Flexible ML Experiment Tracking System for Python Coders with DVC and St...
summary
Builds a reproducible Python ML tracking stack that actually wires DVC to Git and Streamlit, trains a TensorFlow cat-vs-dog model on ~3,000 images, and demonstrates running pipelines with dvc repro, queuing experiments and inspecting non‑linear experiment commits. Shows code that DVC experiments create git refs and that outputs and prediction CSVs are versioned and retrievable; demos include a Streamlit app loads models from commits and a diffing UI that compares model predictions and works live.
AI-written summary · about
chapters (5)
- 0:00Start
- 9:01DVC Reproducible Pipelines
- 10:59Run an Experiment, Second Try
- 12:53DVC Experiments
- 14:50Run an Experiment, Third Try
related talks
46mFlexible ML Experiment Tracking System for Python Coders with DVC and Streamlit
37mOpen Source Tools for ML Experiments Management
27mLeveraging Git for ML Experiment Management
42mApplications in ML Drug Discovery pipelines
26mTransforming a Jupyter Notebook into a reproducible pipeline for ML experiments
01h 32mHow to easily set up and version control your Machine Learning Pipelines
59mDevOps for Machine Learning Projects
33mHow to Track your Development Process with DVC