Hugging Face's logo

qualcomm
/
TrOCR

Image-to-Text
PyTorch
LiteRT
ONNX
android
Model card Files 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 (TrOCREncoder): 23.0M
    • Model size (TrOCREncoder): 87.8 MB
    • Number of parameters (TrOCRDecoder): 38.3M
    • Model size (TrOCRDecoder): 146 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 4.266 ms 0 - 82 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 4.161 ms 5 - 73 MB NPU TrOCR.dlc
TrOCRDecoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 2.446 ms 0 - 143 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 2.592 ms 3 - 130 MB NPU TrOCR.dlc
TrOCRDecoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 2.05 ms 0 - 455 MB NPU TrOCR.tflite
TrOCRDecoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 2.042 ms 3 - 29 MB NPU TrOCR.dlc
TrOCRDecoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 2.928 ms 0 - 82 MB NPU TrOCR.tflite
TrOCRDecoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 2.809 ms 0 - 64 MB NPU TrOCR.dlc
TrOCRDecoder float SA7255P ADP Qualcomm® SA7255P TFLITE 4.266 ms 0 - 82 MB NPU TrOCR.tflite
TrOCRDecoder float SA7255P ADP Qualcomm® SA7255P QNN_DLC 4.161 ms 5 - 73 MB NPU TrOCR.dlc
TrOCRDecoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 2.061 ms 0 - 509 MB NPU TrOCR.tflite
TrOCRDecoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 2.091 ms 3 - 27 MB NPU TrOCR.dlc
TrOCRDecoder float SA8295P ADP Qualcomm® SA8295P TFLITE 2.883 ms 0 - 75 MB NPU TrOCR.tflite
TrOCRDecoder float SA8295P ADP Qualcomm® SA8295P QNN_DLC 2.934 ms 7 - 65 MB NPU TrOCR.dlc
TrOCRDecoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 2.046 ms 0 - 446 MB NPU TrOCR.tflite
TrOCRDecoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 2.05 ms 1 - 25 MB NPU TrOCR.dlc
TrOCRDecoder float SA8775P ADP Qualcomm® SA8775P TFLITE 2.928 ms 0 - 82 MB NPU TrOCR.tflite
TrOCRDecoder float SA8775P ADP Qualcomm® SA8775P QNN_DLC 2.809 ms 0 - 64 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile TFLITE 2.07 ms 0 - 515 MB NPU TrOCR.tflite
TrOCRDecoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile QNN_DLC 2.034 ms 2 - 27 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile ONNX 2.61 ms 0 - 243 MB NPU TrOCR.onnx
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 1.497 ms 0 - 156 MB NPU TrOCR.tflite
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 1.499 ms 0 - 144 MB NPU TrOCR.dlc
TrOCRDecoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 2.059 ms 0 - 145 MB NPU TrOCR.onnx
TrOCRDecoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile TFLITE 1.366 ms 0 - 79 MB NPU TrOCR.tflite
TrOCRDecoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile QNN_DLC 1.357 ms 1 - 155 MB NPU TrOCR.dlc
TrOCRDecoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile ONNX 1.675 ms 1 - 157 MB NPU TrOCR.onnx
TrOCRDecoder float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 2.279 ms 675 - 675 MB NPU TrOCR.dlc
TrOCRDecoder float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 2.371 ms 68 - 68 MB NPU TrOCR.onnx
TrOCREncoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 77.528 ms 7 - 167 MB NPU TrOCR.tflite
TrOCREncoder float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 72.833 ms 2 - 149 MB NPU TrOCR.dlc
TrOCREncoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 52.189 ms 7 - 171 MB NPU TrOCR.tflite
TrOCREncoder float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 64.665 ms 1 - 149 MB NPU TrOCR.dlc
TrOCREncoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 40.253 ms 7 - 26 MB NPU TrOCR.tflite
TrOCREncoder float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 38.966 ms 2 - 37 MB NPU TrOCR.dlc
TrOCREncoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 42.5 ms 7 - 167 MB NPU TrOCR.tflite
TrOCREncoder float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 39.725 ms 2 - 150 MB NPU TrOCR.dlc
TrOCREncoder float SA7255P ADP Qualcomm® SA7255P TFLITE 77.528 ms 7 - 167 MB NPU TrOCR.tflite
TrOCREncoder float SA7255P ADP Qualcomm® SA7255P QNN_DLC 72.833 ms 2 - 149 MB NPU TrOCR.dlc
TrOCREncoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 40.72 ms 7 - 30 MB NPU TrOCR.tflite
TrOCREncoder float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 38.837 ms 2 - 35 MB NPU TrOCR.dlc
TrOCREncoder float SA8295P ADP Qualcomm® SA8295P TFLITE 53.48 ms 7 - 165 MB NPU TrOCR.tflite
TrOCREncoder float SA8295P ADP Qualcomm® SA8295P QNN_DLC 50.008 ms 2 - 148 MB NPU TrOCR.dlc
TrOCREncoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 40.894 ms 7 - 32 MB NPU TrOCR.tflite
TrOCREncoder float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 39.043 ms 1 - 35 MB NPU TrOCR.dlc
TrOCREncoder float SA8775P ADP Qualcomm® SA8775P TFLITE 42.5 ms 7 - 167 MB NPU TrOCR.tflite
TrOCREncoder float SA8775P ADP Qualcomm® SA8775P QNN_DLC 39.725 ms 2 - 150 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile TFLITE 40.941 ms 7 - 31 MB NPU TrOCR.tflite
TrOCREncoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile QNN_DLC 38.753 ms 2 - 37 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile ONNX 40.111 ms 14 - 140 MB NPU TrOCR.onnx
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 32.336 ms 134 - 296 MB NPU TrOCR.tflite
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 31.86 ms 2 - 155 MB NPU TrOCR.dlc
TrOCREncoder float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 32.19 ms 12 - 169 MB NPU TrOCR.onnx
TrOCREncoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile TFLITE 28.294 ms 5 - 165 MB NPU TrOCR.tflite
TrOCREncoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile QNN_DLC 28.687 ms 2 - 176 MB NPU TrOCR.dlc
TrOCREncoder float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile ONNX 27.961 ms 14 - 203 MB NPU TrOCR.onnx
TrOCREncoder float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 37.201 ms 102 - 102 MB NPU TrOCR.dlc
TrOCREncoder float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 37.174 ms 50 - 50 MB NPU TrOCR.onnx

Installation

Install the package via pip:

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 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

Profiling Results
------------------------------------------------------------
TrOCRDecoder
Device                          : cs_8275 (ANDROID 14)                 
Runtime                         : TFLITE                               
Estimated inference time (ms)   : 4.3                                  
Estimated peak memory usage (MB): [0, 82]                              
Total # Ops                     : 399                                  
Compute Unit(s)                 : npu (399 ops) gpu (0 ops) cpu (0 ops)

------------------------------------------------------------
TrOCREncoder
Device                          : cs_8275 (ANDROID 14)                 
Runtime                         : TFLITE                               
Estimated inference time (ms)   : 77.5                                 
Estimated peak memory usage (MB): [7, 167]                             
Total # Ops                     : 603                                  
Compute Unit(s)                 : npu (603 ops) gpu (0 ops) cpu (0 ops)

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 S24")

# 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. 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
277
Inference Providers NEW
Image-to-Text
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Space using qualcomm/TrOCR 1