Hugging Face's logo

qualcomm
/
TrOCR

Image-to-Text
PyTorch
LiteRT
android
Model card Files
xet
Community
3

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.

This repository provides scripts to run TrOCR on Qualcomm® devices. More details on model performance across various devices, can be found here.

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 4.374 ms 0 - 75 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 4.125 ms 5 - 74 MB NPU TrOCR.dlc
TrOCRDecoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 2.424 ms 0 - 132 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 2.846 ms 5 - 134 MB NPU TrOCR.dlc
TrOCRDecoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 1.996 ms 0 - 382 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 1.967 ms 1 - 20 MB NPU TrOCR.dlc
TrOCRDecoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 2.616 ms 0 - 174 MB NPU TrOCR.onnx.zip
TrOCRDecoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 2.828 ms 0 - 74 MB NPU TrOCR.tflite
TrOCRDecoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 8.658 ms 5 - 74 MB NPU TrOCR.dlc
TrOCRDecoder float SA7255P ADP Qualcomm® SA7255P TFLITE 4.374 ms 0 - 75 MB NPU TrOCR.tflite
TrOCRDecoder float SA7255P ADP Qualcomm® SA7255P QNN_DLC 4.125 ms 5 - 74 MB NPU TrOCR.dlc
TrOCRDecoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 2.047 ms 0 - 223 MB NPU TrOCR.tflite
TrOCRDecoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 1.983 ms 2 - 23 MB NPU TrOCR.dlc
TrOCRDecoder float SA8295P ADP Qualcomm® SA8295P TFLITE 2.885 ms 0 - 65 MB NPU TrOCR.tflite
TrOCRDecoder float SA8295P ADP Qualcomm® SA8295P QNN_DLC 2.815 ms 0 - 59 MB NPU TrOCR.dlc
TrOCRDecoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 2.001 ms 0 - 214 MB NPU TrOCR.tflite
TrOCRDecoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 2.005 ms 2 - 21 MB NPU TrOCR.dlc
TrOCRDecoder float SA8775P ADP Qualcomm® SA8775P TFLITE 2.828 ms 0 - 74 MB NPU TrOCR.tflite
TrOCRDecoder float SA8775P ADP Qualcomm® SA8775P QNN_DLC 8.658 ms 5 - 74 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 1.434 ms 0 - 145 MB NPU TrOCR.tflite
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 1.419 ms 0 - 143 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 1.734 ms 0 - 149 MB NPU TrOCR.onnx.zip
TrOCRDecoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 1.286 ms 0 - 79 MB NPU TrOCR.tflite
TrOCRDecoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 1.245 ms 0 - 155 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 1.853 ms 0 - 154 MB NPU TrOCR.onnx.zip
TrOCRDecoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile TFLITE 1.16 ms 0 - 73 MB NPU TrOCR.tflite
TrOCRDecoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile QNN_DLC 1.106 ms 3 - 142 MB NPU TrOCR.dlc
TrOCRDecoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile ONNX 1.4 ms 3 - 143 MB NPU TrOCR.onnx.zip
TrOCRDecoder float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 2.118 ms 654 - 654 MB NPU TrOCR.dlc
TrOCRDecoder float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 2.233 ms 68 - 68 MB NPU TrOCR.onnx.zip
TrOCREncoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 41.885 ms 7 - 116 MB NPU TrOCR.tflite
TrOCREncoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 45.923 ms 2 - 123 MB NPU TrOCR.dlc
TrOCREncoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 22.706 ms 7 - 127 MB NPU TrOCR.tflite
TrOCREncoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 28.73 ms 2 - 134 MB NPU TrOCR.dlc
TrOCREncoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 13.282 ms 6 - 25 MB NPU TrOCR.tflite
TrOCREncoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 16.8 ms 2 - 24 MB NPU TrOCR.dlc
TrOCREncoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 17.835 ms 0 - 137 MB NPU TrOCR.onnx.zip
TrOCREncoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 16.032 ms 7 - 115 MB NPU TrOCR.tflite
TrOCREncoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 19.786 ms 2 - 127 MB NPU TrOCR.dlc
TrOCREncoder float SA7255P ADP Qualcomm® SA7255P TFLITE 41.885 ms 7 - 116 MB NPU TrOCR.tflite
TrOCREncoder float SA7255P ADP Qualcomm® SA7255P QNN_DLC 45.923 ms 2 - 123 MB NPU TrOCR.dlc
TrOCREncoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 13.269 ms 7 - 22 MB NPU TrOCR.tflite
TrOCREncoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 16.832 ms 1 - 21 MB NPU TrOCR.dlc
TrOCREncoder float SA8295P ADP Qualcomm® SA8295P TFLITE 24.334 ms 7 - 124 MB NPU TrOCR.tflite
TrOCREncoder float SA8295P ADP Qualcomm® SA8295P QNN_DLC 25.832 ms 2 - 130 MB NPU TrOCR.dlc
TrOCREncoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 13.284 ms 7 - 22 MB NPU TrOCR.tflite
TrOCREncoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 16.759 ms 2 - 17 MB NPU TrOCR.dlc
TrOCREncoder float SA8775P ADP Qualcomm® SA8775P TFLITE 16.032 ms 7 - 115 MB NPU TrOCR.tflite
TrOCREncoder float SA8775P ADP Qualcomm® SA8775P QNN_DLC 19.786 ms 2 - 127 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 9.448 ms 6 - 119 MB NPU TrOCR.tflite
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 11.472 ms 2 - 136 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 11.6 ms 15 - 184 MB NPU TrOCR.onnx.zip
TrOCREncoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 6.36 ms 6 - 118 MB NPU TrOCR.tflite
TrOCREncoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 9.285 ms 2 - 153 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 9.978 ms 15 - 165 MB NPU TrOCR.onnx.zip
TrOCREncoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile TFLITE 5.489 ms 0 - 111 MB NPU TrOCR.tflite
TrOCREncoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile QNN_DLC 6.947 ms 2 - 129 MB NPU TrOCR.dlc
TrOCREncoder float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile ONNX 7.281 ms 15 - 145 MB NPU TrOCR.onnx.zip
TrOCREncoder float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 17.518 ms 203 - 203 MB NPU TrOCR.dlc
TrOCREncoder float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 18.147 ms 49 - 49 MB NPU TrOCR.onnx.zip

Installation

Install the package via pip:

# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[trocr]"

Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub Workbench 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.

qai-hub configure --api_token API_TOKEN

Navigate to 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.

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.
python -m qai_hub_models.models.trocr.export

How does this work?

This export script leverages Qualcomm® AI Hub 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.

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.

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.

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 Workbench. Sign up for access.

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

  • TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

  • QNN (.so export ): This sample app 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. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of TrOCR can be found here.
  • The license for the compiled assets for on-device deployment can be found here

References

Community

Downloads last month
653
Inference Providers NEW
Image-to-Text
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Space using qualcomm/TrOCR 1