Nikolaos Vlachostergios@nikolaosvlachostergios81618

AMD ROCm OCR Challenge Submission This project implements an OCR inference application packaged as a Docker container using the required AMD ROCm PyTorch base image. The container provides a command-line interface that accepts an input image and an optional output directory, performs OCR processing, and produces the corresponding JSON output. The solution is packaged for execution in the challenge evaluation environment and uses the required ROCm/PyTorch environment as its base. The Docker image has been built locally and tested to verify that the container starts correctly and that the application exposes the required command-line interface. Container interface: python3 /app/app.py --input-image <path> --output-dir <path> The submitted Docker image is publicly pullable from the specified container registry so that it can be fetched and executed directly by the evaluation infrastructure.

Six Matches. One leaderboard. Build on AMD, earn points and compete for prizes.