File size: 318 Bytes
			
			02af082 05b398a  | 
								1 2 3 4 5 6 7 8 9 10 11  | 
								# Jamba
- β
 qlora w/ deepspeed Zero-2 needs at least 2x GPUs and
  - 35GiB VRAM per GPU w minimal context length
  - 56GiB VRAM per GPU (w multipack enabled)
- β
 qlora w/ deepspeed Zero-3 needs at least 2x GPUs and 67GiB VRAM (wtf?)
- β
 qlora single-gpu, ~51GiB VRAM
- β
 multipack
- β FSDP
- β 8-bit LoRA
 |