When i import the model, It gives me this Error ValueError: Layer count mismatch when loading weights from file. Model expected 4 layers, found 0 saved layers

[AaribAzfar]

I tried using doctr, for OCR and when i try to load model, it gives this error

from doctr.models import ocr_predictor
model = ocr_predictor(
    det_arch="db_resnet50", reco_arch="crnn_vgg16_bn", pretrained=True
)

This Error:
ValueError: Layer count mismatch when loading weights from file. Model expected 4 layers, found 0 saved layers

[AaribAzfar]

This is the Whole Error

ValueError                                Traceback (most recent call last)
Cell In[2], [line 3](vscode-notebook-cell:?execution_count=2&line=3)
      [1](vscode-notebook-cell:?execution_count=2&line=1) # Create an OCR predictor
----> [3](vscode-notebook-cell:?execution_count=2&line=3) model = ocr_predictor(
      [4](vscode-notebook-cell:?execution_count=2&line=4)     det_arch="db_resnet50", reco_arch="crnn_vgg16_bn", pretrained=True
      [5](vscode-notebook-cell:?execution_count=2&line=5) )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\zoo.py:114, in ocr_predictor(det_arch, reco_arch, pretrained, pretrained_backbone, assume_straight_pages, preserve_aspect_ratio, symmetric_pad, export_as_straight_boxes, detect_orientation, straighten_pages, detect_language, **kwargs)
     [63](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:63) def ocr_predictor(
     [64](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:64)     det_arch: Any = "fast_base",
     [65](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:65)     reco_arch: Any = "crnn_vgg16_bn",
   (...)
     [75](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:75)     **kwargs: Any,
     [76](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:76) ) -> OCRPredictor:
     [77](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:77)     """End-to-end OCR architecture using one model for localization, and another for text recognition.
     [78](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:78) 
     [79](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:79)     >>> import numpy as np
   (...)
    [112](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:112)         OCR predictor
    [113](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:113)     """
--> [114](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:114)     return _predictor(
    [115](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:115)         det_arch,
    [116](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:116)         reco_arch,
    [117](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:117)         pretrained,
    [118](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:118)         pretrained_backbone=pretrained_backbone,
    [119](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:119)         assume_straight_pages=assume_straight_pages,
    [120](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:120)         preserve_aspect_ratio=preserve_aspect_ratio,
    [121](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:121)         symmetric_pad=symmetric_pad,
    [122](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:122)         export_as_straight_boxes=export_as_straight_boxes,
    [123](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:123)         detect_orientation=detect_orientation,
    [124](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:124)         straighten_pages=straighten_pages,
    [125](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:125)         detect_language=detect_language,
    [126](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:126)         **kwargs,
    [127](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:127)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\zoo.py:32, in _predictor(det_arch, reco_arch, pretrained, pretrained_backbone, assume_straight_pages, preserve_aspect_ratio, symmetric_pad, det_bs, reco_bs, detect_orientation, straighten_pages, detect_language, **kwargs)
     [16](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:16) def _predictor(
     [17](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:17)     det_arch: Any,
     [18](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:18)     reco_arch: Any,
   (...)
     [30](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:30) ) -> OCRPredictor:
     [31](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:31)     # Detection
---> [32](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:32)     det_predictor = detection_predictor(
     [33](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:33)         det_arch,
     [34](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:34)         pretrained=pretrained,
     [35](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:35)         pretrained_backbone=pretrained_backbone,
     [36](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:36)         batch_size=det_bs,
     [37](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:37)         assume_straight_pages=assume_straight_pages,
     [38](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:38)         preserve_aspect_ratio=preserve_aspect_ratio,
     [39](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:39)         symmetric_pad=symmetric_pad,
     [40](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:40)     )
     [42](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:42)     # Recognition
     [43](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:43)     reco_predictor = recognition_predictor(
     [44](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:44)         reco_arch,
     [45](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:45)         pretrained=pretrained,
     [46](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:46)         pretrained_backbone=pretrained_backbone,
     [47](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:47)         batch_size=reco_bs,
     [48](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:48)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\zoo.py:103, in detection_predictor(arch, pretrained, assume_straight_pages, **kwargs)
     [78](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:78) def detection_predictor(
     [79](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:79)     arch: Any = "fast_base",
     [80](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:80)     pretrained: bool = False,
     [81](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:81)     assume_straight_pages: bool = True,
     [82](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:82)     **kwargs: Any,
     [83](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:83) ) -> DetectionPredictor:
     [84](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:84)     """Text detection architecture.
     [85](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:85) 
     [86](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:86)     >>> import numpy as np
   (...)
    [101](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:101)         Detection predictor
    [102](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:102)     """
--> [103](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:103)     return _predictor(arch, pretrained, assume_straight_pages, **kwargs)

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\zoo.py:50, in _predictor(arch, pretrained, assume_straight_pages, **kwargs)
     [47](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:47) if arch not in ARCHS:
     [48](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:48)     raise ValueError(f"unknown architecture '{arch}'")
---> [50](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:50) _model = detection.__dict__[arch](
     [51](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:51)     pretrained=pretrained,
     [52](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:52)     pretrained_backbone=kwargs.get("pretrained_backbone", True),
     [53](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:53)     assume_straight_pages=assume_straight_pages,
     [54](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:54) )
     [55](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:55) # Reparameterize FAST models by default to lower inference latency and memory usage
     [56](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:56) if isinstance(_model, detection.FAST):

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\differentiable_binarization\tensorflow.py:390, in db_resnet50(pretrained, **kwargs)
    [371](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:371) def db_resnet50(pretrained: bool = False, **kwargs: Any) -> DBNet:
    [372](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:372)     """DBNet as described in `"Real-time Scene Text Detection with Differentiable Binarization"
    [373](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:373)     <[https://arxiv.org/pdf/1911.08947.pdf>`_](https://arxiv.org/pdf/1911.08947.pdf%3E%60_), using a ResNet-50 backbone.
    [374](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:374) 
   (...)
    [388](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:388)         text detection architecture
    [389](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:389)     """
--> [390](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:390)     return _db_resnet(
    [391](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:391)         "db_resnet50",
    [392](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:392)         pretrained,
    [393](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:393)         ResNet50,
    [394](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:394)         ["conv2_block3_out", "conv3_block4_out", "conv4_block6_out", "conv5_block3_out"],
    [395](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:395)         **kwargs,
    [396](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:396)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\differentiable_binarization\tensorflow.py:318, in _db_resnet(arch, pretrained, backbone_fn, fpn_layers, pretrained_backbone, input_shape, **kwargs)
    [315](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:315) # Load pretrained parameters
    [316](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:316) if pretrained:
    [317](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:317)     # The given class_names differs from the pretrained model => skip the mismatching layers for fine tuning
--> [318](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:318)     load_pretrained_params(
    [319](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:319)         model,
    [320](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:320)         _cfg["url"],
    [321](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:321)         skip_mismatch=kwargs["class_names"] != default_cfgs[arch].get("class_names", [CLASS_NAME]),
    [322](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:322)     )
    [324](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:324) return model

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\utils\tensorflow.py:73, in load_pretrained_params(model, url, hash_prefix, skip_mismatch, **kwargs)
     [71](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:71) archive_path = download_from_url(url, hash_prefix=hash_prefix, cache_subdir="models", **kwargs)
     [72](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:72) # Load weights
---> [73](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:73) model.load_weights(archive_path, skip_mismatch=skip_mismatch)

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\keras\utils\traceback_utils.py:70, in filter_traceback.<locals>.error_handler(*args, **kwargs)
     [67](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:67)     filtered_tb = _process_traceback_frames(e.__traceback__)
     [68](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:68)     # To get the full stack trace, call:
     [69](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:69)     # `tf.debugging.disable_traceback_filtering()`
---> [70](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:70)     raise e.with_traceback(filtered_tb) from None
     [71](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:71) finally:
     [72](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/utils/traceback_utils.py:72)     del filtered_tb

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\keras\saving\hdf5_format.py:817, in load_weights_from_hdf5_group(f, model)
    [815](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:815) layer_names = filtered_layer_names
    [816](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:816) if len(layer_names) != len(filtered_layers):
--> [817](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:817)     raise ValueError(
    [818](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:818)         f"Layer count mismatch when loading weights from file. "
    [819](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:819)         f"Model expected {len(filtered_layers)} layers, found "
    [820](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:820)         f"{len(layer_names)} saved layers."
    [821](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:821)     )
    [823](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:823) # We batch weight value assignments in a single backend call
    [824](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:824) # which provides a speedup in TensorFlow.
    [825](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/keras/saving/hdf5_format.py:825) weight_value_tuples = []

ValueError: Layer count mismatch when loading weights from file. Model expected 4 layers, found 0 saved layers.

[felixdittrich92]

Hi @AaribAzfar 👋,

Which backend do you use TensorFlow or PyTorch and which docTR version ? :)

Could you please also provide the informations you get by executing: https://github.com/mindee/doctr/blob/main/scripts/collect_env.py ?

[AaribAzfar]

Thanks for fast replies.
I cannot change my pytorch, since i need cuda on that, to run other model, and also want to use OCR with this

[AaribAzfar]

I have tried other ocr, however they are not accurate for my use case, tried finetuning them, but still not accurate enough.
i am checking if doctr gives better results and maybe finetune or train it if possible to my exact use case

[felixdittrich92]

Ok that's strange ..

which tf-keras version have you installed ?

Could you try to put the following on top instead ?

import os
from doctr.file_utils import ensure_keras_v2

ensure_keras_v2()

os.environ["USE_TF"] = "1"

from doctr.models import ocr_predictor

model = ocr_predictor(
    det_arch="db_resnet50", reco_arch="crnn_vgg16_bn", pretrained=True
)

An alternative would be to use https://github.com/felixdittrich92/OnnxTR (at the end docTR but without the tensorflow or pytorch requirement) - If you have conflicts with other code

[AaribAzfar]

Same error as before, OnnxTr is for documents/images. However i have a video, from which i have to detect text.

[felixT2K]

which tf-keras version have you installed ?

If you use docTR for videos you can do the same with OnnxTR (user interface is also mostly the same) :)

[AaribAzfar]

PS C:\Users\areeb\Documents\Data_Preprocessing_and_Training> pip show keras                                                                                                                                     
Name: keras                                                                                                                                                                                                     
Version: 3.6.0                                                                                                                                                                                                  
Summary: Multi-backend Keras.                                                                                                                                                                                   
Home-page: https://github.com/keras-team/keras                                                                                                                                                                  
Author: Keras team                                                                                                                                                                                              
Author-email: keras-users@googlegroups.com                                                                                                                                                                      
License: Apache License 2.0                                                                                                                                                                                     
Location: c:\users\areeb\appdata\local\programs\python\python310\lib\site-packages                                                                                                                              
Requires: absl-py, h5py, ml-dtypes, namex, numpy, optree, packaging, rich                                                                                                                                       
Required-by: tensorflow_intel                     

OK i will check for OnnxTR

[AaribAzfar]

PS C:\Users\areeb\Documents\Data_Preprocessing_and_Training> pip show tf-keras                                                                                                                                  
Name: tf_keras                                                                                                                                                                                                  
Version: 2.18.0
Summary: Deep learning for humans.
Home-page: https://keras.io/
Author: Keras team
Author-email: keras-users@googlegroups.com
License: Apache 2.0
Location: c:\users\areeb\appdata\roaming\python\python310\site-packages
Requires: tensorflow
Required-by:

Same Error ValueError: Layer 'conv2d_9' expected 1 variables, but received 0 variables during loading. Expected: ['conv2d_9/kernel:0']

[AaribAzfar]

Whole Error for reference:

ValueError                                Traceback (most recent call last)
Cell In[1], [line 10](vscode-notebook-cell:?execution_count=1&line=10)
      [6](vscode-notebook-cell:?execution_count=1&line=6) os.environ["USE_TF"] = "1"
      [8](vscode-notebook-cell:?execution_count=1&line=8) from doctr.models import ocr_predictor
---> [10](vscode-notebook-cell:?execution_count=1&line=10) model = ocr_predictor(
     [11](vscode-notebook-cell:?execution_count=1&line=11)     det_arch="db_resnet50", reco_arch="crnn_vgg16_bn", pretrained=True
     [12](vscode-notebook-cell:?execution_count=1&line=12) )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\zoo.py:114, in ocr_predictor(det_arch, reco_arch, pretrained, pretrained_backbone, assume_straight_pages, preserve_aspect_ratio, symmetric_pad, export_as_straight_boxes, detect_orientation, straighten_pages, detect_language, **kwargs)
     [63](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:63) def ocr_predictor(
     [64](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:64)     det_arch: Any = "fast_base",
     [65](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:65)     reco_arch: Any = "crnn_vgg16_bn",
   (...)
     [75](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:75)     **kwargs: Any,
     [76](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:76) ) -> OCRPredictor:
     [77](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:77)     """End-to-end OCR architecture using one model for localization, and another for text recognition.
     [78](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:78) 
     [79](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:79)     >>> import numpy as np
   (...)
    [112](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:112)         OCR predictor
    [113](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:113)     """
--> [114](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:114)     return _predictor(
    [115](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:115)         det_arch,
    [116](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:116)         reco_arch,
    [117](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:117)         pretrained,
    [118](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:118)         pretrained_backbone=pretrained_backbone,
    [119](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:119)         assume_straight_pages=assume_straight_pages,
    [120](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:120)         preserve_aspect_ratio=preserve_aspect_ratio,
    [121](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:121)         symmetric_pad=symmetric_pad,
    [122](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:122)         export_as_straight_boxes=export_as_straight_boxes,
    [123](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:123)         detect_orientation=detect_orientation,
    [124](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:124)         straighten_pages=straighten_pages,
    [125](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:125)         detect_language=detect_language,
    [126](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:126)         **kwargs,
    [127](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:127)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\zoo.py:32, in _predictor(det_arch, reco_arch, pretrained, pretrained_backbone, assume_straight_pages, preserve_aspect_ratio, symmetric_pad, det_bs, reco_bs, detect_orientation, straighten_pages, detect_language, **kwargs)
     [16](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:16) def _predictor(
     [17](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:17)     det_arch: Any,
     [18](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:18)     reco_arch: Any,
   (...)
     [30](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:30) ) -> OCRPredictor:
     [31](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:31)     # Detection
---> [32](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:32)     det_predictor = detection_predictor(
     [33](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:33)         det_arch,
     [34](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:34)         pretrained=pretrained,
     [35](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:35)         pretrained_backbone=pretrained_backbone,
     [36](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:36)         batch_size=det_bs,
     [37](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:37)         assume_straight_pages=assume_straight_pages,
     [38](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:38)         preserve_aspect_ratio=preserve_aspect_ratio,
     [39](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:39)         symmetric_pad=symmetric_pad,
     [40](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:40)     )
     [42](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:42)     # Recognition
     [43](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:43)     reco_predictor = recognition_predictor(
     [44](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:44)         reco_arch,
     [45](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:45)         pretrained=pretrained,
     [46](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:46)         pretrained_backbone=pretrained_backbone,
     [47](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:47)         batch_size=reco_bs,
     [48](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/zoo.py:48)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\zoo.py:103, in detection_predictor(arch, pretrained, assume_straight_pages, **kwargs)
     [78](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:78) def detection_predictor(
     [79](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:79)     arch: Any = "fast_base",
     [80](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:80)     pretrained: bool = False,
     [81](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:81)     assume_straight_pages: bool = True,
     [82](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:82)     **kwargs: Any,
     [83](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:83) ) -> DetectionPredictor:
     [84](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:84)     """Text detection architecture.
     [85](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:85) 
     [86](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:86)     >>> import numpy as np
   (...)
    [101](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:101)         Detection predictor
    [102](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:102)     """
--> [103](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:103)     return _predictor(arch, pretrained, assume_straight_pages, **kwargs)

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\zoo.py:50, in _predictor(arch, pretrained, assume_straight_pages, **kwargs)
     [47](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:47) if arch not in ARCHS:
     [48](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:48)     raise ValueError(f"unknown architecture '{arch}'")
---> [50](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:50) _model = detection.__dict__[arch](
     [51](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:51)     pretrained=pretrained,
     [52](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:52)     pretrained_backbone=kwargs.get("pretrained_backbone", True),
     [53](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:53)     assume_straight_pages=assume_straight_pages,
     [54](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:54) )
     [55](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:55) # Reparameterize FAST models by default to lower inference latency and memory usage
     [56](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/zoo.py:56) if isinstance(_model, detection.FAST):

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\differentiable_binarization\tensorflow.py:390, in db_resnet50(pretrained, **kwargs)
    [371](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:371) def db_resnet50(pretrained: bool = False, **kwargs: Any) -> DBNet:
    [372](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:372)     """DBNet as described in `"Real-time Scene Text Detection with Differentiable Binarization"
    [373](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:373)     <[https://arxiv.org/pdf/1911.08947.pdf>`_](https://arxiv.org/pdf/1911.08947.pdf%3E%60_), using a ResNet-50 backbone.
    [374](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:374) 
   (...)
    [388](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:388)         text detection architecture
    [389](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:389)     """
--> [390](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:390)     return _db_resnet(
    [391](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:391)         "db_resnet50",
    [392](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:392)         pretrained,
    [393](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:393)         ResNet50,
    [394](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:394)         ["conv2_block3_out", "conv3_block4_out", "conv4_block6_out", "conv5_block3_out"],
    [395](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:395)         **kwargs,
    [396](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:396)     )

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\detection\differentiable_binarization\tensorflow.py:318, in _db_resnet(arch, pretrained, backbone_fn, fpn_layers, pretrained_backbone, input_shape, **kwargs)
    [315](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:315) # Load pretrained parameters
    [316](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:316) if pretrained:
    [317](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:317)     # The given class_names differs from the pretrained model => skip the mismatching layers for fine tuning
--> [318](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:318)     load_pretrained_params(
    [319](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:319)         model,
    [320](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:320)         _cfg["url"],
    [321](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:321)         skip_mismatch=kwargs["class_names"] != default_cfgs[arch].get("class_names", [CLASS_NAME]),
    [322](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:322)     )
    [324](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/detection/differentiable_binarization/tensorflow.py:324) return model

File c:\Users\areeb\AppData\Local\Programs\Python\Python310\lib\site-packages\doctr\models\utils\tensorflow.py:73, in load_pretrained_params(model, url, hash_prefix, skip_mismatch, **kwargs)
     [71](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:71) archive_path = download_from_url(url, hash_prefix=hash_prefix, cache_subdir="models", **kwargs)
     [72](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:72) # Load weights
---> [73](file:///C:/Users/areeb/AppData/Local/Programs/Python/Python310/lib/site-packages/doctr/models/utils/tensorflow.py:73) model.load_weights(archive_path, skip_mismatch=skip_mismatch)

File ~\AppData\Roaming\Python\Python310\site-packages\tf_keras\src\utils\traceback_utils.py:70, in filter_traceback.<locals>.error_handler(*args, **kwargs)
     [67](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:67)     filtered_tb = _process_traceback_frames(e.__traceback__)
     [68](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:68)     # To get the full stack trace, call:
     [69](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:69)     # `tf.debugging.disable_traceback_filtering()`
---> [70](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:70)     raise e.with_traceback(filtered_tb) from None
     [71](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:71) finally:
     [72](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/utils/traceback_utils.py:72)     del filtered_tb

File ~\AppData\Roaming\Python\Python310\site-packages\tf_keras\src\engine\base_layer.py:3518, in Layer.load_own_variables(self, store)
   [3516](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3516) all_vars = self._trainable_weights + self._non_trainable_weights
   [3517](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3517) if len(store.keys()) != len(all_vars):
-> [3518](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3518)     raise ValueError(
   [3519](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3519)         f"Layer '{self.name}' expected {len(all_vars)} variables, "
   [3520](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3520)         "but received "
   [3521](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3521)         f"{len(store.keys())} variables during loading. "
   [3522](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3522)         f"Expected: {[v.name for v in all_vars]}"
   [3523](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3523)     )
   [3524](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3524) for i, v in enumerate(all_vars):
   [3525](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3525)     # TODO(rchao): check shapes and raise errors.
   [3526](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/areeb/Documents/Data_Preprocessing_and_Training/~/AppData/Roaming/Python/Python310/site-packages/tf_keras/src/engine/base_layer.py:3526)     v.assign(store[f"{i}"])

ValueError: Layer 'conv2d_9' expected 1 variables, but received 0 variables during loading. Expected: ['conv2d_9/kernel:0']

[AaribAzfar]

To fix this simply install pip3 install tf-keras and afterwards run again :

import os
from doctr.file_utils import ensure_keras_v2

ensure_keras_v2()

os.environ["USE_TF"] = "1"

from doctr.models import ocr_predictor

model = ocr_predictor(
    det_arch="db_resnet50", reco_arch="crnn_vgg16_bn", pretrained=True
)

I am running this script

[felixT2K]

Does this happen also with another detection model for example "db_mobilenet_v3_large" ?

[AaribAzfar]

Yes same :(