LlmsContinued LLM Pretraining with Unsloth
Make a model learn a new language by doing continued pretraining with Unsloth using Llama 3, Phi-3 and Mistral.
LlmsInterface for local and remote LLM (llama/chatgpt)
An interface for llama.cpp and ChatGPT. Contribute to Merkoba/Meltdown development by creating an account on GitHub.
LlmsStanford researchers unveil the Foundation Model Transparency Index, with 100 indicators; Llama 2 leads at 54%, GPT-4 is third at 48%, PaLM 2 is fifth at 40% (Kevin Roose/New York Times)
Kevin Roose / New York Times:nStanford researchers unveil the Foundation Model Transparency Index, with 100 indicators; Llama 2 leads at 54%, GPT-4 is …
LlmsPurpose of Pipeline/Chain Abstractions in LLM Frameworks?
n Article URL: https://old.reddit.com/r/LocalLLaMA/comments/174eckz/purpose_of_pipelinechain_abstractions/ n Comments URL: …
LlmsAfter using the 32k GPT4 model, I've lost interest in 4K/8K models
Using GPT4 with a massive long-ass context window is honestly the absolutely best I've seen AI do anything. The quality shoots up massively, and...
LlmsOllama – Easily Run LLMs on your laptop
n Article URL: https://ollama.ai/ n Comments URL: https://news.ycombinator.com/item?id=36808754 n Points: 22 n # Comments: 2 n