bartowski/Hercules-4.0-Yi-34B-GGUF

Llamacpp Quantizations of Hercules-4.0-Yi-34B

Using llama.cpp release b2589 for quantization.

Original model: https://huggingface.co/Locutusque/Hercules-4.0-Yi-34B

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

Filename	Quant type	File Size	Description
Hercules-4.0-Yi-34B-Q8_0.gguf	Q8_0	36.54GB	Extremely high quality, generally unneeded but max available quant.
Hercules-4.0-Yi-34B-Q6_K.gguf	Q6_K	28.21GB	Very high quality, near perfect, recommended.
Hercules-4.0-Yi-34B-Q5_K_M.gguf	Q5_K_M	24.32GB	High quality, recommended.
Hercules-4.0-Yi-34B-Q5_K_S.gguf	Q5_K_S	23.70GB	High quality, recommended.
Hercules-4.0-Yi-34B-Q5_0.gguf	Q5_0	23.70GB	High quality, older format, generally not recommended.
Hercules-4.0-Yi-34B-Q4_K_M.gguf	Q4_K_M	20.65GB	Good quality, uses about 4.83 bits per weight, recommended.
Hercules-4.0-Yi-34B-Q4_K_S.gguf	Q4_K_S	19.59GB	Slightly lower quality with small space savings.
Hercules-4.0-Yi-34B-IQ4_NL.gguf	IQ4_NL	19.65GB	Decent quality, similar to Q4_K_S, new method of quanting, recommended.
Hercules-4.0-Yi-34B-IQ4_XS.gguf	IQ4_XS	18.63GB	Decent quality, new method with similar performance to Q4.
Hercules-4.0-Yi-34B-Q4_0.gguf	Q4_0	19.46GB	Decent quality, older format, generally not recommended.
Hercules-4.0-Yi-34B-Q3_K_L.gguf	Q3_K_L	18.13GB	Lower quality but usable, good for low RAM availability.
Hercules-4.0-Yi-34B-Q3_K_M.gguf	Q3_K_M	16.65GB	Even lower quality.
Hercules-4.0-Yi-34B-IQ3_M.gguf	IQ3_M	15.56GB	Medium-low quality, new method with decent performance.
Hercules-4.0-Yi-34B-IQ3_S.gguf	IQ3_S	15.01GB	Lower quality, new method with decent performance, recommended over Q3 quants.
Hercules-4.0-Yi-34B-Q3_K_S.gguf	Q3_K_S	14.96GB	Low quality, not recommended.
Hercules-4.0-Yi-34B-Q2_K.gguf	Q2_K	12.82GB	Extremely low quality, not recommended.

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