verify-training-pipeline

What problem does it solve?

Verify a CVlization training example is properly structured, can build, trains successfully, and logs appropriate metrics.

Core Features & Use Cases

• Structure validation: ensures essential files (example.yaml, Dockerfile, build.sh, train.sh, train.py) exist and are well-formed.
• Build and training verification: can be run via build.sh / train.sh or through cvl run presets to confirm end-to-end execution.
• Metrics & logging checks: validates that training logs and outputs (checkpoints, logs) are produced and meaningful metrics are captured.
• Debugging & CI readiness: provides a repeatable workflow for debugging issues and integrating into CI pipelines.
• Lazy downloading & caching: ensures datasets and weights are cached and reused to optimize iterations.

Quick Start

Run the verification workflow on a training example to confirm structure, build, training, and metrics logging.