# Image Classification Model Card

## Overview
This repository contains an ONNX image classification model for recognizing visual symbol classes in small cropped images.

## Files
- `model_fp32.onnx`: FP32 ONNX model artifact
- `eval_onnx_fp32.json`: Evaluation summary for the ONNX model
- `training_metrics.json`: Training/validation metric logs
- `class_map.json`: Label mapping used by the model

## Intended Use
- Inference in OCR or visual classification pipelines
- Benchmarking and experimentation

## Limitations
- Performance depends on image quality and class balance
- May underperform on out-of-domain samples

## Ethical Considerations
Use this model in lawful and privacy-respecting contexts.