My pytorch conversion to mobile is not working...!

I have OCR NN really…huge and
cuz I want to use pytorch on android…
I tried to convert pytorch model to pytorch mobile…
(But… If I can just uses pytorch library in JAVA??)
Here is what I tried…

recognizer2 = recognizer
recognizer2.eval()
example = img_cv_grey, horizontal_list, free_list,\
                            decoder, beamWidth, batch_size,\
                            workers, allowlist, blocklist, detail,\
                            paragraph, contrast_ths, adjust_contrast,\
                            filter_ths, False
    torch.jit.trace(recognizer2, example)

The example is sample input(altough I don know why it is needed… Can’t we just convert to the mobile without sample input? it is very anoying…)

the error message is…


RuntimeError Traceback (most recent call last)
in ()
----> 1 result = convert_torch_mobile(“drive/My Drive/colab/easy_ocr_test/Capture1.PNG”)

2 frames
in convert_torch_mobile(image, decoder, beamWidth, batch_size, workers, allowlist, blocklist, detail, paragraph, min_size, contrast_ths, adjust_contrast, filter_ths, text_threshold, low_text, link_threshold, canvas_size, mag_ratio, slope_ths, ycenter_ths, height_ths, width_ths, add_margin)
15 detector2.eval()
16 example = img_cv_grey, horizontal_list, free_list, decoder, beamWidth, batch_size, workers, allowlist, blocklist, detail, paragraph, contrast_ths, adjust_contrast, filter_ths, False
—> 17 torch.jit.trace(recognizer2, example)
18
19

/usr/local/lib/python3.6/dist-packages/torch/jit/init.py in trace(func, example_inputs, optimize, check_trace, check_inputs, check_tolerance, strict, _force_outplace, _module_class, _compilation_unit)
953 return trace_module(func, {‘forward’: example_inputs}, None,
954 check_trace, wrap_check_inputs(check_inputs),
–> 955 check_tolerance, strict, _force_outplace, _module_class)
956
957 if (hasattr(func, ‘self’) and isinstance(func.self, torch.nn.Module) and

/usr/local/lib/python3.6/dist-packages/torch/jit/init.py in trace_module(mod, inputs, optimize, check_trace, check_inputs, check_tolerance, strict, _force_outplace, _module_class, _compilation_unit)
1107 func = mod if method_name == “forward” else getattr(mod, method_name)
1108 example_inputs = make_tuple(example_inputs)
-> 1109 module._c._create_method_from_trace(method_name, func, example_inputs, var_lookup_fn, strict, _force_outplace)
1110 check_trace_method = module._c._get_method(method_name)
1111

RuntimeError: Tracer cannot infer type of (array([[255, 255, 255, …, 255, 255, 255],
[255, 255, 255, …, 255, 255, 255],
[255, 255, 255, …, 255, 255, 255],
…,
[255, 255, 255, …, 255, 255, 255],
[255, 255, 255, …, 255, 255, 255],
[255, 255, 255, …, 255, 255, 255]], dtype=uint8), [[59, 89, 15, 31], [97, 295, 11, 31], [58, 326, 32, 58], [332, 746, 34, 58], [61, 169, 59, 79], [197, 295, 59, 79], [318, 460, 56, 80], [481, 559, 59, 79], [581, 663, 59, 79], [99, 131, 83, 99], [145, 210, 77, 103], [231, 445, 83, 99], [58, 504, 128, 154], [11, 45, 143, 159], [58, 774, 152, 176], [58, 172, 174, 198], [197, 295, 177, 197], [318, 404, 174, 198], [411, 471, 177, 197], [494, 602, 174, 198], [625, 711, 177, 197], [65, 181, 199, 219], [199, 275, 199, 219], [297, 489, 203, 219], [58, 352, 244, 273], [59, 551, 273, 293], [59, 169, 295, 315], [197, 279, 295, 315], [305, 445, 295, 315], [469, 575, 295, 315], [599, 683, 295, 315], [67, 181, 317, 337], [197, 305, 317, 337], [327, 569, 321, 337], [60, 288, 366, 390], [58, 724, 390, 414], [59, 169, 413, 433], [197, 281, 413, 433], [305, 437, 413, 433], [461, 565, 413, 433], [591, 675, 413, 433], [96, 182, 434, 458], [199, 307, 437, 457], [327, 523, 439, 457], [58, 242, 484, 508], [58, 472, 508, 532], [58, 172, 530, 554], [197, 273, 533, 553], [297, 437, 533, 553], [461, 535, 533, 551], [561, 645, 533, 553], [97, 181, 555, 575], [199, 305, 555, 575], [362, 440, 552, 576], [456, 506, 552, 576], [532, 795, 555, 575], [635, 659, 613, 631], [696, 734, 610, 634], [335, 499, 605, 641], [539, 561, 615, 631]], [], ‘greedy’, 5, 1, 0, None, None, 1, False, 0.1, 0.5, 0.003, False)
:Only tensors and (possibly nested) tuples of tensors, lists, or dictsare supported as inputs or outputs of traced functions, but instead got value of type ndarray.

And my model is

DataParallel(
(module): Model(
(FeatureExtraction): ResNet_FeatureExtractor(
(ConvNet): ResNet(
(conv0_1): Conv2d(1, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn0_1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv0_2): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn0_2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): BasicBlock(
(conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(64, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
)
(conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(maxpool2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(layer2): Sequential(
(0): BasicBlock(
(conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(128, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(maxpool3): MaxPool2d(kernel_size=2, stride=(2, 1), padding=(0, 1), dilation=1, ceil_mode=False)
(layer3): Sequential(
(0): BasicBlock(
(conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(3): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(4): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(conv3): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(layer4): Sequential(
(0): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(1): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
(2): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
)
)
(conv4_1): Conv2d(512, 512, kernel_size=(2, 2), stride=(2, 1), padding=(0, 1), bias=False)
(bn4_1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(conv4_2): Conv2d(512, 512, kernel_size=(2, 2), stride=(1, 1), bias=False)
(bn4_2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(AdaptiveAvgPool): AdaptiveAvgPool2d(output_size=(None, 1))
(SequenceModeling): Sequential(
(0): BidirectionalLSTM(
(rnn): LSTM(512, 512, batch_first=True, bidirectional=True)
(linear): Linear(in_features=1024, out_features=512, bias=True)
)
(1): BidirectionalLSTM(
(rnn): LSTM(512, 512, batch_first=True, bidirectional=True)
(linear): Linear(in_features=1024, out_features=512, bias=True)
)
)
(Prediction): Linear(in_features=512, out_features=1568, bias=True)
)
)

Why it happens? What I want is just use pytorch in JAVA… any help please? Thanks…

My simpled error output is this

torch.jit.trace(recognizer2, example)

RuntimeError: Type ‘Tuple[int, int, float, float, float, int, float, float, float, float, float, float, bool]’ cannot be traced. Only Tensors and (possibly nested) Lists, Dicts, and Tuples of Tensors can be traced