Aura-4B-GGUFオープンソースモデル - 無料でのデプロイでテキスト生成タスクをサポート、複数タイプの量子化に対応

ホーム

Aura 4B GGUF

bartowskiによって開発

Aura-4BはAuraIndustries/Aura-4Bをベースにした量子化バージョンで、llama.cppを使用してimatrix量子化を行い、複数の量子化タイプをサポートし、テキスト生成タスクに適しています。

大規模言語モデル英語オープンソースライセンス:Apache-2.0 #マルチ量子化バージョン #低メモリ最適化 #テキスト生成

ダウンロード数 290

リリース時間 : 12/23/2024

モデル概要

このモデルは4Bパラメータのテキスト生成モデルで、量子化処理後、リソースが限られたデバイスでも動作し、高い生成品質を維持できます。

モデル特徴

複数量子化オプション

f16からQ2_Kまでの複数量子化バージョンを提供し、さまざまなハードウェアと性能ニーズに対応します。

高品質テキスト生成

量子化後もモデルは高いテキスト生成品質を維持し、特にQ6_K_LやQ5_K_Lなどのバージョンが優れています。

ARM/AVXデバイスサポート

オンライン再パッケージング技術により、ARMおよびAVXデバイスでの実行性能を最適化します。

モデル能力

テキスト生成

指示追従

ロールプレイ

使用事例

テキスト生成

クリエイティブライティング

ストーリーや詩などのクリエイティブなテキストを生成します。

高品質なクリエイティブテキスト出力。

指示追従

ユーザーの指示に基づいて対応するテキスト応答を生成します。

ユーザーの指示を正確に理解し実行します。

ロールプレイ

会話シミュレーション

特定のキャラクターの会話スタイルと内容をシミュレートします。

リアルなキャラクター会話体験。

quantized_by: bartowski pipeline_tag: text-generation language:

en datasets:
Mielikki/Erebus-87k
FourOhFour/Instruct_Phase
FourOhFour/RP_Phase
anthracite-core/full-opus-chosen-hermes-rejected-kto-v1 base_model: AuraIndustries/Aura-4B license: apache-2.0

Llamacpp imatrix Quantizations of Aura-4B

Using llama.cpp release b4381 for quantization.

Original model: https://huggingface.co/AuraIndustries/Aura-4B

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
Aura-4B-f16.gguf	f16	9.03GB	false	Full F16 weights.
Aura-4B-Q8_0.gguf	Q8_0	4.80GB	false	Extremely high quality, generally unneeded but max available quant.
Aura-4B-Q6_K_L.gguf	Q6_K_L	3.90GB	false	Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Aura-4B-Q6_K.gguf	Q6_K	3.71GB	false	Very high quality, near perfect, recommended.
Aura-4B-Q5_K_L.gguf	Q5_K_L	3.47GB	false	Uses Q8_0 for embed and output weights. High quality, recommended.
Aura-4B-Q5_K_M.gguf	Q5_K_M	3.23GB	false	High quality, recommended.
Aura-4B-Q5_K_S.gguf	Q5_K_S	3.16GB	false	High quality, recommended.
Aura-4B-Q4_K_L.gguf	Q4_K_L	3.07GB	false	Uses Q8_0 for embed and output weights. Good quality, recommended.
Aura-4B-Q4_1.gguf	Q4_1	2.91GB	false	Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
Aura-4B-Q3_K_XL.gguf	Q3_K_XL	2.81GB	false	Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability.
Aura-4B-Q4_K_M.gguf	Q4_K_M	2.78GB	false	Good quality, default size for most use cases, recommended.
Aura-4B-Q4_K_S.gguf	Q4_K_S	2.66GB	false	Slightly lower quality with more space savings, recommended.
Aura-4B-Q4_0.gguf	Q4_0	2.66GB	false	Legacy format, offers online repacking for ARM and AVX CPU inference.
Aura-4B-IQ4_NL.gguf	IQ4_NL	2.66GB	false	Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference.
Aura-4B-IQ4_XS.gguf	IQ4_XS	2.54GB	false	Decent quality, smaller than Q4_K_S with similar performance, recommended.
Aura-4B-Q3_K_L.gguf	Q3_K_L	2.46GB	false	Lower quality but usable, good for low RAM availability.
Aura-4B-Q3_K_M.gguf	Q3_K_M	2.30GB	false	Low quality.
Aura-4B-Q2_K_L.gguf	Q2_K_L	2.22GB	false	Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
Aura-4B-IQ3_M.gguf	IQ3_M	2.18GB	false	Medium-low quality, new method with decent performance comparable to Q3_K_M.
Aura-4B-Q3_K_S.gguf	Q3_K_S	2.10GB	false	Low quality, not recommended.
Aura-4B-IQ3_XS.gguf	IQ3_XS	2.03GB	false	Lower quality, new method with decent performance, slightly better than Q3_K_S.
Aura-4B-Q2_K.gguf	Q2_K	1.84GB	false	Very low quality but surprisingly usable.
Aura-4B-IQ2_M.gguf	IQ2_M	1.72GB	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/Aura-4B-GGUF --include "Aura-4B-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/Aura-4B-GGUF --include "Aura-4B-Q8_0/*" --local-dir ./

You can either specify a new local-dir (Aura-4B-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.