RichardErkhov/double7_-_vicuna-160m-gguf

Quantization made by Richard Erkhov.

vicuna-160m - GGUF

Model creator: https://huggingface.co/double7/
Original model: https://huggingface.co/double7/vicuna-160m/

Name	Quant method	Size
vicuna-160m.Q2_K.gguf	Q2_K	0.07GB
vicuna-160m.IQ3_XS.gguf	IQ3_XS	0.07GB
vicuna-160m.IQ3_S.gguf	IQ3_S	0.07GB
vicuna-160m.Q3_K_S.gguf	Q3_K_S	0.07GB
vicuna-160m.IQ3_M.gguf	IQ3_M	0.08GB
vicuna-160m.Q3_K.gguf	Q3_K	0.08GB
vicuna-160m.Q3_K_M.gguf	Q3_K_M	0.08GB
vicuna-160m.Q3_K_L.gguf	Q3_K_L	0.08GB
vicuna-160m.IQ4_XS.gguf	IQ4_XS	0.09GB
vicuna-160m.Q4_0.gguf	Q4_0	0.09GB
vicuna-160m.IQ4_NL.gguf	IQ4_NL	0.09GB
vicuna-160m.Q4_K_S.gguf	Q4_K_S	0.09GB
vicuna-160m.Q4_K.gguf	Q4_K	0.1GB
vicuna-160m.Q4_K_M.gguf	Q4_K_M	0.1GB
vicuna-160m.Q4_1.gguf	Q4_1	0.1GB
vicuna-160m.Q5_0.gguf	Q5_0	0.11GB
vicuna-160m.Q5_K_S.gguf	Q5_K_S	0.11GB
vicuna-160m.Q5_K.gguf	Q5_K	0.11GB
vicuna-160m.Q5_K_M.gguf	Q5_K_M	0.11GB
vicuna-160m.Q5_1.gguf	Q5_1	0.12GB
vicuna-160m.Q6_K.gguf	Q6_K	0.12GB
vicuna-160m.Q8_0.gguf	Q8_0	0.16GB

Original model description:

license: apache-2.0 datasets: - anon8231489123/ShareGPT_Vicuna_unfiltered language: - en pipeline_tag: text-generation

Model description

This is a Vicuna-like model with only 160M parameters, which is fine-tuned from LLaMA-160m on ShareGPT data.

The training setup follows the Vicuna suite.

The model is mainly developed as a base Small Speculative Model in MCSD paper. As a comparison, it can be better aligned to the Vicuna models than LLaMA-160m with little loss of alignment to the LLaMA models.

Draft Model	Target Model	Alignment
LLaMA-68/160M	LLaMA-13/33B	😃
LLaMA-68/160M	Vicuna-13/33B	😟
Vicuna-68/160M	LLaMA-13/33B	😃
Vicuna-68/160M	Vicuna-13/33B	😃