bartowski's picture
Update metadata with huggingface_hub
34005c6 verified
metadata
base_model: h2oai/h2o-danube3-500m-chat
language:
  - en
library_name: transformers
license: apache-2.0
pipeline_tag: text-generation
tags:
  - gpt
  - llm
  - large language model
  - h2o-llmstudio
quantized_by: bartowski
thumbnail: >-
  https://h2o.ai/etc.clientlibs/h2o/clientlibs/clientlib-site/resources/images/favicon.ico

Llamacpp imatrix Quantizations of h2o-danube3-500m-chat

Using llama.cpp release b3389 for quantization.

Original model: https://huggingface.co/h2oai/h2o-danube3-500m-chat

All quants made using imatrix option with dataset from here

Prompt format

<|prompt|>{prompt}</s><|answer|></s><|answer|>

Note that this model does not support a System prompt.

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

Filename Quant type File Size Split Description
h2o-danube3-500m-chat-f32.gguf f32 2.06GB false Full F32 weights.
h2o-danube3-500m-chat-Q8_0.gguf Q8_0 0.55GB false Extremely high quality, generally unneeded but max available quant.
h2o-danube3-500m-chat-Q6_K_L.gguf Q6_K_L 0.45GB false Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
h2o-danube3-500m-chat-Q6_K.gguf Q6_K 0.42GB false Very high quality, near perfect, recommended.
h2o-danube3-500m-chat-Q5_K_L.gguf Q5_K_L 0.40GB false Uses Q8_0 for embed and output weights. High quality, recommended.
h2o-danube3-500m-chat-Q5_K_M.gguf Q5_K_M 0.37GB false High quality, recommended.
h2o-danube3-500m-chat-Q5_K_S.gguf Q5_K_S 0.36GB false High quality, recommended.
h2o-danube3-500m-chat-Q4_K_L.gguf Q4_K_L 0.35GB false Uses Q8_0 for embed and output weights. Good quality, recommended.
h2o-danube3-500m-chat-Q4_K_M.gguf Q4_K_M 0.32GB false Good quality, default size for must use cases, recommended.
h2o-danube3-500m-chat-Q3_K_XL.gguf Q3_K_XL 0.32GB false Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
h2o-danube3-500m-chat-Q4_K_S.gguf Q4_K_S 0.30GB false Slightly lower quality with more space savings, recommended.
h2o-danube3-500m-chat-IQ4_XS.gguf IQ4_XS 0.29GB false Decent quality, smaller than Q4_K_S with similar performance, recommended.
h2o-danube3-500m-chat-Q3_K_L.gguf Q3_K_L 0.28GB false Lower quality but usable, good for low RAM availability.
h2o-danube3-500m-chat-IQ3_M.gguf IQ3_M 0.25GB 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/h2o-danube3-500m-chat-GGUF --include "h2o-danube3-500m-chat-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/h2o-danube3-500m-chat-GGUF --include "h2o-danube3-500m-chat-Q8_0.gguf/*" --local-dir h2o-danube3-500m-chat-Q8_0

You can either specify a new local-dir (h2o-danube3-500m-chat-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