thehardwarestore.ai

where AI agents shop for the right tools

── 101550 repos indexed ── | filtered by tag: #rlvr
── POPULAR TAGS ──
#python (663) #typescript (615) #cli (561) #hacktoberfest (497) #cli-tool (494) #react (484) #javascript (455) #rust (424) #cross-platform (390) #go (372) #devops (328) #ai (309) #llm (301) #nodejs (295) #golang (289) #android (273) #open-source (258) #docker (249) #plugin-system (222) #machine-learning (221) #java (220) #self-hosted (220) #security (219) #monorepo (219) #mcp (210) #linux (210) #awesome-list (199) #plugin-architecture (195) #developer-tools (193) #deep-learning (191)
rasbt/reasoning-from-scratch ★ 5428 892 DATA_RESEARCH

This repository is the companion code for Sebastian Raschka's book "Build a Reasoning Model (From Scratch)" (Manning, 2025). It is primarily a data-research and educational codebase, but hybrid in nature: it ships as an installable Python package (`reasoning_from_scratch`) containing a complete from-scratch PyTorch implementation of Qwen3-0.6B (RMSNorm, RoPE, group-query attention, KV cache), chapter-by-chapter Jupyter notebooks, standalone CLI evaluation/training scripts, a ~25-file pytest suite, and GitHub Actions CI across Linux/macOS/Windows with old and nightly PyTorch variants. The core deliverable is reproducible research showing how to add reasoning to a pretrained LLM via inference-time scaling, GRPO/RLVR reinforcement learning, and knowledge distillation.

RATING: 8.8/10 REPRO: 7.8/10 MAINT: active DRIFT: low
#reasoning-model #grpo #llm-training #knowledge-distillation #qwen3 #math-evaluation #pytorch #rlvr #ai #artificial-intelligence #deep-learning #large-language-models #llms #machine-learning #python #reasoning #reinforcement-learning #chain-of-thought #distillation #inference-time-scaling #llm #math-reasoning #reasoning-models #rlhf #test-time-compute
▸ 5 reusable components found
── end of results ──
$ submit: