detr_handwriten_cursive_text_detection

Model Card for Model ID

DETR allows to detect and generate the bounding boxes for handwritten and cursive text. This model was finetuned using the base model facebook/detr-resnet-101. The dataset used is still under development and possible released in future versions. Mainly, the model detects spanish text. Note: The default value of generated bounding boxes was used (num_queries: 100). Modifying this value when using the model could lead to unexpected behavior.

Model Details

Model Description

<!-- Provide a longer summary of what this model is. -->

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

• Developed by: Rodrigo Alvarez
• Funded by [optional]: [More Information Needed]
• Shared by [optional]: [More Information Needed]
• Model type: Text Detection / Bounding Box generation
• Language(s) (NLP): en (default), es-MX (finetuned)
• License: [More Information Needed]
• Finetuned from model [optional]: facebook/detr-resnet-101

Model Sources [optional]

<!-- Provide the basic links for the model. -->

• Repository: [https://github.com/rodrigoalvarez-20/detr_trocr_handwritten_text/development](DETR TROCR Lab)
• Paper [optional]: Work in progress
• Demo [optional]: https://github.com/rodrigoalvarez-20/detr_trocr_handwritten_text/blob/development/detr_lab.ipynb

Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

Direct Use

from transformers import DetrForObjectDetection, DetrImageProcessor
import torch
import cv2
import supervision as sv
# User defined constants
MODEL_CHECKPOINT = "Rodr16020/detr_handwriten_cursive_text_detection"
DEVICE = "cuda"
CONFIDENCE_TRESHOLD = 0.5 # This parameter allows to filter the generated boxes with a confidence score >= to this value
IOU_TRESHOLD = 0.5
TEST_IMAGE = "demo.jpeg" # Path to the test image
#Load the model and preprocessor
img_proc = DetrImageProcessor.from_pretrained(MODEL_CHECKPOINT)
detr_model = DetrForObjectDetection.from_pretrained(
    pretrained_model_name_or_path=MODEL_CHECKPOINT,
    ignore_mismatched_sizes=True
).to(DEVICE)
# Get the pixel values of the image (matrix)
image = cv2.imread(TEST_IMAGE)
# inference
with torch.no_grad():
    # load image and predict
    inputs = img_proc(images=image, return_tensors='pt').to(DEVICE)
    outputs = detr_model(**inputs)
    # post-process
    # Resize the generated Bounding Boxes coords to the image original size
    target_sizes = torch.tensor([image.shape[:2]]).to(DEVICE)
    results = img_proc.post_process_object_detection(
        outputs=outputs, 
        threshold=CONFIDENCE_TRESHOLD, 
        target_sizes=target_sizes
    )[0]

# To extract all the generated bboxes
boxes = results["boxes"].tolist()[0]
# With supervision lib, use the generated coords to annotate the image and preview the boxes
box_annotator = sv.BoxAnnotator()
detections = sv.Detections.from_transformers(transformers_results=results).with_nms(threshold=0.1)
labels = [f"{confidence:.2f}" for _,_, confidence, class_id, _ in detections]
frame = box_annotator.annotate(scene=image.copy(), detections=detections, labels=labels)
sv.plot_image(frame, (16, 16))

[More Information Needed]

Downstream Use [optional]

<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->

[More Information Needed]

Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->

[More Information Needed]

Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

[More Information Needed]

Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

Training Details

Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

[More Information Needed]

Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->

Preprocessing [optional]

[More Information Needed]

Training Hyperparameters

• Dataset Format: COCO
• Device: CUDA
• WEIGHT_DECAY = 3e-3
• CLIP_GRAD = 1e-4 #0.001
• BATCH_SIZE = 8
• ACC_BATCH = BATCH_SIZE * 4
• MODEL_LR = 5e-4 # In some articles, they set the value to 5e-4, but, in my case, it doesn't work, so I try with this and works "well"
• BB_LR = 5e-4 # Same as above
• MAX_EPOCHS = 300 # Use >= 50 . But it stops learning near the step 70

Speeds, Sizes, Times [optional]

<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->

[More Information Needed]

Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

Testing Data, Factors & Metrics

Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]

Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

[More Information Needed]

Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[More Information Needed]

Results

[More Information Needed]

Summary

Model Examination [optional]

<!-- Relevant interpretability work for the model goes here -->

[More Information Needed]

Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

• Hardware Type: [More Information Needed]
• Hours used: [More Information Needed]
• Cloud Provider: [More Information Needed]
• Compute Region: [More Information Needed]
• Carbon Emitted: [More Information Needed]

Technical Specifications [optional]

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

A simple and a tiny computer at CIC research lab.

When finetuning, the model and data used a total of

Hardware

• ASRock-placa base Z370/OEM
• Gabinete Corsair 4000D Airflow
• Procesador Intel Core i7 i7-8700K
• Memoria RAM XPG Spectrix DDR4, 3200MHz, 16GB (x4)
• SSD Externo Western Digital WD My Passport, 1TB
• NVIDIA GeForce RTX 4090 24GB
• Corsair Serie RMX, RM1000x, 1000 W

Software

• transformers
• pytorch
• tensorboard
• cv2
• supervision

And possibly others

Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Glossary [optional]

<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]