-python")
 class PDFReaderBackend(_BaseReaderBackend):
 
-    def __init__(self, rotate=0, zoom_x=2.0, zoom_y=2.0):
+    def __init__(self, rotate=0, zoom=2.0):
         super().__init__()
-        self.mat = fitz.Matrix(zoom_x, zoom_y).prerotate(rotate)
+        self._rotation = rotate
+        self._scale = zoom
 
     def read_file(self, in_path):
-        for page in fitz.open(in_path):
-            pixmap = page.get_pixmap(matrix=self.mat, alpha=False)
-            img_cv = np.frombuffer(pixmap.samples, dtype=np.uint8).reshape(
-                pixmap.h, pixmap.w, pixmap.n
-            )
+        for page in pdfium.PdfDocument(in_path):
+            image = page.render(scale=self._scale, rotation=self._rotation).to_pil()
+            image = image.convert("RGB")
+            img_cv = np.array(image)
             img_cv = cv2.cvtColor(img_cv, cv2.COLOR_RGB2BGR)
             yield img_cv
 
@@ -361,7 +361,7 @@ class DecordVideoReaderBackend(_VideoReaderBackend):
         self.valid_mode = True
         self._fps = 0
 
-        # XXX(gaotingquan): There is a confict with `paddle` when import `decord` globally.
+        # XXX(gaotingquan): There is a conflict with `paddle` when import `decord` globally.
         try:
             import decord
 

paddlex/inference/utils/misc.py

@@ -0,0 +1,20 @@
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+def is_mkldnn_available():
+    # XXX: Not sure if this is the best way to check if MKL-DNN is available
+    from paddle.inference import Config
+
+    return hasattr(Config, "set_mkldnn_cache_capacity")

paddlex/inference/utils/mkldnn_blocklist.py

@@ -0,0 +1,59 @@
+# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+MKLDNN_BLOCKLIST = [
+    "LaTeX_OCR_rec",
+    "PP-FormulaNet-L",
+    "PP-FormulaNet-S",
+    "UniMERNet",
+    "UVDoc",
+    "Cascade-MaskRCNN-ResNet50-FPN",
+    "Cascade-MaskRCNN-ResNet50-vd-SSLDv2-FPN",
+    "Mask-RT-DETR-M",
+    "Mask-RT-DETR-S",
+    "MaskRCNN-ResNeXt101-vd-FPN",
+    "MaskRCNN-ResNet101-FPN",
+    "MaskRCNN-ResNet101-vd-FPN",
+    "MaskRCNN-ResNet50-FPN",
+    "MaskRCNN-ResNet50-vd-FPN",
+    "MaskRCNN-ResNet50",
+    "SOLOv2",
+    "PP-TinyPose_128x96",
+    "PP-TinyPose_256x192",
+    "Cascade-FasterRCNN-ResNet50-FPN",
+    "Cascade-FasterRCNN-ResNet50-vd-SSLDv2-FPN",
+    "Co-DINO-Swin-L",
+    "Co-Deformable-DETR-Swin-T",
+    "FasterRCNN-ResNeXt101-vd-FPN",
+    "FasterRCNN-ResNet101-FPN",
+    "FasterRCNN-ResNet101",
+    "FasterRCNN-ResNet34-FPN",
+    "FasterRCNN-ResNet50-FPN",
+    "FasterRCNN-ResNet50-vd-FPN",
+    "FasterRCNN-ResNet50-vd-SSLDv2-FPN",
+    "FasterRCNN-ResNet50",
+    "FasterRCNN-Swin-Tiny-FPN",
+    "MaskFormer_small",
+    "MaskFormer_tiny",
+    "SLANeXt_wired",
+    "SLANeXt_wireless",
+    "SLANet",
+    "SLANet_plus",
+    "YOWO",
+    "SAM-H_box",
+    "SAM-H_point",
+    "PP-FormulaNet_plus-L",
+    "PP-FormulaNet_plus-M",
+    "PP-FormulaNet_plus-S",
+]

paddlex/inference/utils/official_models.py

@@ -12,11 +12,22 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import os
+import shutil
+import tempfile
+from functools import lru_cache
 from pathlib import Path
 
+import huggingface_hub as hf_hub
+
+hf_hub.logging.set_verbosity_error()
+
+import requests
+
 from ...utils import logging
 from ...utils.cache import CACHE_DIR
 from ...utils.download import download_and_extract
+from ...utils.flags import MODEL_SOURCE
 
 OFFICIAL_MODELS = {
     "ResNet18": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/ResNet18_infer.tar",
@@ -37,6 +48,7 @@ OFFICIAL_MODELS = {
     "PP-LCNet_x0_75": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x0_75_infer.tar",
     "PP-LCNet_x1_0": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x1_0_infer.tar",
     "PP-LCNet_x1_0_doc_ori": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x1_0_doc_ori_infer.tar",
+    "PP-LCNet_x1_0_textline_ori": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x1_0_textline_ori_infer.tar",
     "PP-LCNet_x1_5": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x1_5_infer.tar",
     "PP-LCNet_x2_5": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x2_5_infer.tar",
     "PP-LCNet_x2_0": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-LCNet_x2_0_infer.tar",
@@ -250,6 +262,9 @@ PP-LCNet_x1_0_vehicle_attribute_infer.tar",
     "UniMERNet": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/UniMERNet_infer.tar",
     "PP-FormulaNet-S": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-FormulaNet-S_infer.tar",
     "PP-FormulaNet-L": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-FormulaNet-L_infer.tar",
+    "PP-FormulaNet_plus-S": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-FormulaNet_plus-S_infer.tar",
+    "PP-FormulaNet_plus-M": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-FormulaNet_plus-M_infer.tar",
+    "PP-FormulaNet_plus-L": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-FormulaNet_plus-L_infer.tar",
     "FasterRCNN-ResNet34-FPN": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/FasterRCNN-ResNet34-FPN_infer.tar",
     "FasterRCNN-ResNet50": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/FasterRCNN-ResNet50_infer.tar",
     "FasterRCNN-ResNet50-FPN": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/FasterRCNN-ResNet50-FPN_infer.tar",
@@ -331,24 +346,144 @@ PP-LCNet_x1_0_vehicle_attribute_infer.tar",
     "PP-DocLayout-L": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocLayout-L_infer.tar",
     "PP-DocLayout-M": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocLayout-M_infer.tar",
     "PP-DocLayout-S": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocLayout-S_infer.tar",
+    "PP-DocLayout_plus-L": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocLayout_plus-L_infer.tar",
+    "PP-DocBlockLayout": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocBlockLayout_infer.tar",
     "BEVFusion": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/BEVFusion_infer.tar",
     "YOLO-Worldv2-L": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/YOLO-Worldv2-L_infer.tar",
     "PP-DocBee-2B": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocBee-2B_infer.tar",
     "PP-DocBee-7B": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocBee-7B_infer.tar",
+    "PP-Chart2Table": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-Chart2Table_infer.tar",
+    "PP-OCRv5_server_det": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-OCRv5_server_det_infer.tar",
+    "PP-OCRv5_mobile_det": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-OCRv5_mobile_det_infer.tar",
+    "PP-OCRv5_server_rec": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/\
+PP-OCRv5_server_rec_infer.tar",
+    "PP-OCRv5_mobile_rec": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/\
+PP-OCRv5_mobile_rec_infer.tar",
+    "PP-DocBee2-3B": "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0.0/PP-DocBee2-3B_infer.tar",
 }
 
 
+HUGGINGFACE_MODELS = [
+    "arabic_PP-OCRv3_mobile_rec",
+    "chinese_cht_PP-OCRv3_mobile_rec",
+    "ch_RepSVTR_rec",
+    "ch_SVTRv2_rec",
+    "cyrillic_PP-OCRv3_mobile_rec",
+    "devanagari_PP-OCRv3_mobile_rec",
+    "en_PP-OCRv3_mobile_rec",
+    "en_PP-OCRv4_mobile_rec",
+    "japan_PP-OCRv3_mobile_rec",
+    "ka_PP-OCRv3_mobile_rec",
+    "korean_PP-OCRv3_mobile_rec",
+    "LaTeX_OCR_rec",
+    "latin_PP-OCRv3_mobile_rec",
+    "PicoDet_layout_1x",
+    "PicoDet_layout_1x_table",
+    "PicoDet-L_layout_17cls",
+    "PicoDet-L_layout_3cls",
+    "PicoDet-S_layout_17cls",
+    "PicoDet-S_layout_3cls",
+    "PP-DocBee2-3B",
+    "PP-DocBee-2B",
+    "PP-DocBee-7B",
+    "PP-DocBlockLayout",
+    "PP-DocLayout-L",
+    "PP-DocLayout-M",
+    "PP-DocLayout_plus-L",
+    "PP-DocLayout-S",
+    "PP-FormulaNet-L",
+    "PP-FormulaNet_plus-L",
+    "PP-FormulaNet_plus-M",
+    "PP-FormulaNet_plus-S",
+    "PP-FormulaNet-S",
+    "PP-LCNet_x1_0_doc_ori",
+    "PP-LCNet_x1_0_table_cls",
+    "PP-OCRv3_mobile_det",
+    "PP-OCRv3_mobile_rec",
+    "PP-OCRv3_server_det",
+    "PP-OCRv4_mobile_det",
+    "PP-OCRv4_mobile_rec",
+    "PP-OCRv4_mobile_seal_det",
+    "PP-OCRv4_server_det",
+    "PP-OCRv4_server_rec_doc",
+    "PP-OCRv4_server_rec",
+    "PP-OCRv4_server_seal_det",
+    "PP-OCRv5_mobile_det",
+    "PP-OCRv5_mobile_rec",
+    "PP-OCRv5_server_det",
+    "PP-OCRv5_server_rec",
+    "RT-DETR-H_layout_17cls",
+    "RT-DETR-H_layout_3cls",
+    "RT-DETR-L_wired_table_cell_det",
+    "RT-DETR-L_wireless_table_cell_det",
+    "SLANet",
+    "SLANet_plus",
+    "SLANeXt_wired",
+    "SLANeXt_wireless",
+    "ta_PP-OCRv3_mobile_rec",
+    "te_PP-OCRv3_mobile_rec",
+    "UniMERNet",
+    "UVDoc",
+]
+
+
+@lru_cache(1)
+def is_huggingface_accessible():
+    try:
+        response = requests.get("https://huggingface.co", timeout=1)
+        return response.ok == True
+    except requests.exceptions.RequestException as e:
+        return False
+
+
 class OfficialModelsDict(dict):
     """Official Models Dict"""
 
+    _save_dir = Path(CACHE_DIR) / "official_models"
+
     def __getitem__(self, key):
-        url = super().__getitem__(key)
-        save_dir = Path(CACHE_DIR) / "official_models"
+        def _download_from_bos():
+            url = super(OfficialModelsDict, self).__getitem__(key)
+            download_and_extract(url, self._save_dir, f"{key}", overwrite=False)
+            return self._save_dir / f"{key}"
+
+        def _download_from_hf():
+            local_dir = self._save_dir / f"{key}"
+            try:
+                if os.path.exists(local_dir):
+                    hf_hub.snapshot_download(
+                        repo_id=f"PaddlePaddle/{key}", local_dir=local_dir
+                    )
+                else:
+                    with tempfile.TemporaryDirectory() as td:
+                        temp_dir = os.path.join(td, "temp_dir")
+                        hf_hub.snapshot_download(
+                            repo_id=f"PaddlePaddle/{key}", local_dir=temp_dir
+                        )
+                        shutil.move(temp_dir, local_dir)
+            except Exception as e:
+                logging.warning(
+                    f"Encounter exception when download model from huggingface: \n{e}.\nPaddleX would try to download from BOS."
+                )
+                return _download_from_bos()
+            return local_dir
+
         logging.info(
-            f"Using official model ({key}), the model files will be automatically downloaded and saved in {save_dir}."
+            f"Using official model ({key}), the model files will be automatically downloaded and saved in {self._save_dir}."
         )
-        download_and_extract(url, save_dir, f"{key}", overwrite=False)
-        return save_dir / f"{key}"
+
+        if (
+            MODEL_SOURCE.lower() == "huggingface"
+            and is_huggingface_accessible()
+            and key in HUGGINGFACE_MODELS
+        ):
+            return _download_from_hf()
+        elif MODEL_SOURCE.lower() == "modelscope":
+            raise Exception(
+                f"ModelScope is not supported! Please use `HuggingFace` or `BOS`."
+            )
+        else:
+            return _download_from_bos()
 
 
 official_models = OfficialModelsDict(OFFICIAL_MODELS)

paddlex/inference/utils/pp_option.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import os
+from copy import deepcopy
 from typing import Dict, List
 
 from ...utils import logging
@@ -22,12 +23,34 @@ from ...utils.device import (
     parse_device,
     set_env_for_device_type,
 )
-from ...utils.flags import USE_PIR_TRT
+from ...utils.flags import (
+    DISABLE_MKLDNN_MODEL_BL,
+    DISABLE_TRT_MODEL_BL,
+    ENABLE_MKLDNN_BYDEFAULT,
+    USE_PIR_TRT,
+)
+from .misc import is_mkldnn_available
+from .mkldnn_blocklist import MKLDNN_BLOCKLIST
 from .new_ir_blocklist import NEWIR_BLOCKLIST
 from .trt_blocklist import TRT_BLOCKLIST
 from .trt_config import TRT_CFG_SETTING, TRT_PRECISION_MAP
 
 
+def get_default_run_mode(model_name, device_type):
+    if not model_name:
+        return "paddle"
+    if device_type != "cpu":
+        return "paddle"
+    if (
+        ENABLE_MKLDNN_BYDEFAULT
+        and is_mkldnn_available()
+        and model_name not in MKLDNN_BLOCKLIST
+    ):
+        return "mkldnn"
+    else:
+        return "paddle"
+
+
 class PaddlePredictorOption(object):
     """Paddle Inference Engine Option"""
 
@@ -44,7 +67,7 @@ class PaddlePredictorOption(object):
     )
     SUPPORT_DEVICE = ("gpu", "cpu", "npu", "xpu", "mlu", "dcu", "gcu")
 
-    def __init__(self, model_name, **kwargs):
+    def __init__(self, model_name=None, **kwargs):
         super().__init__()
         self._model_name = model_name
         self._cfg = {}
@@ -55,6 +78,10 @@ class PaddlePredictorOption(object):
     def model_name(self):
         return self._model_name
 
+    @model_name.setter
+    def model_name(self, model_name):
+        self._model_name = model_name
+
     @property
     def changed(self):
         return self._changed
@@ -64,6 +91,13 @@ class PaddlePredictorOption(object):
         assert isinstance(v, bool)
         self._changed = v
 
+    def copy(self):
+        obj = type(self)(self._model_name)
+        obj._cfg = deepcopy(self._cfg)
+        if hasattr(self, "trt_cfg_setting"):
+            obj.trt_cfg_setting = self.trt_cfg_setting
+        return obj
+
     def _init_option(self, **kwargs):
         for k, v in kwargs.items():
             if self._has_setter(k):
@@ -91,7 +125,7 @@ class PaddlePredictorOption(object):
         device_type, device_ids = parse_device(get_default_device())
 
         default_config = {
-            "run_mode": "paddle",
+            "run_mode": get_default_run_mode(self.model_name, device_type),
             "device_type": device_type,
             "device_id": None if device_ids is None else device_ids[0],
             "cpu_threads": 8,
@@ -106,6 +140,7 @@ class PaddlePredictorOption(object):
             "trt_dynamic_shape_input_data": None,  # only for trt
             "trt_shape_range_info_path": None,  # only for trt
             "trt_allow_rebuild_at_runtime": True,  # only for trt
+            "mkldnn_cache_capacity": 10,
         }
         return default_config
 
@@ -125,13 +160,35 @@ class PaddlePredictorOption(object):
             raise ValueError(
                 f"`run_mode` must be {support_run_mode_str}, but received {repr(run_mode)}."
             )
-        # TRT Blocklist
-        if run_mode.startswith("trt") and self._model_name in TRT_BLOCKLIST:
-            logging.warning(
-                f"The model({self._model_name}) is not supported to run in trt mode! Using `paddle` instead!"
-            )
+
+        if run_mode.startswith("mkldnn") and not is_mkldnn_available():
+            logging.warning("MKL-DNN is not available. Using `paddle` instead.")
             run_mode = "paddle"
 
+        # TODO: Check if trt is available
+
+        if self._model_name is not None:
+            # TRT Blocklist
+            if (
+                not DISABLE_TRT_MODEL_BL
+                and run_mode.startswith("trt")
+                and self._model_name in TRT_BLOCKLIST
+            ):
+                logging.warning(
+                    f"The model({self._model_name}) is not supported to run in trt mode! Using `paddle` instead!"
+                )
+                run_mode = "paddle"
+            # MKLDNN Blocklist
+            elif (
+                not DISABLE_MKLDNN_MODEL_BL
+                and run_mode.startswith("mkldnn")
+                and self._model_name in MKLDNN_BLOCKLIST
+            ):
+                logging.warning(
+                    f"The model({self._model_name}) is not supported to run in MKLDNN mode! Using `paddle` instead!"
+                )
+                run_mode = "paddle"
+
         self._update("run_mode", run_mode)
 
     @property
@@ -206,7 +263,7 @@ class PaddlePredictorOption(object):
         """set trt config"""
         assert isinstance(
             config, dict
-        ), f"The trt_cfg_setting must be `dict` type, but recived `{type(config)}` type!"
+        ), f"The trt_cfg_setting must be `dict` type, but received `{type(config)}` type!"
         self._update("trt_cfg_setting", config)
 
     @property
@@ -273,6 +330,14 @@ class PaddlePredictorOption(object):
     def trt_allow_rebuild_at_runtime(self, trt_allow_rebuild_at_runtime):
         self._update("trt_allow_rebuild_at_runtime", trt_allow_rebuild_at_runtime)
 
+    @property
+    def mkldnn_cache_capacity(self):
+        return self._cfg["mkldnn_cache_capacity"]
+
+    @mkldnn_cache_capacity.setter
+    def mkldnn_cache_capacity(self, capacity: int):
+        self._update("mkldnn_cache_capacity", capacity)
+
     # For backward compatibility
     # TODO: Issue deprecation warnings
     @property

paddlex/model.py

@@ -17,7 +17,7 @@ from copy import deepcopy
 from .inference import PaddlePredictorOption, create_predictor
 from .modules import (
     build_dataset_checker,
-    build_evaluater,
+    build_evaluator,
     build_exportor,
     build_trainer,
 )
@@ -119,7 +119,7 @@ class _ModelBasedConfig(_BaseModel):
         trainer.train()
 
     def evaluate(self):
-        evaluator = build_evaluater(self._config)
+        evaluator = build_evaluator(self._config)
         return evaluator.evaluate()
 
     def export(self):

paddlex/modules/__init__.py

@@ -14,7 +14,7 @@
 from importlib import import_module
 
 from .anomaly_detection import UadDatasetChecker, UadEvaluator, UadExportor, UadTrainer
-from .base import build_dataset_checker, build_evaluater, build_exportor, build_trainer
+from .base import build_dataset_checker, build_evaluator, build_exportor, build_trainer
 from .face_recognition import (
     FaceRecDatasetChecker,
     FaceRecEvaluator,

paddlex/modules/anomaly_detection/evaluator.py

@@ -26,7 +26,7 @@ class UadEvaluator(BaseEvaluator):
     entities = MODELS
 
     def update_config(self):
-        """update evalution config"""
+        """update evaluation config"""
         self.pdx_config.update_dataset(self.global_config.dataset_dir, "SegDataset")
         self.pdx_config.update_pretrained_weights(None, is_backbone=True)
 
@@ -47,7 +47,7 @@ class UadEvaluator(BaseEvaluator):
         return config_path
 
     def get_eval_kwargs(self) -> dict:
-        """get key-value arguments of model evalution function
+        """get key-value arguments of model evaluation function
 
         Returns:
             dict: the arguments of evaluation function.

paddlex/modules/base/__init__.py

@@ -13,6 +13,6 @@
 # limitations under the License.
 
 from .dataset_checker import BaseDatasetChecker, build_dataset_checker
-from .evaluator import BaseEvaluator, build_evaluater
+from .evaluator import BaseEvaluator, build_evaluator
 from .exportor import BaseExportor, build_exportor
 from .trainer import BaseTrainer, build_trainer

paddlex/modules/base/evaluator.py

@@ -27,14 +27,14 @@ from ...utils.misc import AutoRegisterABCMetaClass
 from .build_model import build_model
 
 
-def build_evaluater(config: AttrDict) -> "BaseEvaluator":
-    """build model evaluater
+def build_evaluator(config: AttrDict) -> "BaseEvaluator":
+    """build model evaluator
 
     Args:
         config (AttrDict): PaddleX pipeline config, which is loaded from pipeline yaml file.
 
     Returns:
-        BaseEvaluator: the evaluater, which is subclass of BaseEvaluator.
+        BaseEvaluator: the evaluator, which is subclass of BaseEvaluator.
     """
     model_name = config.Global.model
     try:
@@ -161,10 +161,10 @@ evaling!"
 
     @abstractmethod
     def update_config(self):
-        """update evalution config"""
+        """update evaluation config"""
         raise NotImplementedError
 
     @abstractmethod
     def get_eval_kwargs(self):
-        """get key-value arguments of model evalution function"""
+        """get key-value arguments of model evaluation function"""
         raise NotImplementedError

paddlex/modules/base/trainer.py

@@ -21,7 +21,7 @@ from ...utils.device import (
     set_env_for_device,
     update_device_num,
 )
-from ...utils.flags import FLAGS_json_format_model, DISABLE_CINN_MODEL_WL
+from ...utils.flags import DISABLE_CINN_MODEL_WL, FLAGS_json_format_model
 from ...utils.misc import AutoRegisterABCMetaClass
 from .build_model import build_model
 from .utils.cinn_setting import CINN_WHITELIST, enable_cinn_backend

paddlex/modules/doc_vlm/dataset_checker.py

@@ -18,12 +18,12 @@ from .model_list import MODELS
 
 
 class DocVLMDatasetChecker(BaseDatasetChecker):
-    """Dataset Checker for Document Vision Languange Model"""
+    """Dataset Checker for Document Vision Language Model"""
 
     entities = MODELS
 
     def __init__(self, config):
         # not support for now
         raise UnsupportedAPIError(
-            "document vision languange models do not support data check for now."
+            "document vision language models do not support data check for now."
         )

paddlex/modules/doc_vlm/evaluator.py

@@ -18,12 +18,12 @@ from .model_list import MODELS
 
 
 class DocVLMEvaluator(BaseEvaluator):
-    """Document Vision Languange Model Evaluator"""
+    """Document Vision Language Model Evaluator"""
 
     entities = MODELS
 
     def __init__(self, config):
         # not support for now
         raise UnsupportedAPIError(
-            "document vision languange models do not support evaluate for now."
+            "document vision language models do not support evaluate for now."
         )

paddlex/modules/doc_vlm/exportor.py

@@ -18,12 +18,12 @@ from .model_list import MODELS
 
 
 class DocVLMExportor(BaseExportor):
-    """Document Vision Languange Model Exportor"""
+    """Document Vision Language Model Exportor"""
 
     entities = MODELS
 
     def __init__(self, config):
         # not support for now
         raise UnsupportedAPIError(
-            "document vision languange models do not support export for now."
+            "document vision language models do not support export for now."
         )

paddlex/modules/doc_vlm/model_list.py

@@ -13,4 +13,4 @@
 # limitations under the License.
 
 
-MODELS = ["PP-DocBee-2B", "PP-DocBee-7B"]
+MODELS = ["PP-DocBee-2B", "PP-DocBee-7B", "PP-Chart2Table", "PP-DocBee2-3B"]

paddlex/modules/doc_vlm/trainer.py

@@ -18,14 +18,14 @@ from .model_list import MODELS
 
 
 class DocVLMTrainer(BaseTrainer):
-    """Document Vision Languange Model Trainer"""
+    """Document Vision Language Model Trainer"""
 
     entities = MODELS
 
     def __init__(self, config):
         # not support for now
         raise UnsupportedAPIError(
-            "Document vision languange models do not support train for now."
+            "Document vision language models do not support train for now."
         )
 
     def update_config(self):