Hackernews posts about llama2-70B

  1. Post-transformer inference: 224× compression of Llama-70B with improved accuracy (zenodo.org)
    72 points by anima-core 23 days ago | 56 comments
  2. Reverse-engineered a Sextortion Bot: Llama-7B instance with 2048 token window (old.reddit.com)
    15 points by perelin 1 day ago | 1 comments
  3. SEQUOIA: Exact Llama2-70B on an RTX4090 with half-second per-token latency (infini-ai-lab.github.io)
    131 points by zinccat over 1 year ago | 61 comments
  4. EfficientQAT: LLM Quantization, gets a 2-bit llama2-70B outperform regular 13B (old.reddit.com)
    21 points by jackbravo over 1 year ago | discuss
  5. LLMs: What if we did unlimited tokens instead?
    1 points by torrmal over 1 year ago | discuss
  6. Meta AI releases Code Llama 70B (twitter.com)
    598 points by albert_e almost 2 years ago | 294 comments
  7. Ask HN: Code Llama 70B on a dedicated server
    67 points by lavren1974 almost 2 years ago | 42 comments
  8. Run the strongest open-source LLM model: Llama3 70B with just a single 4GB GPU (ai.gopubby.com)
    57 points by qiakai over 1 year ago | 29 comments
  9. The AI Hype Fraud of Reflection-Llama-70B (twitter.com)
    43 points by lnyan over 1 year ago | 3 comments
  10. Show HN: Perplexity (llama3 70B) Inline Bot on Telegram (www.telegrambots.ai)
    13 points by sandoche over 1 year ago | 12 comments
  11. Nvidia has published a competitive llama3-70B QA/RAG fine tune (old.reddit.com)
    12 points by miles over 1 year ago | 2 comments
  12. Editing Files at 1000 tokens/s with llama-70B (www.cursor.com)
    10 points by amanrs over 1 year ago | 4 comments
  13. A Llama 70B finetune that has reflection baked into it's weights (huggingface.co)
    6 points by s-macke over 1 year ago | discuss
  14. Fine-tuning Llama3-70B for GradPilot – easily beats Claude Sonnet (www.alignedhq.ai)
    5 points by pmmucsd over 1 year ago | 1 comments
  15. EagleX 7B with 1.7T tokens beats llama2 7B (2T tokens) on English evals (twitter.com)
    4 points by tosh almost 2 years ago | 2 comments
  16. GroqCloud Makes DeepSeek R1 Distill Llama 70B Available (groq.com)
    4 points by LorenDB 11 months ago | discuss
  17. RWKV-5 "Eagle" 7B: beats Mistral-7B at multilingual,reaches Llama2-7B at English (twitter.com)
    4 points by bratao almost 2 years ago | discuss
  18. Deploying Llama3 70B on AWS – GPU Requirement, Cost and Step-by-Step Guide (www.slashml.com)
    3 points by JJneid over 1 year ago | 5 comments
  19. Llama3-70B mentions "sentience" independently (twitter.com)
    2 points by indigodaddy over 1 year ago | 3 comments
  20. Star Trek prompt optimal for grade school math on Llama-70B (twitter.com)
    2 points by convexstrictly almost 2 years ago | 1 comments
  21. EagleX v2: Soaring past LLaMA2 7B in both English and Multi-lang evals (RWKV-v5) (blog.rwkv.com)
    2 points by oakpond over 1 year ago | discuss
  22. Mixtral or Llama 70B on Google Spreadsheet Thanks to Hugging Face's API (huggingface.co)
    1 points by brulenaudet over 1 year ago | 1 comments
  23. Ask HN: Code Llama 70B on a dedicated server
    1 points by lavren1974 almost 2 years ago | 1 comments
  24. Show HN: I Used Llama-70B Logprobs for Better, Cheaper and Faster Chunking (github.com)
    1 points by ghita_ about 1 year ago | discuss
  25. Show HN: Running llama3-70B on ruby OpenAI client via Cloudflare worker (twitter.com)
    1 points by jackculpan over 1 year ago | discuss
  26. Show HN: DeepSeek Your HN Profile (hn-wrapped.kadoa.com)
    121 points by hubraumhugo 11 months ago | 82 comments
  27. Show HN: Llm2sh – Translate plain-language requests into shell commands (github.com)
    67 points by RandomBK over 1 year ago | 21 comments
  28. Show HN: Wat Dat – A Firefox Extension for Instant Text Explanations (addons.mozilla.org)
    32 points by bsgada over 1 year ago | 18 comments
  29. How fast can one reasonably expect to get inference on a ~70B model?
    9 points by yungtriggz over 1 year ago | 6 comments
  30. Ask HN: Build spec for home LLM box?
    8 points by ActorNightly over 1 year ago | 1 comments
  31. Show HN: GitHub Wrapped – analyzing every file change in every commit (www.wrapped.dev)
    6 points by hellovai about 1 year ago | discuss
  32. Show HN: Meetily – Open-Source AI Meeting Assistant (Alt to Otter.ai)
    4 points by zackriya 10 months ago | 4 comments