VOOZH about

URL: https://huggingface.co/bartowski/goppa-ai_Goppa-LogiLlama-GGUF

⇱ bartowski/goppa-ai_Goppa-LogiLlama-GGUF · Hugging Face

Llamacpp imatrix Quantizations of Goppa-LogiLlama by goppa-ai

Using llama.cpp release b4792 for quantization.

Original model: https://huggingface.co/goppa-ai/Goppa-LogiLlama

All quants made using imatrix option with dataset from here

Run them in LM Studio

Run them directly with llama.cpp, or any other llama.cpp based project

Prompt format 

<|begin_of_text|><|start_header_id|>system<|end_header_id|>

Cutting Knowledge Date: December 2023
Today Date: 06 Mar 2025

{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>

{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>

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

Filename Quant type File Size Split Description
Goppa-LogiLlama-f16.gguf f16 2.48GB false Full F16 weights.
Goppa-LogiLlama-Q8_0.gguf Q8_0 1.32GB false Extremely high quality, generally unneeded but max available quant.
Goppa-LogiLlama-Q6_K_L.gguf Q6_K_L 1.09GB false Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
Goppa-LogiLlama-Q6_K.gguf Q6_K 1.02GB false Very high quality, near perfect, recommended.
Goppa-LogiLlama-Q5_K_L.gguf Q5_K_L 0.98GB false Uses Q8_0 for embed and output weights. High quality, recommended.
Goppa-LogiLlama-Q5_K_M.gguf Q5_K_M 0.91GB false High quality, recommended.
Goppa-LogiLlama-Q5_K_S.gguf Q5_K_S 0.89GB false High quality, recommended.
Goppa-LogiLlama-Q4_K_L.gguf Q4_K_L 0.87GB false Uses Q8_0 for embed and output weights. Good quality, recommended.
Goppa-LogiLlama-Q4_1.gguf Q4_1 0.83GB false Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
Goppa-LogiLlama-Q4_K_M.gguf Q4_K_M 0.81GB false Good quality, default size for most use cases, recommended.
Goppa-LogiLlama-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.
Goppa-LogiLlama-Q4_K_S.gguf Q4_K_S 0.78GB false Slightly lower quality with more space savings, recommended.
Goppa-LogiLlama-Q4_0.gguf Q4_0 0.77GB false Legacy format, offers online repacking for ARM and AVX CPU inference.
Goppa-LogiLlama-IQ4_NL.gguf IQ4_NL 0.77GB false Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference.
Goppa-LogiLlama-IQ4_XS.gguf IQ4_XS 0.74GB false Decent quality, smaller than Q4_K_S with similar performance, recommended.
Goppa-LogiLlama-Q3_K_L.gguf Q3_K_L 0.73GB false Lower quality but usable, good for low RAM availability.
Goppa-LogiLlama-Q3_K_M.gguf Q3_K_M 0.69GB false Low quality.
Goppa-LogiLlama-IQ3_M.gguf IQ3_M 0.66GB false Medium-low quality, new method with decent performance comparable to Q3_K_M.
Goppa-LogiLlama-Q3_K_S.gguf Q3_K_S 0.64GB false Low quality, not recommended.
Goppa-LogiLlama-Q2_K_L.gguf Q2_K_L 0.64GB false Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable.
Goppa-LogiLlama-IQ3_XS.gguf IQ3_XS 0.62GB false Lower quality, new method with decent performance, slightly better than Q3_K_S.
Goppa-LogiLlama-Q2_K.gguf Q2_K 0.58GB false Very low quality but surprisingly usable.
Goppa-LogiLlama-IQ3_XXS.gguf IQ3_XXS 0.56GB false Lower quality, new method with decent performance, comparable to Q3 quants.
Goppa-LogiLlama-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

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.

Which file should I choose?

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

Downloads last month
1,639
GGUF
Model size
1B params
Architecture
llama
Hardware compatibility
Log In to add your hardware

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Model tree for bartowski/goppa-ai_Goppa-LogiLlama-GGUF

Quantized
(11)
this model

Dataset used to train bartowski/goppa-ai_Goppa-LogiLlama-GGUF