When you ask some tasks to be done, code interpreter understands the task and write the code to the task done. And then, it runs the code for you to prvide the result. It was once provided only for the paid user in ChatGPT. While I was talking to perplexity, I found the library called open-interpreter by chance, a python library freely available.
It is pretty easy to install and use by connecting open-interpreter with ollama. It seems that the library was designed for jupyter notebook (or any similar program). I need to do lot of data analysis. So I provided directory of the file and the file name to do some analysis. Open-interpreter uses LLMs in ollama to write the code and run it to do the task. When it writes a code, it first asks me to confirm the code before running. The below link shows how to install and use open-interpreter.
https://docs.openinterpreter.com/language-models/local-models/ollama
Ollama - Open Interpreter
If Ollama is producing strange output, make sure to update to the latest version
docs.openinterpreter.com
I was always short for hands, and it feels like I have extra hand coding while I'm busy wth other thinkgs to do. Ths is pretty nice. Hugging face can be also connected to open-interpreter. When I use open-interpreter in my laptop, I use small LLM, which is good enough for simple coding. When I can't really focus on coding, open-interpreter is a great help to do the task.
'GenAI' 카테고리의 다른 글
Coding agent making a game within five minutes (0) | 2025.04.15 |
---|---|
Easy way of making RAG (AI for document summary: AnythingLLM with Ollama) (0) | 2025.04.13 |
Cartoon generated by AI (a guide on Gen AI and copyright) (0) | 2025.04.12 |
MUST USE AI services for startups and investors (TOP 5) (1) | 2024.11.26 |
Is AI AGI? (1) | 2024.11.14 |