Hackernews posts about Llama3

Llama3 is an AI language model that was implemented from scratch and can be used for tasks such as text generation, translation, and question-answering.

1. Llama 3.1 405B now runs at 969 tokens/s on Cerebras Inference (cerebras.ai)
   426 points by benchmarkist 3 days ago | 155 comments
2. Show HN: Llama 3.2 Interpretability with Sparse Autoencoders (github.com)
   319 points by PaulPauls about 9 hours ago | 29 comments
3. 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
4. Cerebras Inference now 3x faster: Llama3.1-70B breaks 2,100 tokens/s (cerebras.ai)
   147 points by campers 28 days ago | 84 comments
5. HuggingFace - Tencent launches Hunyuan Large which outperforms Llama 3.1 405B (huggingface.co)
   23 points by janik-io 17 days ago | 1 comments
6. Nvidia presents Llama-Mesh: Generating 3D Mesh with Llama 3.1 8B (research.nvidia.com)
   20 points by BUFU 5 days ago | 1 comments
7. Scaling Automatic Neuron Description (Describing Every Neuron in Llama 3) (transluce.org)
   8 points by ekzhang 30 days ago | 1 comments
8. Cerebras runs Llama 3.1 70B at 2,100 tokens per second, live demo available (twitter.com)
   6 points by modeless 29 days ago | discuss
9. Cerebras Inference now runs Llama 3.1-70B at 2100 tokens/s (cerebras.ai)
   6 points by cs-fan-101 29 days ago | discuss
10. Show HN: Finetune Llama 3.2 Vision in a Colab (colab.research.google.com)
    4 points by danielhanchen about 10 hours ago | discuss
11. LLama3.2-vision as almost perfect OCR? (demo.doctractor.com)
    3 points by pkarwatka 4 days ago | 1 comments
12. Chat with Llama 3.1 8B (yosunai.com)
    3 points by vijayst 6 days ago | 1 comments
13. CurlDuck, curl script for DuckDuckGo LLMs like Llama3 (github.com)
    2 points by zoobab 4 days ago | discuss
14. Llama3.2-Vision on Ollama (ollama.com)
    2 points by taran_narat 14 days ago | discuss
15. Llama-3.1-Nemotron-70B-Instruct Model (build.nvidia.com)
    2 points by lapnect 30 days ago | discuss
16. Best Open Source LLM:Tencent Hunyuan-Large Outperforming Llama 3 and DeepSeek V2 (techwavearena.com)
    1 points by tiime 14 days ago | discuss
17. On Device Llama 3.1 with Core ML (machinelearning.apple.com)
    1 points by 2bit 21 days ago | discuss
18. How to Run Meta Llama 3.1 405B with Nebius AI Studio API (nebius.com)
    1 points by dsaed 24 days ago | discuss
19. Ask HN: Local RAG with private knowledge base
    108 points by zephodb 25 days ago | 40 comments
20. Ask HN: AI Cloud Computing, is it cheaper than the OpenAI API in the end?
    2 points by calipsow 29 days ago | 4 comments
21. Show HN: Open-Source Alternative for GitHub Copilot (marketplace.visualstudio.com)
    1 points by suneelak 19 days ago | 4 comments
22. Show HN: Senti – Offline LLM Voice Chat (testflight.apple.com)
    1 points by syousif 22 days ago | 1 comments
23. Quantized Llama models with increased speed and a reduced memory footprint (ai.meta.com)
    508 points by egnehots 28 days ago | 122 comments
24. Llama-OCR: Document to Markdown (llamaocr.com)
    293 points by lapnect 6 days ago | 96 comments
25. Go library for in-process vector search and embeddings with llama.cpp (github.com)
    114 points by kelindar 25 days ago | 29 comments
26. Show HN: LlamaPReview – AI GitHub PR reviewer that learns your codebase (github.com)
    102 points by Jet_Xu 23 days ago | 42 comments
27. Cerebras Trains Llama Models to Leap over GPUs (www.nextplatform.com)
    64 points by rbanffy 27 days ago | 33 comments
28. LlamaChunk: Better RAG Chunking Than LlamaIndex (github.com)
    15 points by npip99 7 days ago | 5 comments
29. Chinese researchers develop AI model for military use on back of Meta's Llama (www.reuters.com)
    11 points by fnordpiglet 21 days ago | 1 comments
30. 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