README.md

---
base_model:
- facebook/detr-resnet-101
language:
- en
- es
library_name: transformers
---

# 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](Demo)

## 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

```python
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](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **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]