Method Bits 7B 13B 30B 65B 8x7B
Full 16 160GB 320GB 600GB 1200GB 1000GB
Freeze 16 20GB 40GB 120GB 240GB 200GB
LoRA 16 16GB 32GB 80GB 160GB 120GB
QLoRA 8 10GB 16GB 40GB 80GB 80GB
QLoRA 4 6GB 12GB 24GB 48GB 32GB
Large models are not easily accessible
Model Inference memory Fine-tuning memory
Т5-11B 22 GB 176 GB
LLaMA-33B 66 GB 396 GB
LLaMA-65B 130 GB 780 GB
Last active
May 2, 2024 01:12
-
-
Save bigsnarfdude/fac2bacc78e98714ce6e94335b93e37b to your computer and use it in GitHub Desktop.
tableVRAM.md
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment