Dolphin3.0-Llama3.2-1B-GGUFオープンソースモデル - 複数量化版でテキスト生成タスクに対応

Dolphin3.0 Llama3.2 1B GGUF

bartowskiによって開発

Llama3.2アーキテクチャに基づく1Bパラメータ量子化モデル、テキスト生成タスクをサポート、様々な量子化バージョン選択に対応

大規模言語モデル英語#マルチタスクテキスト生成 #効率的な量子化デプロイ #コードと数学的推論

ダウンロード数 1,134

リリース時間 : 1/5/2025

モデル概要

これは量子化されたテキスト生成モデルで、cognitivecomputations/Dolphin3.0-Llama3.2-1Bモデルを基に、llama.cppツールを使用して量子化処理されています。モデルは複数の量子化レベルをサポートし、異なるハードウェア環境に適しています。

モデル特徴

複数量子化バージョン

F32からQ2_Kまでの複数量子化バージョンを提供し、異なるハードウェアと性能ニーズに対応

imatrix量子化

llama.cppのimatrixオプションを使用した量子化で、量子化品質を向上

ARM/AVX最適化

ARMおよびAVX CPUのオンライン再パッケージ機能をサポートし、推論性能を最適化

埋め込み/出力重み最適化

一部の量子化バージョンでは埋め込みと出力重みにQ8_0量子化を使用し、モデル品質を向上

モデル能力

テキスト生成

インストラクション追従

コード生成

数学問題解答

使用事例

プログラミング支援

コード生成

自然言語記述に基づいてコードスニペットを生成

コードフィードバック

コード改善の提案とフィードバックを提供

教育

数学問題解答

数学の応用問題や計算問題を解答

汎用AIアシスタント

対話インタラクション

インテリジェントアシスタントとして自然言語対話を行う

quantized_by: bartowski pipeline_tag: text-generation language:

en base_model: cognitivecomputations/Dolphin3.0-Llama3.2-1B license: llama3.2 datasets:
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

Llamacpp imatrix Quantizations of Dolphin3.0-Llama3.2-1B

Using llama.cpp release b4418 for quantization.

Original model: https://huggingface.co/cognitivecomputations/Dolphin3.0-Llama3.2-1B

All quants made using imatrix option with dataset from here

Run them in LM Studio

Prompt format

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

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

Filename	Quant type	File Size	Split	Description
Dolphin3.0-Llama3.2-1B-f32.gguf	f32	4.95GB	false	Full F32 weights.
Dolphin3.0-Llama3.2-1B-f16.gguf	f16	2.48GB	false	Full F16 weights.
Dolphin3.0-Llama3.2-1B-Q8_0.gguf	Q8_0	1.32GB	false	Extremely high quality, generally unneeded but max available quant.
Dolphin3.0-Llama3.2-1B-Q6_K_L.gguf	Q6_K_L	1.09GB	false	Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Dolphin3.0-Llama3.2-1B-Q6_K.gguf	Q6_K	1.02GB	false	Very high quality, near perfect, recommended.
Dolphin3.0-Llama3.2-1B-Q5_K_L.gguf	Q5_K_L	0.98GB	false	Uses Q8_0 for embed and output weights. High quality, recommended.
Dolphin3.0-Llama3.2-1B-Q5_K_M.gguf	Q5_K_M	0.91GB	false	High quality, recommended.
Dolphin3.0-Llama3.2-1B-Q5_K_S.gguf	Q5_K_S	0.89GB	false	High quality, recommended.
Dolphin3.0-Llama3.2-1B-Q4_K_L.gguf	Q4_K_L	0.87GB	false	Uses Q8_0 for embed and output weights. Good quality, recommended.
Dolphin3.0-Llama3.2-1B-Q4_1.gguf	Q4_1	0.83GB	false	Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
Dolphin3.0-Llama3.2-1B-Q4_K_M.gguf	Q4_K_M	0.81GB	false	Good quality, default size for most use cases, recommended.
Dolphin3.0-Llama3.2-1B-Q3_K_XL.gguf	Q3_K_XL	0.80GB	false	Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
Dolphin3.0-Llama3.2-1B-Q4_K_S.gguf	Q4_K_S	0.78GB	false	Slightly lower quality with more space savings, recommended.
Dolphin3.0-Llama3.2-1B-Q4_0.gguf	Q4_0	0.77GB	false	Legacy format, offers online repacking for ARM and AVX CPU inference.
Dolphin3.0-Llama3.2-1B-IQ4_NL.gguf	IQ4_NL	0.77GB	false	Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference.
Dolphin3.0-Llama3.2-1B-IQ4_XS.gguf	IQ4_XS	0.74GB	false	Decent quality, smaller than Q4_K_S with similar performance, recommended.
Dolphin3.0-Llama3.2-1B-Q3_K_L.gguf	Q3_K_L	0.73GB	false	Lower quality but usable, good for low RAM availability.
Dolphin3.0-Llama3.2-1B-Q3_K_M.gguf	Q3_K_M	0.69GB	false	Low quality.
Dolphin3.0-Llama3.2-1B-IQ3_M.gguf	IQ3_M	0.66GB	false	Medium-low quality, new method with decent performance comparable to Q3_K_M.
Dolphin3.0-Llama3.2-1B-Q3_K_S.gguf	Q3_K_S	0.64GB	false	Low quality, not recommended.
Dolphin3.0-Llama3.2-1B-Q2_K_L.gguf	Q2_K_L	0.64GB	false	Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
Dolphin3.0-Llama3.2-1B-IQ3_XS.gguf	IQ3_XS	0.62GB	false	Lower quality, new method with decent performance, slightly better than Q3_K_S.
Dolphin3.0-Llama3.2-1B-Q2_K.gguf	Q2_K	0.58GB	false	Very low quality but surprisingly usable.
Dolphin3.0-Llama3.2-1B-IQ2_M.gguf	IQ2_M	0.52GB	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-Llama3.2-1B-GGUF --include "Dolphin3.0-Llama3.2-1B-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-Llama3.2-1B-GGUF --include "Dolphin3.0-Llama3.2-1B-Q8_0/*" --local-dir ./

You can either specify a new local-dir (Dolphin3.0-Llama3.2-1B-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.