--- library_name: pytorch license: other tags: - android pipeline_tag: image-to-text --- ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/trocr/web-assets/model_demo.png) # TrOCR: Optimized for Mobile Deployment ## Transformer based model for state-of-the-art optical character recognition (OCR) on both printed and handwritten text End-to-end text recognition approach with pre-trained image transformer and text transformer models for both image understanding and wordpiece-level text generation. This model is an implementation of TrOCR found [here](https://huggingface.co/microsoft/trocr-small-stage1). This repository provides scripts to run TrOCR on Qualcomm® devices. More details on model performance across various devices, can be found [here](https://aihub.qualcomm.com/models/trocr). ### Model Details - **Model Type:** Model_use_case.image_to_text - **Model Stats:** - Model checkpoint: trocr-small-stage1 - Input resolution: 320x320 - Number of parameters (TrOCRDecoder): 38.3M - Model size (TrOCRDecoder) (float): 146 MB - Number of parameters (TrOCREncoder): 23.0M - Model size (TrOCREncoder) (float): 87.8 MB | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model |---|---|---|---|---|---|---|---|---| | TrOCRDecoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 4.312 ms | 0 - 72 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.076 ms | 5 - 73 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 2.444 ms | 0 - 134 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 2.807 ms | 5 - 131 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 2.096 ms | 0 - 214 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 2.018 ms | 1 - 26 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 2.463 ms | 0 - 176 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCRDecoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 2.954 ms | 0 - 72 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 2.829 ms | 5 - 72 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 4.312 ms | 0 - 72 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 4.076 ms | 5 - 73 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 2.085 ms | 0 - 224 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 2.006 ms | 2 - 25 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 3.082 ms | 0 - 64 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.872 ms | 0 - 58 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 2.169 ms | 0 - 381 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 2.04 ms | 1 - 21 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 2.954 ms | 0 - 72 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.829 ms | 5 - 72 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.504 ms | 0 - 142 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.587 ms | 0 - 146 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.768 ms | 0 - 146 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCRDecoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 1.215 ms | 0 - 77 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.156 ms | 0 - 153 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 1.476 ms | 0 - 154 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCRDecoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 1.158 ms | 0 - 72 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCRDecoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 1.129 ms | 1 - 138 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 1.357 ms | 3 - 144 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.194 ms | 654 - 654 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.274 ms | 67 - 67 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 41.818 ms | 7 - 114 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 46.598 ms | 0 - 138 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 22.611 ms | 6 - 123 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 28.219 ms | 2 - 135 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 13.232 ms | 7 - 20 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 17.438 ms | 2 - 24 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 17.86 ms | 0 - 131 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCREncoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 16.001 ms | 7 - 113 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 19.767 ms | 2 - 141 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 41.818 ms | 7 - 114 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 46.598 ms | 0 - 138 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 13.31 ms | 7 - 29 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 17.519 ms | 2 - 24 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 24.233 ms | 7 - 122 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 25.777 ms | 2 - 130 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 13.18 ms | 7 - 20 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 17.353 ms | 2 - 22 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 16.001 ms | 7 - 113 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 19.767 ms | 2 - 141 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 9.413 ms | 5 - 119 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 11.726 ms | 2 - 147 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 11.608 ms | 15 - 185 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCREncoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 6.357 ms | 6 - 117 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 8.382 ms | 2 - 151 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 10.011 ms | 15 - 165 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCREncoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 5.471 ms | 0 - 110 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) | | TrOCREncoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 6.952 ms | 2 - 128 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 7.249 ms | 15 - 144 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 18.104 ms | 214 - 214 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) | | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 18.126 ms | 49 - 49 MB | NPU | [TrOCR.onnx.zip](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx.zip) | ## Installation Install the package via pip: ```bash pip install "qai-hub-models[trocr]" ``` ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`. With this API token, you can configure your client to run models on the cloud hosted devices. ```bash qai-hub configure --api_token API_TOKEN ``` Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information. ## Demo off target The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input. ```bash python -m qai_hub_models.models.trocr.demo ``` The above demo runs a reference implementation of pre-processing, model inference, and post processing. **NOTE**: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above). ``` %run -m qai_hub_models.models.trocr.demo ``` ### Run model on a cloud-hosted device In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following: * Performance check on-device on a cloud-hosted device * Downloads compiled assets that can be deployed on-device for Android. * Accuracy check between PyTorch and on-device outputs. ```bash python -m qai_hub_models.models.trocr.export ``` ## How does this work? This [export script](https://aihub.qualcomm.com/models/trocr/qai_hub_models/models/TrOCR/export.py) leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model on-device. Lets go through each step below in detail: Step 1: **Compile model for on-device deployment** To compile a PyTorch model for on-device deployment, we first trace the model in memory using the `jit.trace` and then call the `submit_compile_job` API. ```python import torch import qai_hub as hub from qai_hub_models.models.trocr import Model # Load the model torch_model = Model.from_pretrained() # Device device = hub.Device("Samsung Galaxy S25") # Trace model input_shape = torch_model.get_input_spec() sample_inputs = torch_model.sample_inputs() pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()]) # Compile model on a specific device compile_job = hub.submit_compile_job( model=pt_model, device=device, input_specs=torch_model.get_input_spec(), ) # Get target model to run on-device target_model = compile_job.get_target_model() ``` Step 2: **Performance profiling on cloud-hosted device** After compiling models from step 1. Models can be profiled model on-device using the `target_model`. Note that this scripts runs the model on a device automatically provisioned in the cloud. Once the job is submitted, you can navigate to a provided job URL to view a variety of on-device performance metrics. ```python profile_job = hub.submit_profile_job( model=target_model, device=device, ) ``` Step 3: **Verify on-device accuracy** To verify the accuracy of the model on-device, you can run on-device inference on sample input data on the same cloud hosted device. ```python input_data = torch_model.sample_inputs() inference_job = hub.submit_inference_job( model=target_model, device=device, inputs=input_data, ) on_device_output = inference_job.download_output_data() ``` With the output of the model, you can compute like PSNR, relative errors or spot check the output with expected output. **Note**: This on-device profiling and inference requires access to Qualcomm® AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup). ## Deploying compiled model to Android The models can be deployed using multiple runtimes: - TensorFlow Lite (`.tflite` export): [This tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a guide to deploy the .tflite model in an Android application. - QNN (`.so` export ): This [sample app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html) provides instructions on how to use the `.so` shared library in an Android application. ## View on Qualcomm® AI Hub Get more details on TrOCR's performance across various devices [here](https://aihub.qualcomm.com/models/trocr). Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/) ## License * The license for the original implementation of TrOCR can be found [here](https://github.com/microsoft/unilm/blob/master/LICENSE). * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf) ## References * [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) * [Source Model Implementation](https://huggingface.co/microsoft/trocr-small-stage1) ## Community * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).