R_REDDYX.XYZ
AIPython★ 856⑂ 4 forksBSD-3-Clause

YeonwooSung/llm-architecture-from-scratch

IMPLEMENTING LLM ARCHITECTURES FROM SCRATCH

The project contains clean implementations of modern LLM architectures in PyTorch. Includes DeepSeek-R1, DeepSeek-V3, Gemma 3, Kimi K2, and Llama 4. Perfect for studying model internals.

// KEY FEATURES

  • Pure PyTorch code without external frameworks for learning
  • Implemented DeepSeek-R1, V3, Gemma 3, Kimi K2, Llama 4
  • Detailed comments for every architecture module
  • 856 stars on GitHub, up-to-date SOTA models
#deepseek-r1#deepseek-v3#gemma3#kimi-k2#llama4#llm#moe#qwen3
Open on GitHub →

New repositories every 30 minutes

REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.

Join on Telegram
← Full catalog·Full index

// SIMILAR REPOSITORIES

mcp-searxng - MCP Server for SearXNG★ 857CREATE OH-MY-TAIYIFORGE PROJECT★ 855BIGMOEONEDGE - SUPER MOE ON PHONE★ 855PROTECTION AGAINST OVERDEFENSE OF AGENTS★ 855SELF-LEARNING QA AGENT★ 857Local Knowledge Graph MCP★ 858

← FULL CATALOG

YeonwooSung/llm-architecture-from-scratch — Implementing LLM…