vanilj/ smaug-llama-3-70b-instruct

582 2 years ago

This model was built using a new Smaug recipe for improving performance on real world multi-turn conversations applied to meta-llama/Meta-Llama-3-70B-Instruct.

11 models

smaug-llama-3-70b-instruct:latest
ad9579e5734e • 43GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q2_K
d6b40dce253a • 26GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q3_K_S
f4d96e27f090 • 31GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q3_K_M
2e1b19eb5137 • 34GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q3_K_L
1c54d8aea32a • 37GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q4_K_S
e580b7cae026 • 40GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q4_K_M
ad9579e5734e • 43GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q5_K_S
7b5249fb809d • 49GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q5_K_M
3c3354de6304 • 50GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:Q6_K
3abcd2574a85 • 58GB • 8K context window •
Text input • 2 years ago
smaug-llama-3-70b-instruct:f16
a05eb0e9a72b • 282GB • 8K context window •
Text input • 2 years ago