alabulei's picture
Update README.md
8c6d275 verified
metadata
base_model: bigcode/starcoder2-15b-instruct-v0.1
inference: false
license: bigcode-openrail-m
library_name: transformers
model_creator: bigcode
model_name: StarCoder2-15B-Instruct-v0.1
pipeline_tag: text-generation
quantized_by: Second State Inc.
tags:
  - code

StarCoder2-15B-Instruct-v0.1-GGUF

Original Model

bigcode/starcoder2-15b-instruct-v0.1

Run with Gaianet

Prompt template:

prompt template: comming soon

Context size:

chat_ctx_size: 6144

Run with GaiaNet:

Quantized GGUF Models

Name Quant method Bits Size Use case
starcoder2-15b-instruct-v0.1-Q2_K.gguf Q2_K 2 6.19 GB smallest, significant quality loss - not recommended for most purposes
starcoder2-15b-instruct-v0.1-Q3_K_L.gguf Q3_K_L 3 8.97 GB small, substantial quality loss
starcoder2-15b-instruct-v0.1-Q3_K_M.gguf Q3_K_M 3 8.1 GB very small, high quality loss
starcoder2-15b-instruct-v0.1-Q3_K_S.gguf Q3_K_S 3 6.99 GB very small, high quality loss
starcoder2-15b-instruct-v0.1-Q4_0.gguf Q4_0 4 9.07 GB legacy; small, very high quality loss - prefer using Q3_K_M
starcoder2-15b-instruct-v0.1-Q4_K_M.gguf Q4_K_M 4 9.86 GB medium, balanced quality - recommended
starcoder2-15b-instruct-v0.1-Q4_K_S.gguf Q4_K_S 4 9.26 GB small, greater quality loss
starcoder2-15b-instruct-v0.1-Q5_0.gguf Q5_0 5 11 GB legacy; medium, balanced quality - prefer using Q4_K_M
starcoder2-15b-instruct-v0.1-Q5_K_M.gguf Q5_K_M 5 11.4 GB large, very low quality loss - recommended
starcoder2-15b-instruct-v0.1-Q5_K_S.gguf Q5_K_S 5 11 GB large, low quality loss - recommended
starcoder2-15b-instruct-v0.1-Q6_K.gguf Q6_K 6 13.1 GB very large, extremely low quality loss
starcoder2-15b-instruct-v0.1-Q8_0.gguf Q8_0 8 17 GB very large, extremely low quality loss - not recommended
starcoder2-15b-instruct-v0.1-f16.gguf f16 16 31.9 GB

Quantized with llama.cpp b2824