%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Model Overview
Model Features
Model Capabilities
Use Cases
🚀 Llamacpp imatrix Quantizations of Cydonia-24B-v2.1 by TheDrummer
This project provides llama.cpp imatrix quantizations of the Cydonia-24B-v2.1 model, enabling efficient deployment and use on various platforms.
🚀 Quick Start
- Use llama.cpp release b4792 for quantization.
- Original model: https://huggingface.co/TheDrummer/Cydonia-24B-v2.1
- All quants are made using the imatrix option with the dataset from here.
- Run them in LM Studio.
- Run them directly with llama.cpp, or any other llama.cpp based project.
✨ Features
- Multiple quantization types are available to meet different performance and quality requirements.
- Support for running in LM Studio and llama.cpp based projects.
- Online repacking for weights to improve performance on ARM and AVX machines.
📦 Installation
Install huggingface-cli
pip install -U "huggingface_hub[cli]"
Download a specific file
huggingface-cli download bartowski/TheDrummer_Cydonia-24B-v2.1-GGUF --include "TheDrummer_Cydonia-24B-v2.1-Q4_K_M.gguf" --local-dir ./
Download split files
If the model is bigger than 50GB, it will have been split into multiple files. To download them all to a local folder, run:
huggingface-cli download bartowski/TheDrummer_Cydonia-24B-v2.1-GGUF --include "TheDrummer_Cydonia-24B-v2.1-Q8_0/*" --local-dir ./
💻 Usage Examples
Basic Usage
Prompt format
<s>[SYSTEM_PROMPT]{system_prompt}[/SYSTEM_PROMPT][INST]{prompt}[/INST]
📚 Documentation
Download a file
| Filename | Quant type | File Size | Split | Description |
|---|---|---|---|---|
| Cydonia-24B-v2.1-Q8_0.gguf | Q8_0 | 25.05GB | false | Extremely high quality, generally unneeded but max available quant. |
| Cydonia-24B-v2.1-Q6_K_L.gguf | Q6_K_L | 19.67GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended. |
| Cydonia-24B-v2.1-Q6_K.gguf | Q6_K | 19.35GB | false | Very high quality, near perfect, recommended. |
| Cydonia-24B-v2.1-Q5_K_L.gguf | Q5_K_L | 17.18GB | false | Uses Q8_0 for embed and output weights. High quality, recommended. |
| Cydonia-24B-v2.1-Q5_K_M.gguf | Q5_K_M | 16.76GB | false | High quality, recommended. |
| Cydonia-24B-v2.1-Q5_K_S.gguf | Q5_K_S | 16.30GB | false | High quality, recommended. |
| Cydonia-24B-v2.1-Q4_1.gguf | Q4_1 | 14.87GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
| Cydonia-24B-v2.1-Q4_K_L.gguf | Q4_K_L | 14.83GB | false | Uses Q8_0 for embed and output weights. Good quality, recommended. |
| Cydonia-24B-v2.1-Q4_K_M.gguf | Q4_K_M | 14.33GB | false | Good quality, default size for most use cases, recommended. |
| Cydonia-24B-v2.1-Q4_K_S.gguf | Q4_K_S | 13.55GB | false | Slightly lower quality with more space savings, recommended. |
| Cydonia-24B-v2.1-Q4_0.gguf | Q4_0 | 13.49GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
| Cydonia-24B-v2.1-IQ4_NL.gguf | IQ4_NL | 13.47GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
| Cydonia-24B-v2.1-Q3_K_XL.gguf | Q3_K_XL | 12.99GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
| Cydonia-24B-v2.1-IQ4_XS.gguf | IQ4_XS | 12.76GB | false | Decent quality, smaller than Q4_K_S with similar performance, recommended. |
| Cydonia-24B-v2.1-Q3_K_L.gguf | Q3_K_L | 12.40GB | false | Lower quality but usable, good for low RAM availability. |
| Cydonia-24B-v2.1-Q3_K_M.gguf | Q3_K_M | 11.47GB | false | Low quality. |
| Cydonia-24B-v2.1-IQ3_M.gguf | IQ3_M | 10.65GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
| Cydonia-24B-v2.1-Q3_K_S.gguf | Q3_K_S | 10.40GB | false | Low quality, not recommended. |
| Cydonia-24B-v2.1-IQ3_XS.gguf | IQ3_XS | 9.91GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
| Cydonia-24B-v2.1-Q2_K_L.gguf | Q2_K_L | 9.55GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
| Cydonia-24B-v2.1-IQ3_XXS.gguf | IQ3_XXS | 9.28GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
| Cydonia-24B-v2.1-Q2_K.gguf | Q2_K | 8.89GB | false | Very low quality but surprisingly usable. |
| Cydonia-24B-v2.1-IQ2_M.gguf | IQ2_M | 8.11GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
| Cydonia-24B-v2.1-IQ2_S.gguf | IQ2_S | 7.48GB | false | Low quality, uses SOTA techniques to be usable. |
| Cydonia-24B-v2.1-IQ2_XS.gguf | IQ2_XS | 7.21GB | false | Low quality, uses SOTA techniques to be 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.
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, 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 increase.
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:
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.
🔧 Technical Details
- Quantization is performed using llama.cpp release b4792.
- The imatrix option is used for quantization, with the calibration dataset from here.
📄 License
This project is licensed under the other license.
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.
- Thank you to LM Studio for sponsoring my work.
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
Transformers Transformers Supports Multiple Languages