Extremely simple and fast word2vec implementation with Negative Sampling + Sub-sampling
-
Updated
Jan 21, 2021 - Python
Extremely simple and fast word2vec implementation with Negative Sampling + Sub-sampling
Implements https://arxiv.org/abs/1711.05101 AdamW optimizer, cosine learning rate scheduler and "Cyclical Learning Rates for Training Neural Networks" https://arxiv.org/abs/1506.01186 for PyTorch framework
sharpDARTS: Faster and More Accurate Differentiable Architecture Search
SGDR: Stochastic Gradient Descent with Warm Restarts and Cosine Annealing learning rate schedule engine.
SGDR: Stochastic Gradient Descent with Warm Restarts and Cosine Annealing learning rate schedule engine.
End-to-end Image Classification using Deep Learning toolkit for custom image datasets. Features include Pre-Processing, Training with Multiple CNN Architectures and Statistical Inference Tools. Special utilities for RAM optimization, Learning Rate Scheduling, and Detailed Code Comments are included.
Implemented Deep Residual Learning for Image Recognition Paper and achieved lower error rate by customizing different parts of the architecture.
A comprehensive, research-driven collection of learning rate schedulers for PyTorch — 17 schedulers, composable warmup, opinionated presets, and first-class paper references.
self-used pytorch utilities
Short-Term Wind Power Forecasting utilizing a Transformer model and a seasonally aware custom cosine annealing scheduler.
Framework-agnostic learning-rate schedules as pure functions, in Python with zero dependencies.
TinyYoloV2 imagenet 1K results.
Add some useful functions based on AlexeyAB darknet
Minimal PyTorch implementation of cosine annealing with warm-up for super-convergence training schedules.
REST API for classifying 100 object categories. VGG-style CNN in PyTorch (70% accuracy), FastAPI, trained from scratch without transfer learning. Inference < 100 ms.
High-performance PyTorch LR schedulers with cosine annealing, flexible waypoints, plateau steps, and LR scaling. Unified API with pre-computed segments for zero runtime overhead.
To associate your repository with the cosine-annealing topic, visit your repo's landing page and select "manage topics."