Hackernews posts about Llama2

Llama2 is a large language model that has achieved impressive performance in natural language processing tasks, outperforming ChatGPT and other models in various benchmarks.

1. Help needed to improve Haskell port of Andrej Karpathy's llama2.c (github.com)
   3 points by CertumIter 24 days ago | discuss
2. Sagence emerges from stealth promising Llama 2 at 10 percent power (spectrum.ieee.org)
   1 points by pseudolus 2 days ago | discuss
3. Quantized Llama models with increased speed and a reduced memory footprint (ai.meta.com)
   508 points by egnehots 28 days ago | 122 comments
4. Llama 3.1 405B now runs at 969 tokens/s on Cerebras Inference (cerebras.ai)
   426 points by benchmarkist 3 days ago | 155 comments
5. Llama-OCR: Document to Markdown (llamaocr.com)
   293 points by lapnect 5 days ago | 96 comments
6. Ollama 0.4 is released with support for Meta's Llama 3.2 Vision models locally (ollama.com)
   182 points by BUFU 15 days ago | 25 comments
7. Cerebras Inference now 3x faster: Llama3.1-70B breaks 2,100 tokens/s (cerebras.ai)
   147 points by campers 28 days ago | 84 comments
8. Go library for in-process vector search and embeddings with llama.cpp (github.com)
   114 points by kelindar 24 days ago | 29 comments
9. Show HN: LlamaPReview – AI GitHub PR reviewer that learns your codebase (github.com)
   102 points by Jet_Xu 22 days ago | 42 comments
10. Cerebras Trains Llama Models to Leap over GPUs (www.nextplatform.com)
    64 points by rbanffy 26 days ago | 33 comments
11. HuggingFace - Tencent launches Hunyuan Large which outperforms Llama 3.1 405B (huggingface.co)
    23 points by janik-io 16 days ago | 1 comments
12. Nvidia presents Llama-Mesh: Generating 3D Mesh with Llama 3.1 8B (research.nvidia.com)
    20 points by BUFU 5 days ago | 1 comments
13. LlamaChunk: Better RAG Chunking Than LlamaIndex (github.com)
    15 points by npip99 6 days ago | 5 comments
14. Chinese researchers develop AI model for military use on back of Meta's Llama (www.reuters.com)
    11 points by fnordpiglet 20 days ago | 1 comments
15. Meta is using more than 100k Nvidia H100 AI GPUs to train Llama-4 (www.tomshardware.com)
    8 points by turtlegrids 20 days ago | 1 comments
16. Scaling Automatic Neuron Description (Describing Every Neuron in Llama 3) (transluce.org)
    8 points by ekzhang 29 days ago | 1 comments
17. Defense Llama: The LLM Purpose-Built for American National Security (scale.com)
    8 points by ushakov 16 days ago | discuss
18. Cerebras runs Llama 3.1 70B at 2,100 tokens per second, live demo available (twitter.com)
    6 points by modeless 28 days ago | discuss
19. Cerebras Inference now runs Llama 3.1-70B at 2100 tokens/s (cerebras.ai)
    6 points by cs-fan-101 28 days ago | discuss
20. Chinese researchers develop AI model for military use on back of Meta's Llama (www.reuters.com)
    5 points by consumer451 20 days ago | 1 comments
21. Pdf-extract-API – OCR with marker-pdf and transform with Llama – single API call (github.com)
    5 points by pkarwatka 23 days ago | 1 comments
22. Meta to let US national security agencies and defense contractors use Llama AI (www.theguardian.com)
    5 points by c420 16 days ago | discuss
23. Chinese researchers build military AI using Meta's open-source Llama model (www.tomshardware.com)
    5 points by AnhTho_FR 17 days ago | discuss
24. Llama-Mesh: Unifying 3D Mesh Generation with Language Models (research.nvidia.com)
    4 points by lnyan 6 days ago | discuss
25. Show HN: Fast and Cheap Llama-405B (centml.ai)
    4 points by akbarnur 17 days ago | discuss
26. LLama3.2-vision as almost perfect OCR? (demo.doctractor.com)
    3 points by pkarwatka 3 days ago | 1 comments
27. Chat with Llama 3.1 8B (yosunai.com)
    3 points by vijayst 6 days ago | 1 comments
28. PRC Adapts Meta's Llama for Military and Security AI Applications (jamestown.org)
    3 points by IndrekR 16 days ago | discuss
29. Llama 405B up to 142 tok/s on Nvidia H200 SXM (old.reddit.com)
    2 points by avianion 24 days ago | 1 comments
30. Show HN: LlamaPReview – AI code reviewer trusted by 2000 repos, 40%+ effective (jetxu-llm.github.io)
    2 points by Jet_Xu 2 days ago | discuss