bartowski's picture
Llamacpp quants
871d03b verified
metadata
license: apache-2.0
pipeline_tag: text-generation
inference: true
quantized_by: bartowski

Llamacpp Quantizations of Mistral-7B-v0.2-hf

Using llama.cpp release b2440 for quantization.

Original model: https://huggingface.co/alpindale/Mistral-7B-v0.2-hf/

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

Filename Quant type File Size Description
Mistral-7B-v0.2-hf-Q8_0.gguf Q8_0 7.69GB Extremely high quality, generally unneeded but max available quant.
Mistral-7B-v0.2-hf-Q6_K.gguf Q6_K 5.94GB Very high quality, near perfect, recommended.
Mistral-7B-v0.2-hf-Q5_K_M.gguf Q5_K_M 5.13GB High quality, very usable.
Mistral-7B-v0.2-hf-Q5_K_S.gguf Q5_K_S 4.99GB High quality, very usable.
Mistral-7B-v0.2-hf-Q5_0.gguf Q5_0 4.99GB High quality, older format, generally not recommended.
Mistral-7B-v0.2-hf-Q4_K_M.gguf Q4_K_M 4.36GB Good quality, similar to 4.25 bpw.
Mistral-7B-v0.2-hf-Q4_K_S.gguf Q4_K_S 4.14GB Slightly lower quality with small space savings.
Mistral-7B-v0.2-hf-IQ4_NL.gguf IQ4_NL 4.15GB Good quality, similar to Q4_K_S, new method of quanting,
Mistral-7B-v0.2-hf-IQ4_XS.gguf IQ4_XS 3.94GB Decent quality, new method with similar performance to Q4.
Mistral-7B-v0.2-hf-Q4_0.gguf Q4_0 4.10GB Decent quality, older format, generally not recommended.
Mistral-7B-v0.2-hf-IQ3_M.gguf IQ3_M 3.28GB Medium-low quality, new method with decent performance.
Mistral-7B-v0.2-hf-IQ3_S.gguf IQ3_S 3.18GB Lower quality, new method with decent performance, recommended over Q3 quants.
Mistral-7B-v0.2-hf-Q3_K_L.gguf Q3_K_L 3.82GB Lower quality but usable, good for low RAM availability.
Mistral-7B-v0.2-hf-Q3_K_M.gguf Q3_K_M 3.51GB Even lower quality.
Mistral-7B-v0.2-hf-Q3_K_S.gguf Q3_K_S 3.16GB Low quality, not recommended.
Mistral-7B-v0.2-hf-Q2_K.gguf Q2_K 2.71GB Extremely low quality, not recommended.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski