Dolphin3.0-Llama3.1-8B-GGUF Open-source Model - Efficient Operation in Resource-constrained Environments, Multiple Quantization Types Available

Dolphin3.0 Llama3.1 8B GGUF

Developed by bartowski

A quantized version of the Dolphin3.0-Llama3.1-8B model, which supports efficient operation in resource-constrained environments and provides multiple quantization types to choose from.

Large Language Model English#Efficient Quantitative Model #Multi-task Text Generation #Low-resource Deployment

Downloads 947

Release Time : 1/5/2025

Model Overview

This project provides a quantized version of the Dolphin3.0-Llama3.1-8B model, which solves the problem of efficiently running large models in resource-constrained environments and enables users to perform text generation tasks with less resource consumption.

Model Features

Multiple Quantization Types

It provides a rich variety of quantization types for users to choose from to meet different performance and quality requirements.

Online Repackaging

Some quantization types support online repackaging, which can automatically optimize hardware performance.

Detailed Documentation

It provides detailed instructions on downloading, using, and selecting quantized models.

Model Capabilities

Text Generation

Efficient Inference

Resource Optimization

Use Cases

Text Generation

Efficient Text Generation

Efficiently generate text in resource-constrained environments

Complete text generation tasks with less resource consumption

🚀 Llamacpp imatrix Quantizations of Dolphin3.0-Llama3.1-8B

This project provides quantized versions of the Dolphin3.0-Llama3.1-8B model using the llama.cpp library. It offers various quantization types to balance between model quality and resource usage, allowing users to run the model on different hardware configurations.

Key Information

Property	Details
Quantized By	bartowski
Pipeline Tag	text-generation
Language	en
Base Model	cognitivecomputations/Dolphin3.0-Llama3.1-8B
License	llama3.1
Training Data	OpenCoder-LLM/opc-sft-stage1, OpenCoder-LLM/opc-sft-stage2, microsoft/orca-agentinstruct-1M-v1, microsoft/orca-math-word-problems-200k, NousResearch/hermes-function-calling-v1, AI-MO/NuminaMath-CoT, AI-MO/NuminaMath-TIR, allenai/tulu-3-sft-mixture, cognitivecomputations/dolphin-coder, HuggingFaceTB/smoltalk, cognitivecomputations/samantha-data, m-a-p/CodeFeedback-Filtered-Instruction, m-a-p/Code-Feedback

🚀 Quick Start

Use llama.cpp release b4418 for quantization.
The original model can be found at https://huggingface.co/cognitivecomputations/Dolphin3.0-Llama3.1-8B.
All quantizations are made using the imatrix option with the dataset from here.
You can run the quantized models in LM Studio.

Prompt format

<|im_start|>system
{system_prompt}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

📦 Download

You can download a specific file (not the whole branch) from the following table:

Filename	Quant type	File Size	Split	Description
Dolphin3.0-Llama3.1-8B-f32.gguf	f32	32.13GB	false	Full F32 weights.
Dolphin3.0-Llama3.1-8B-f16.gguf	f16	16.07GB	false	Full F16 weights.
Dolphin3.0-Llama3.1-8B-Q8_0.gguf	Q8_0	8.54GB	false	Extremely high quality, generally unneeded but max available quant.
Dolphin3.0-Llama3.1-8B-Q6_K_L.gguf	Q6_K_L	6.85GB	false	Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Dolphin3.0-Llama3.1-8B-Q6_K.gguf	Q6_K	6.60GB	false	Very high quality, near perfect, recommended.
Dolphin3.0-Llama3.1-8B-Q5_K_L.gguf	Q5_K_L	6.06GB	false	Uses Q8_0 for embed and output weights. High quality, recommended.
Dolphin3.0-Llama3.1-8B-Q5_K_M.gguf	Q5_K_M	5.73GB	false	High quality, recommended.
Dolphin3.0-Llama3.1-8B-Q5_K_S.gguf	Q5_K_S	5.60GB	false	High quality, recommended.
Dolphin3.0-Llama3.1-8B-Q4_K_L.gguf	Q4_K_L	5.31GB	false	Uses Q8_0 for embed and output weights. Good quality, recommended.
Dolphin3.0-Llama3.1-8B-Q4_1.gguf	Q4_1	5.13GB	false	Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
Dolphin3.0-Llama3.1-8B-Q4_K_M.gguf	Q4_K_M	4.92GB	false	Good quality, default size for most use cases, recommended.
Dolphin3.0-Llama3.1-8B-Q3_K_XL.gguf	Q3_K_XL	4.78GB	false	Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
Dolphin3.0-Llama3.1-8B-Q4_K_S.gguf	Q4_K_S	4.69GB	false	Slightly lower quality with more space savings, recommended.
Dolphin3.0-Llama3.1-8B-Q4_0.gguf	Q4_0	4.68GB	false	Legacy format, offers online repacking for ARM and AVX CPU inference.
Dolphin3.0-Llama3.1-8B-IQ4_NL.gguf	IQ4_NL	4.68GB	false	Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference.
Dolphin3.0-Llama3.1-8B-IQ4_XS.gguf	IQ4_XS	4.45GB	false	Decent quality, smaller than Q4_K_S with similar performance, recommended.
Dolphin3.0-Llama3.1-8B-Q3_K_L.gguf	Q3_K_L	4.32GB	false	Lower quality but usable, good for low RAM availability.
Dolphin3.0-Llama3.1-8B-Q3_K_M.gguf	Q3_K_M	4.02GB	false	Low quality.
Dolphin3.0-Llama3.1-8B-IQ3_M.gguf	IQ3_M	3.78GB	false	Medium-low quality, new method with decent performance comparable to Q3_K_M.
Dolphin3.0-Llama3.1-8B-Q2_K_L.gguf	Q2_K_L	3.69GB	false	Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
Dolphin3.0-Llama3.1-8B-Q3_K_S.gguf	Q3_K_S	3.66GB	false	Low quality, not recommended.
Dolphin3.0-Llama3.1-8B-IQ3_XS.gguf	IQ3_XS	3.52GB	false	Lower quality, new method with decent performance, slightly better than Q3_K_S.
Dolphin3.0-Llama3.1-8B-Q2_K.gguf	Q2_K	3.18GB	false	Very low quality but surprisingly usable.
Dolphin3.0-Llama3.1-8B-IQ2_M.gguf	IQ2_M	2.95GB	false	Relatively low quality, uses SOTA techniques to be surprisingly usable.

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

Downloading using huggingface-cli

Click to view download instructions

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/Dolphin3.0-Llama3.1-8B-GGUF --include "Dolphin3.0-Llama3.1-8B-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/Dolphin3.0-Llama3.1-8B-GGUF --include "Dolphin3.0-Llama3.1-8B-Q8_0/*" --local-dir ./

You can either specify a new local-dir (Dolphin3.0-Llama3.1-8B-Q8_0) or download them all in place (./)

ARM/AVX information

Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.

Now, however, there is something called "online repacking" for weights. details in this PR. If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.

As of llama.cpp build b4282 you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.

Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to this PR which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.

Click to view Q4_0_X_X information (deprecated)

I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.

Click to view benchmarks on an AVX2 system (EPYC7702)

model	size	params	backend	threads	test	t/s	% (vs Q4_0)
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	pp512	204.03 ± 1.03	100%
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	pp1024	282.92 ± 0.19	100%
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	pp2048	259.49 ± 0.44	100%
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	tg128	39.12 ± 0.27	100%
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	tg256	39.31 ± 0.69	100%
qwen2 3B Q4_0	1.70 GiB	3.09 B	CPU	64	tg512	40.52 ± 0.03	100%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	pp512	301.02 ± 1.74	147%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	pp1024	287.23 ± 0.20	101%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	pp2048	262.77 ± 1.81	101%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	tg128	18.80 ± 0.99	48%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	tg256	24.46 ± 3.04	83%
qwen2 3B Q4_K_M	1.79 GiB	3.09 B	CPU	64	tg512	36.32 ± 3.59	90%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	pp512	271.71 ± 3.53	133%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	pp1024	279.86 ± 45.63	100%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	pp2048	320.77 ± 5.00	124%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	tg128	43.51 ± 0.05	111%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	tg256	43.35 ± 0.09	110%
qwen2 3B Q4_0_8_8	1.69 GiB	3.09 B	CPU	64	tg512	42.60 ± 0.31	105%

Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation

Which file should I choose?

Click here for details

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.

🔧 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.

If you want to support the author's work, visit the ko-fi page here: https://ko-fi.com/bartowski

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご