File size: 1,097 Bytes
cedbb16 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 |
---
license: mit
language:
- en
pipeline_tag: text-generation
tags:
- astronomy
- science
- LLM
---
# AstroLLaMA.gguf
[AstroLLaMA](https://huggingface.co/universeTBD/astrollama) in GPT-Generated Unified Format.
## What's GPT-Generated Unified Format?
GGUF is file format used for storing models for inference, particularly in the context of language models like GPT. GGUF models **could be run on CPUs**, which made them accessible to a wider range of users.
More detailed description can be found [here](https://medium.com/@phillipgimmi/what-is-gguf-and-ggml-e364834d241c).
## How to play with AstroLLaMA on your laptop?
1. Install [ollama](https://github.com/jmorganca/ollama);
2. Download `astrollama.gguf` to your PC;
3. Create a file named Modelfile, and add a FROM instruction with the local filepath to AstroLLaMA, e.g.,
```
FROM ./astrollama.gguf
```
4. Create the model in Ollama
```
ollama create astrollama -f path_to_modelfile
```
5. Run AstroLLaMA locally
```
ollama run astrollama
``` |