{"name":"samir-vllm","repo":"samir-sarwar/samir-vllm","url":"https://midorreal.com/project/samir-sarwar-samir-vllm","repository":"https://github.com/samir-sarwar/samir-vllm","description":"small LLM inference engine from scratch, w/ 6+ hours of youtube tutorials & 85+ pages of handwritten notes included","facts":{"language":{"value":"C++","source":"https://github.com/samir-sarwar/samir-vllm","checked":"22 Nov 2026"},"releases":{"value":"0","source":"https://github.com/samir-sarwar/samir-vllm/releases","checked":"21 Nov 2026"},"contributors":{"value":"1","source":"https://github.com/samir-sarwar/samir-vllm/graphs/contributors","checked":"21 Nov 2026"},"archived":{"value":"no","source":"https://github.com/samir-sarwar/samir-vllm","checked":"22 Nov 2026"}},"stars":241.0,"writeup":{"what_it_is":"samir-vllm is a small LLM inference engine written in C++17 and CUDA, built to teach how inference works. It loads Llama 3.2 1B Instruct weights from a safetensors file, runs the transformer layers on the GPU, and supports greedy generation with static or continuous batching and a paged KV cache.","audience":"Learners studying how LLM inference works internally","claims":[{"kind":"specific","claim":"Runs the Llama 3.2 1B Instruct model with dense causal prefill across all 16 layers","excerpt":"Dense causal prefill now runs all 16 layers","status":"not_checked"},{"kind":"specific","claim":"Uses a paged KV cache with 16-token pages","excerpt":"Each physical page holds 16 tokens of K and V","status":"not_checked"},{"kind":"specific","claim":"Supports static and continuous batching","excerpt":"Static batching and a continuous scheduler that admit queued prompts according to available page budgets","status":"not_checked"}],"alternatives":[],"written_by":"AI, from the project's README","date":"2026-11-10"},"why_now":null,"owner_supplied":null,"score":{"verdict":"too_early","hype":92,"reality":20,"gap":72,"momentum":null,"confidence":60,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T07:00:00+00:00","verdict_source":"rule"},"early_read":{"estimate":true,"lean":"too_close","gap_low":null,"gap_high":null,"rule":"early-2.0"},"changes":[{"date":"2026-08-27T00:17:33+00:00","text":"Repository created","source":"https://github.com/samir-sarwar/samir-vllm"}],"cite":"samir-vllm is a GitHub project at https://github.com/samir-sarwar/samir-vllm written mainly in C++, described there as \"small LLM inference engine from scratch, w/ 6+ hours of youtube tutorials & 85+ pages of handwritten notes included\".","attribution":"Data from Mid or Real, https://midorreal.com/project/samir-sarwar-samir-vllm, checked 22 November 2026. Attribution with a link is required.","still_real":null}