---
tags:
- Hong Kong
- 香港
- Cantonese
- 廣東話
- 粵語
---
## CantoneseLLM-6B-preview202402 with ExLlamaV2 Quantization
哩個係用 [/hon9kon9ize/CantoneseLLM-6B-preview202402](https://huggingface.co/hon9kon9ize/CantoneseLLM-6B-preview202402) 生成嘅exl2量化模型。<br>
This is a quantizated model from [/hon9kon9ize/CantoneseLLM-6B-preview202402](https://huggingface.co/hon9kon9ize/CantoneseLLM-6B-preview202402) in exl2 format.<br>
這是一個由 [/hon9kon9ize/CantoneseLLM-6B-preview202402](https://huggingface.co/hon9kon9ize/CantoneseLLM-6B-preview202402) 生成的exl2量化模型。


哩度係main branch, 只係放EvLlamaV2量化果陣用到嘅[measurement.json](measurement.json)檔案，請響下面揀量化程度。<br>
You are currently at the [main](https://huggingface.co/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/main) branch, which provides only [measurement.json](measurement.json) used in the ExLlamaV2 quantization. Please take a look of your choices in following table of branches.<br>
這裡是main branch, 只提供EvLlamaV2量化時所用到的[measurement.json](measurement.json)檔案，請在下面選擇量化程度。。


[8.0bpw-h8](/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/8.0bpw-h8) 8 bits per weight.

[6.0bpw-h6](/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/6.0bpw-h6) 6 bits per weight.

[5.0bpw-h6](/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/5.0bpw-h6) 4 bits per weight.

[4.0bpw-h6](/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/4.0bpw-h6) 4 bits per weight.

[3.0bpw-h6](/kennylam/CantoneseLLM-6B-preview202402-exl2/tree/3.0bpw-h6) 3 bits per weight.