File size: 527 Bytes
4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 39fadf3 4c108a5 |
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 |
---
pipeline_tag: text-generation
tags:
- exllama
- exl2
- phi
- code
- math
- tchat
- text-generation
---
**Quantization**: ExLlamaV2 (ExL2) at **2.5 bits** per weight.
## Overview
This is an ExLlamaV2 (ExL2) 2.5 bpw quantized version of [microsoft/phi-4](https://huggingface.co/microsoft/phi-4).
## Quantization By
I often have idle A100 GPUs while building/testing the app, so I put them to use quantizing models.
I hope the community finds these quantizations useful.
Andrew Webby @ [RolePlai](https://roleplai.app/)
|