Llamacpp imatrix Quantizations of Lite-Mistral-150M-v2-Instruct

Using llama.cpp release b3389 for quantization.

Original model: https://huggingface.co/OuteAI/Lite-Mistral-150M-v2-Instruct

All quants made using imatrix option with dataset from here

Prompt format

<s> system
{system_prompt}</s> 
<s> user
{prompt}</s> 
<s> assistant

Download a file (not the whole branch) from below:

Filename Quant type File Size Split Description
Lite-Mistral-150M-v2-Instruct-f32.gguf f32 0.63GB false Full F32 weights.
Lite-Mistral-150M-v2-Instruct-Q8_0.gguf Q8_0 0.17GB false Extremely high quality, generally unneeded but max available quant.
Lite-Mistral-150M-v2-Instruct-Q6_K_L.gguf Q6_K_L 0.14GB false Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Lite-Mistral-150M-v2-Instruct-Q6_K.gguf Q6_K 0.13GB false Very high quality, near perfect, recommended.
Lite-Mistral-150M-v2-Instruct-Q5_K_L.gguf Q5_K_L 0.13GB false Uses Q8_0 for embed and output weights. High quality, recommended.
Lite-Mistral-150M-v2-Instruct-Q4_K_L.gguf Q4_K_L 0.12GB false Uses Q8_0 for embed and output weights. Good quality, recommended.
Lite-Mistral-150M-v2-Instruct-Q5_K_M.gguf Q5_K_M 0.11GB false High quality, recommended.
Lite-Mistral-150M-v2-Instruct-Q3_K_XL.gguf Q3_K_XL 0.11GB false Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
Lite-Mistral-150M-v2-Instruct-Q4_K_M.gguf Q4_K_M 0.10GB false Good quality, default size for must use cases, recommended.
Lite-Mistral-150M-v2-Instruct-Q3_K_L.gguf Q3_K_L 0.09GB false Lower quality but usable, good for low RAM availability.
Lite-Mistral-150M-v2-Instruct-IQ4_XS.gguf IQ4_XS 0.09GB false Decent quality, smaller than Q4_K_S with similar performance, recommended.
Lite-Mistral-150M-v2-Instruct-IQ3_M.gguf IQ3_M 0.08GB false Medium-low quality, new method with decent performance comparable to Q3_K_M.

Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset

Thank you ZeroWw for the inspiration to experiment with embed/output

Downloading using huggingface-cli

First, make sure you have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Then, you can target the specific file you want:

huggingface-cli download bartowski/Lite-Mistral-150M-v2-Instruct-GGUF --include "Lite-Mistral-150M-v2-Instruct-Q4_K_M.gguf" --local-dir ./

If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download bartowski/Lite-Mistral-150M-v2-Instruct-GGUF --include "Lite-Mistral-150M-v2-Instruct-Q8_0.gguf/*" --local-dir Lite-Mistral-150M-v2-Instruct-Q8_0

You can either specify a new local-dir (Lite-Mistral-150M-v2-Instruct-Q8_0) or download them all in place (./)

Which file should I choose?

A great write up with charts showing various performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

The I-quants are not compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

Downloads last month
317
GGUF
Model size
157M params
Architecture
llama

3-bit

4-bit

5-bit

6-bit

8-bit

32-bit

Inference Examples
Unable to determine this model's library. Check the docs .

Model tree for bartowski/Lite-Mistral-150M-v2-Instruct-GGUF

Quantized
(5)
this model