andyyyy64

3 ms·

1. 1.▲
   by andyyyy64 5mo ago

   You're right that it doesn't run anything — it's a pre-download / pre-purchase decision tool, so it estimates rather than measures by design (you can simulate a GPU you don't own with --gpu). That's a genuine l
2. 2.▲
   by andyyyy64 5mo ago

   Good catch that's a real gap. The KV estimate is GQA/MQA-aware (per-model head config) but currently assumes dense full-context attention; it does not model sliding-window / chunked attention, so for SWA models like Mistral
3. 3.▲
   by andyyyy64 5mo ago

   Fair question. llmfit answers "will this model fit in my memory?" — it's a fit/size calculator, and a good one. whichllm answers a different question: "of the models that fit, which is actually best?" It pulls
4. 4.▲
   Show HN: Find the best local LLM for your hardware, ranked by benchmarks(github.com)

   283 points by andyyyy64 5mo ago|68 comments
5. 5.▲
   Show HN: Whichllm – Find and run the best local LLM for your hardware(github.com)

   3 points by andyyyy64 7mo ago|0 comments
6. 6.▲
   Show HN: OpenTiger – Autonomous dev orchestration that never stops(github.com)

   11 points by andyyyy64 7mo ago|2 comments