paddlex 3.0.0rc0__py3-none-any.whl → 3.0.0rc1__py3-none-any.whl

paddlex/.version

@@ -1 +1 @@
-3.0.0.rc0
+3.0.0.rc1

paddlex/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,30 +12,25 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from .utils.lazy_loader import LazyLoader
-import sys
-
-paddle = LazyLoader("lazy_paddle", globals(), "paddle")
-sys.modules["lazy_paddle"] = paddle
-
 import os
+import sys
 
+_SPECIAL_MODS = ["paddle", "paddle_custom_device", "ultra_infer"]
+_loaded_special_mods = []
+for mod in _SPECIAL_MODS:
+    if mod in sys.modules:
+        _loaded_special_mods.append(mod)
 
 from . import version
-from .modules import (
-    build_dataset_checker,
-    build_trainer,
-    build_evaluater,
-)
+from .inference import create_pipeline, create_predictor
 from .model import create_model
-from .inference import create_predictor, create_pipeline
+from .modules import build_dataset_checker, build_evaluater, build_trainer
 
 
 def _initialize():
-    from .utils.logging import setup_logging
+    from . import repo_apis, repo_manager
     from .utils import flags
-    from . import repo_manager
-    from . import repo_apis
+    from .utils.logging import setup_logging
 
     __DIR__ = os.path.abspath(os.path.dirname(os.path.abspath(__file__)))
     repo_manager.set_parent_dirs(
@@ -48,25 +43,13 @@ def _initialize():
         repo_manager.initialize()
 
 
-def _check_paddle_version():
-    """check paddle version"""
-
-    supported_versions = ["3.0", "0.0"]
-    device_type = paddle.device.get_device().split(":")[0]
-    if device_type.lower() == "xpu":
-        supported_versions.append("2.6")
-    version = paddle.__version__
-    # Recognizable version number: major.minor.patch
-    major, minor, patch = version.split(".")
-    # Ignore patch
-    version = f"{major}.{minor}"
-    if version not in supported_versions:
-        raise RuntimeError(
-            f"The {version} version of PaddlePaddle is not supported. "
-            f"Please install one of the following versions of PaddlePaddle: {supported_versions}."
-        )
+__version__ = version.get_pdx_version()
 
 
 _initialize()
 
-__version__ = version.get_pdx_version()
+for mod in _SPECIAL_MODS:
+    if mod in sys.modules and mod not in _loaded_special_mods:
+        raise AssertionError(
+            f"`{mod}` is unexpectedly loaded. Please contact the PaddleX team to report this issue."
+        )

paddlex/__main__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.

paddlex/configs/modules/doc_vlm/PP-DocBee-2B.yaml

@@ -0,0 +1,14 @@
+Global:
+  model: PP-DocBee-2B
+  mode: predict # only support predict
+  device: gpu:0
+  output: "output"
+
+Predict:
+  batch_size: 1
+  model_dir: "/path/to/PP-DocBee-2B"
+  input:
+    image: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/medal_table.png"
+    query: "识别这份表格的内容"
+  kernel_option:
+    run_mode: paddle

paddlex/configs/modules/doc_vlm/PP-DocBee-7B.yaml

@@ -0,0 +1,14 @@
+Global:
+  model: PP-DocBee-7B
+  mode: predict # only support predict
+  device: gpu:0
+  output: "output"
+
+Predict:
+  batch_size: 1
+  model_dir: "/path/to/PP-DocBee-7B"
+  input:
+    image: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/medal_table.png"
+    query: "识别这份表格的内容"
+  kernel_option:
+    run_mode: paddle

paddlex/configs/modules/open_vocabulary_detection/YOLO-Worldv2-L.yaml

@@ -0,0 +1,13 @@
+Global:
+  model: YOLO-Worldv2-L
+  mode: predict # only support predict
+  device: gpu:0
+  output: "output"
+
+Predict:
+  batch_size: 1
+  model_dir: "/ChartQA/PaddleX_refactor/PaddleX/_zzl_test_yolow/YOLO-Worldv2-L"
+  input: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/open_vocabulary_detection.jpg"
+  prompt: "bus, car, rearview mirror"
+  kernel_option:
+    run_mode: paddle

paddlex/configs/pipelines/anomaly_detection.yaml

@@ -5,4 +5,4 @@ SubModules:
     module_name: anomaly_detection
     model_name: STFPM
     model_dir: null
-    batch_size: 1   
+    batch_size: 1   

paddlex/configs/pipelines/doc_understanding.yaml

@@ -0,0 +1,9 @@
+
+pipeline_name: doc_understanding
+
+SubModules:
+  DocUnderstanding:
+    module_name: doc_vlm
+    model_name: PP-DocBee-2B
+    model_dir: null
+    batch_size: 1

paddlex/configs/pipelines/ts_anomaly_detection.yaml

@@ -5,4 +5,4 @@ SubModules:
     module_name: ts_anomaly_detection
     model_name: DLinear_ad
     model_dir: null
-    batch_size: 1   
+    batch_size: 1   

paddlex/configs/pipelines/ts_classification.yaml

@@ -5,4 +5,4 @@ SubModules:
     module_name: ts_classification
     model_name: TimesNet_cls
     model_dir: null
-    batch_size: 1   
+    batch_size: 1   

paddlex/configs/pipelines/ts_forecast.yaml

@@ -5,4 +5,4 @@ SubModules:
     module_name: ts_forecast
     model_name: DLinear
     model_dir: null
-    batch_size: 1   
+    batch_size: 1   

paddlex/constants.py

@@ -0,0 +1,17 @@
+# 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.
+
+from typing import Final
+
+MODEL_FILE_PREFIX: Final[str] = "inference"

paddlex/engine.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,13 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import os
 
+from .model import _ModelBasedConfig
+from .utils.config import get_config, parse_args
+from .utils.errors import raise_unsupported_api_error
+from .utils.flags import INFER_BENCHMARK
 from .utils.lazy_loader import disable_pir_bydefault
 from .utils.result_saver import try_except_decorator
-from .utils.config import parse_args, get_config
-from .utils.errors import raise_unsupported_api_error
-from .model import _ModelBasedConfig
 
 
 class Engine(object):
@@ -47,6 +47,8 @@ class Engine(object):
             return self._model.export()
         elif self._mode == "predict":
             for res in self._model.predict():
+                if INFER_BENCHMARK:
+                    continue
                 res.print()
                 if self._output:
                     res.save_all(save_path=self._output)

paddlex/hpip_links.html

@@ -4,16 +4,28 @@
         <title>PaddleX HPIP Links</title>
     </head>
     <body>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_python-1.0.0.3.0.0rc0-cp38-cp38-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_python-1.0.0.3.0.0rc0-cp39-cp39-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_python-1.0.0.3.0.0rc0-cp310-cp310-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_python-1.0.0.3.0.0rc0-cp311-cp311-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_python-1.0.0.3.0.0rc0-cp312-cp312-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_gpu_python-1.0.0.3.0.0rc0-cp38-cp38-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_gpu_python-1.0.0.3.0.0rc0-cp39-cp39-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_gpu_python-1.0.0.3.0.0rc0-cp310-cp310-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_gpu_python-1.0.0.3.0.0rc0-cp311-cp311-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/ultra_infer_gpu_python-1.0.0.3.0.0rc0-cp312-cp312-linux_x86_64.whl"></a>
-        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/3.0.0rc0/paddlex_hpi-3.0.0rc0-py3-none-any.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_python-1.1.0-cp38-cp38-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_python-1.1.0-cp39-cp39-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_python-1.1.0-cp310-cp310-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_python-1.1.0-cp311-cp311-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_python-1.1.0-cp312-cp312-linux_x86_64.whl"></a>
+
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_gpu_python-1.1.0-cp38-cp38-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_gpu_python-1.1.0-cp39-cp39-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_gpu_python-1.1.0-cp310-cp310-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_gpu_python-1.1.0-cp311-cp311-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_gpu_python-1.1.0-cp312-cp312-linux_x86_64.whl"></a>
+
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp38-cp38-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp39-cp39-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp310-cp310-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp311-cp311-linux_x86_64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp312-cp312-linux_x86_64.whl"></a>
+
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp38-cp38-linux_aarch64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp39-cp39-linux_aarch64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp310-cp310-linux_aarch64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp311-cp311-linux_aarch64.whl"></a>
+        <a href="https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/deploy/hpi/ultra_infer/releases/new_hpi/v1.1.0/ultra_infer_npu_python-1.1.0-cp312-cp312-linux_aarch64.whl"></a>
     </body>
 </html>

paddlex/inference/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -13,7 +13,7 @@
 # limitations under the License.
 
 from ..utils import logging
-
-from .pipelines import create_pipeline, load_pipeline_config
 from .models import create_predictor
+from .pipelines import create_pipeline, load_pipeline_config
+from .utils.hpi import HPIConfig
 from .utils.pp_option import PaddlePredictorOption

paddlex/inference/common/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.

paddlex/inference/common/batch_sampler/__init__.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,9 +12,10 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from .audio_batch_sampler import AudioBatchSampler
 from .base_batch_sampler import BaseBatchSampler
+from .det_3d_batch_sampler import Det3DBatchSampler
+from .doc_vlm_batch_sampler import DocVLMBatchSampler
 from .image_batch_sampler import ImageBatchSampler
-from .video_batch_sampler import VideoBatchSampler
 from .ts_batch_sampler import TSBatchSampler
-from .det_3d_batch_sampler import Det3DBatchSampler
-from .audio_batch_sampler import AudioBatchSampler
+from .video_batch_sampler import VideoBatchSampler

paddlex/inference/common/batch_sampler/audio_batch_sampler.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,10 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import os
-import ast
 from pathlib import Path
-import numpy as np
 
 from ....utils import logging
 from ....utils.cache import CACHE_DIR
@@ -60,9 +57,11 @@ class AudioBatchSampler(BaseBatchSampler):
             if inputs.startswith("http"):
                 inputs = self._download_from_url(inputs)
             yield [inputs]
+        elif isinstance(inputs, list):
+            yield inputs
         else:
-            logging.warning(
-                f"Not supported input data type! Only `str` are supported, but got: {input}."
+            raise TypeError(
+                f"Not supported input data type! Only `str` are supported, but got: {type(inputs)}."
             )
 
     @BaseBatchSampler.batch_size.setter

paddlex/inference/common/batch_sampler/base_batch_sampler.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,14 +12,25 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Union, Tuple, List, Dict, Any, Iterator
-from abc import ABC, abstractmethod
+from abc import abstractmethod
+from typing import Any, Dict, Iterator, List, Tuple
 
-from ....utils.flags import (
-    INFER_BENCHMARK,
-    INFER_BENCHMARK_ITER,
-    INFER_BENCHMARK_DATA_SIZE,
-)
+
+class Batch:
+    def __init__(self):
+        self.instances = []
+        self.input_paths = []
+
+    def append(self, instance, input_path):
+        self.instances.append(instance)
+        self.input_paths.append(input_path)
+
+    def reset(self):
+        self.instances = []
+        self.input_paths = []
+
+    def __len__(self):
+        return len(self.instances)
 
 
 class BaseBatchSampler:
@@ -33,9 +44,6 @@ class BaseBatchSampler:
         """
         super().__init__()
         self._batch_size = batch_size
-        self._benchmark = INFER_BENCHMARK
-        self._benchmark_iter = INFER_BENCHMARK_ITER
-        self._benchmark_data_size = INFER_BENCHMARK_DATA_SIZE
 
     @property
     def batch_size(self) -> int:
@@ -69,11 +77,7 @@ class BaseBatchSampler:
         Yields:
             Iterator[List[Any]]: An iterator yielding the batch data.
         """
-        if input is None and self._benchmark:
-            for _ in range(self._benchmark_iter):
-                yield self._rand_batch(self._benchmark_data_size)
-        else:
-            yield from self.sample(input)
+        yield from self.sample(input)
 
     @abstractmethod
     def sample(self, *args: Tuple[Any], **kwargs: Dict[str, Any]) -> Iterator[list]:

paddlex/inference/common/batch_sampler/det_3d_batch_sampler.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -12,18 +12,15 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Any, Dict, List, Optional, Union
 import os
-import ast
-from pathlib import Path
-import numpy as np
 import pickle
 import tarfile
-
+from pathlib import Path
+from typing import Any, Dict, List, Union
 
 from ....utils import logging
-from ....utils.download import download
 from ....utils.cache import CACHE_DIR
+from ....utils.download import download
 from .base_batch_sampler import BaseBatchSampler
 
 

paddlex/inference/common/batch_sampler/doc_vlm_batch_sampler.py

@@ -0,0 +1,64 @@
+# 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.
+
+
+from ....utils import logging
+from .base_batch_sampler import BaseBatchSampler
+
+
+class DocVLMBatchSampler(BaseBatchSampler):
+    def __init__(self):
+        """Initializes the BaseBatchSampler.
+
+        Args:
+            batch_size (int, optional): The size of each batch. Only support 1.
+        """
+        super().__init__()
+        self.batch_size = 1
+
+    def sample(self, inputs):
+        """Generate list of input file path.
+
+        Args:
+            inputs (str): file path.
+
+        Yields:
+            list: list of file path.
+        """
+        if isinstance(inputs, dict):
+            yield [inputs]
+        elif isinstance(inputs, list) and all(isinstance(i, dict) for i in inputs):
+            yield inputs
+        else:
+            raise TypeError(
+                f"Not supported input data type! Only `dict` are supported, but got: {type(inputs)}."
+            )
+
+    @BaseBatchSampler.batch_size.setter
+    def batch_size(self, batch_size):
+        """Sets the batch size.
+
+        Args:
+            batch_size (int): The batch size to set.
+
+        Raises:
+            Warning: If the batch size is not equal 1.
+        """
+        # only support batch size 1
+        if batch_size != 1:
+            logging.warning(
+                f"doc vlm batch sampler only support batch size 1, but got {batch_size}."
+            )
+        else:
+            self._batch_size = batch_size

paddlex/inference/common/batch_sampler/image_batch_sampler.py

@@ -1,4 +1,4 @@
-# copyright (c) 2024 PaddlePaddle Authors. All Rights Reserve.
+# 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.
@@ -13,36 +13,30 @@
 # limitations under the License.
 
 import os
-import ast
 from pathlib import Path
+
 import numpy as np
 
 from ....utils import logging
-from ....utils.download import download
 from ....utils.cache import CACHE_DIR
+from ....utils.download import download
 from ...utils.io import PDFReader
-from .base_batch_sampler import BaseBatchSampler
+from .base_batch_sampler import BaseBatchSampler, Batch
 
 
-class ImgInstance:
+class ImgBatch(Batch):
     def __init__(self):
-        self.instances = []
-        self.input_paths = []
+        super().__init__()
         self.page_indexes = []
 
     def append(self, instance, input_path, page_index):
-        self.instances.append(instance)
-        self.input_paths.append(input_path)
+        super().append(instance, input_path)
         self.page_indexes.append(page_index)
 
     def reset(self):
-        self.instances = []
-        self.input_paths = []
+        super().reset()
         self.page_indexes = []
 
-    def __len__(self):
-        return len(self.instances)
-
 
 class ImageBatchSampler(BaseBatchSampler):
 
@@ -80,13 +74,13 @@ class ImageBatchSampler(BaseBatchSampler):
         if not isinstance(inputs, list):
             inputs = [inputs]
 
-        batch = ImgInstance()
+        batch = ImgBatch()
         for input in inputs:
             if isinstance(input, np.ndarray):
                 batch.append(input, None, None)
                 if len(batch) == self.batch_size:
                     yield batch
-                    batch.reset()
+                    batch = ImgBatch()
             elif isinstance(input, str) and input.split(".")[-1] in ("PDF", "pdf"):
                 file_path = (
                     self._download_from_url(input)
@@ -97,7 +91,7 @@ class ImageBatchSampler(BaseBatchSampler):
                     batch.append(page_img, file_path, page_idx)
                     if len(batch) == self.batch_size:
                         yield batch
-                        batch.reset()
+                        batch = ImgBatch()
             elif isinstance(input, str):
                 file_path = (
                     self._download_from_url(input)
@@ -109,28 +103,10 @@ class ImageBatchSampler(BaseBatchSampler):
                     batch.append(file_path, file_path, None)
                     if len(batch) == self.batch_size:
                         yield batch
-                        batch.reset()
+                        batch = ImgBatch()
             else:
                 logging.warning(
                     f"Not supported input data type! Only `numpy.ndarray` and `str` are supported! So has been ignored: {input}."
                 )
         if len(batch) > 0:
             yield batch
-
-    def _rand_batch(self, data_size):
-        def parse_size(s):
-            res = ast.literal_eval(s)
-            if isinstance(res, int):
-                return (res, res)
-            else:
-                assert isinstance(res, (tuple, list))
-                assert len(res) == 2
-                assert all(isinstance(item, int) for item in res)
-                return res
-
-        size = parse_size(data_size)
-        rand_batch = [
-            np.random.randint(0, 256, (*size, 3), dtype=np.uint8)
-            for _ in range(self.batch_size)
-        ]
-        return rand_batch