# FareedKhan-dev/train-llm-from-scratch > Educational project: a from-scratch PyTorch Transformer plus a full LLM training pipeline — from raw text through SFT, reward modeling, PPO/DPO/GRPO to chat, without transformers, trl or peft. - Magnitude: 1.4 out of 10 — Cooling - Stars: 11,931 total · +3 stars on 2026-10-03 (UTC), ≈ 11 by evening - Star trust: star growth looks organic - Category: Language models · Language: Python · License: MIT · Created: 2025-01-12 · Last push: 2026-10-02 - GitHub: https://github.com/FareedKhan-dev/train-llm-from-scratch · Homepage: https://fareedkhan-dev.github.io/train-llm-from-scratch/ · Page: https://gitnova.dev/en/r/FareedKhan-dev/train-llm-from-scratch ## Useful for - Train a small LLM on a single GPU following the README steps - Study attention and Transformer internals from the source code - Walk through post-training stages: SFT, reward model, DPO, PPO, GRPO ## How to get started ```sh git clone https://github.com/FareedKhan-dev/train-llm-from-scratch.git cd train-llm-from-scratch pip install -e . ``` Copied as is from the “setup” section of the README. GitNova has not run these commands: check them in the original before running: https://github.com/FareedKhan-dev/train-llm-from-scratch#readme ## Why it’s here - 3 stars so far today, about 11 expected by the end of the day. - Interest is fading: the two-day pace is 9% of the previous week and a half. ## Star trust Star growth looks organic. Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer. ## Numbers - Forks: 1,642 - Issues and pull requests: 41 - Watchers: 87 - Average over the last week: 96 per day - Usual pace: 104 per day ## Stars per day, last 30 days (oldest → newest, today is partial) 2026-10-18 … 2026-11-16: None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None ## Magnitude by day, last 14 days 09-20 6.8, 09-21 7.1, 09-22 6.3, 09-23 4.1, 09-24 2.7, 09-25 3.7, 09-26 3.7, 09-27 3.5, 09-28 3.7, 09-29 3.6, 09-30 3.1, 10-01 2.3, 10-02 1.9, 10-03 1.5 --- Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score. Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast. Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself. Data as of 2026-10-03 10:57 UTC, updated every 30 minutes.