A comprehensive, ready-to-use dataset of 500 simulated machine learning training experiments across 8 major frameworks including PyTorch, TensorFlow, JAX, and Hugging Face. Each record contains 45 fields covering hyperparameters, model architecture, hardware configuration, performance metrics, and training insights — making it ideal for ML practitioners, data science students, and researchers.
Perfect for: hyperparameter optimization research, AutoML prototyping, experiment tracking system development, ML education, and building performance prediction models.
What's included: Learning rate, batch size, optimizer, dropout, weight decay, gradient clipping, train/val loss, accuracy, F1, AUC-ROC, perplexity, mAP, GPU utilization, training time, overfitting flags, and more.
Best for: Data scientists, ML engineers, AI researchers, and bootcamp instructors.
with Signup