ollama
@ollama
https://t.co/rx433zDvXt
ID:1688410127378829312
https://github.com/ollama/ollama 07-08-2023 04:42:08
2,7K Tweets
33,2K Followers
7 Following
L4, L40S, V100, and more are available directly through the platform with dedicated vCPUs and up to 96GB of RAM. π₯
In 60 secs, you can run any Docker Build Cloud is here! π³π§±βοΈ container with GPUs using the CLI, API, or dashboard. Spawn your favorite LLMs with ollama in one call. π
- clickhouse local (ClickHouse)
- llm (@simonw)
- ollama (@ollama)
- llama 3 8b q8 (@AIatMeta)
w/ an apple health file converted to .parquet
using π expanse (github.com/tosh/expanse)
π§ββοΈπͺ witchery!
If you use Ollama, you have your 100% private 'AI Internet' running INSIDE your computer.
You can verify this easily. Simply turn off your wifi and try running an inference. Here I'm chatting with Lllama3 over Open WebUI (Formerly Ollama WebUI) with @Ollama while the WIFI is turned off.
Before we begin, take a look at what we're about to create!
As always, I'll be using my favourite stack:
- @Ollama for locally serving a LLM (Llama-3)
- @Llama_Index for orchestration
- @Streamlit for building the UI
- Lightning AI β‘οΈ for development & hosting
Let's go! π