Openchat 3.6 8b 20240522 IMat GGUF

🚀 openchat-3.6-8b-20240522-IMat-GGUF

Llama.cpp imatrix quantization of openchat/openchat-3.6-8b-20240522

This project offers a quantized version of the openchat/openchat-3.6-8b-20240522 model using the llama.cpp imatrix quantization method. It provides various quantization types in the GGUF format, making it suitable for different resource requirements.

Original Model: openchat/openchat-3.6-8b-20240522
Original dtype: BF16 (bfloat16)
Quantized by: llama.cpp b3006
IMatrix dataset: here

✨ Features

• Multiple Quantization Types: Offers a wide range of quantization types, including IMatrix and common quantizations, to meet different resource and performance needs.
• Easy Download: Supports downloading using the huggingface-cli, with instructions for handling split model files.
• Inference Templates: Provides simple chat templates and chat templates with system prompts for inference.

📦 Installation

If you do not have huggingface-cli installed, you can install it using the following command:

pip install -U "huggingface_hub[cli]"

💻 Usage Examples

Downloading using huggingface-cli

To download a specific file:

huggingface-cli download legraphista/openchat-3.6-8b-20240522-IMat-GGUF --include "openchat-3.6-8b-20240522.Q8_0.gguf" --local-dir ./

If the model file is split into multiple files, you can download them all to a local folder:

huggingface-cli download legraphista/openchat-3.6-8b-20240522-IMat-GGUF --include "openchat-3.6-8b-20240522.Q8_0/*" --local-dir openchat-3.6-8b-20240522.Q8_0
# see FAQ for merging GGUF's

Inference

Simple chat template

<|begin_of_text|><|start_header_id|>GPT4 Correct User<|end_header_id|>

Can you provide ways to eat combinations of bananas and dragonfruits?<|eot_id|><|start_header_id|>GPT4 Correct Assistant<|end_header_id|>

Sure! Here are some ways to eat bananas and dragonfruits together:
 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey.
 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey.<|eot_id|><|start_header_id|>GPT4 Correct User<|end_header_id|>

What about solving an 2x + 3 = 7 equation?<|eot_id|>

Chat template with system prompt

<|begin_of_text|><|start_header_id|>System<|end_header_id|>

You are a helpful AI.<|eot_id|><|start_header_id|>GPT4 Correct User<|end_header_id|>

Can you provide ways to eat combinations of bananas and dragonfruits?<|eot_id|><|start_header_id|>GPT4 Correct Assistant<|end_header_id|>

Sure! Here are some ways to eat bananas and dragonfruits together:
 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey.
 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey.<|eot_id|><|start_header_id|>GPT4 Correct User<|end_header_id|>

What about solving an 2x + 3 = 7 equation?<|eot_id|>

Llama.cpp

llama.cpp/main -m openchat-3.6-8b-20240522.Q8_0.gguf --color -i -p "prompt here (according to the chat template)"

📚 Documentation

Files

IMatrix

Status: ✅ Available
Link: here

Common Quants

All Quants

🔧 Technical Details

According to this investigation, it appears that lower quantizations are the only ones that benefit from the IMatrix input (as per hellaswag results).

📄 License

This project uses the llama3 license.

FAQ

Why is the IMatrix not applied everywhere?

According to this investigation, it appears that lower quantizations are the only ones that benefit from the IMatrix input (as per hellaswag results).

How do I merge a split GGUF?

1. Make sure you have gguf-split available
   • To get hold of gguf-split, navigate to https://github.com/ggerganov/llama.cpp/releases
   • Download the appropriate zip for your system from the latest release
   • Unzip the archive and you should be able to find gguf-split
2. Locate your GGUF chunks folder (ex: openchat-3.6-8b-20240522.Q8_0)
3. Run gguf-split --merge openchat-3.6-8b-20240522.Q8_0/openchat-3.6-8b-20240522.Q8_0-00001-of-XXXXX.gguf openchat-3.6-8b-20240522.Q8_0.gguf
   • Make sure to point gguf-split to the first chunk of the split.

Got a suggestion? Ping me @legraphista!