pictograph/receipt-detection

Receipt Detection

An object detection model (YOLOX-M), trained on 1 class, scoring 87.2% mAP.

View on Pictograph · clearobject · Object Detection · YOLOX-M · FP32 · Apache License 2.0

Overview

This is an object detection model built on YOLOX-M. It expects 640x640 input and runs in FP32 precision. The ONNX weights are in this repo; run them with the Pictograph SDK (below) or open the model on Pictograph to test it in-browser, call the hosted API, or deploy it as an always-on endpoint.

Performance

Headline: 87.2% mAP.

Metric	Value
mAP	87.2%
mAP@50	100.0%
Recall	87.5%

Classes

Class index matches the model's output order.

#	Class
0	receipt

Training

Setting	Value
Input size	640x640
Epochs	20
Batch size	12
Learning rate	0.01
Model size	m
Precision	FP32
Version	1.0.0

Use this model

The weights are exported to ONNX (yolox-c4b9f885.onnx). Run them framework-natively with ONNX Runtime, or with the Pictograph SDK, which applies the model's pre/post-processing for you.

ONNX Runtime

from huggingface_hub import hf_hub_download
import onnxruntime as ort

onnx = hf_hub_download("pictograph/receipt-detection", "yolox-c4b9f885.onnx")
sess = ort.InferenceSession(onnx, providers=["CPUExecutionProvider"])
inp = sess.get_inputs()[0]
print(inp.name, inp.shape)  # NCHW float32; preprocess to the model's input size + normalization
# outputs = sess.run(None, {inp.name: preprocessed_batch})

Pictograph SDK

from huggingface_hub import hf_hub_download
from pictograph import load_model, DetectionModel, DetectionResult

onnx = hf_hub_download("pictograph/receipt-detection", "yolox-c4b9f885.onnx")

model: DetectionModel = load_model(
    onnx, task="object_detection", format="onnx",
)
result: DetectionResult = model.predict(image="photo.jpg")
print(result)

• Test it in-browser, call the hosted API, or deploy an endpoint: open the model on Pictograph.
• SDK + local inference guide: docs.

License

Released under Apache License 2.0.

---

Trained and published with Pictograph - annotate, train, and deploy from one API.