Gemma 3
vision
Pick a size: the video memory shown is the Q4_K_M file (the protocol's quantisation) plus about 1.5 GB for a 4,000-token context.
270M
8 GB card- Parameters
- 0,3B
- Q4_K_M file
- 0,2 Go
- Recommended VRAM
- ≥ 1,7 Go
- Max context
- 32 768
Not measured yet
1B
8 GB card- Parameters
- 1B
- Q4_K_M file
- 0,8 Go
- Recommended VRAM
- ≥ 2,3 Go
- Max context
- 32 768
Not measured yet
4B
8 GB card- Parameters
- 4,3B
- Q4_K_M file
- 2,3 Go
- Recommended VRAM
- ≥ 3,8 Go
- Max context
- 131 072
Not measured yet
12B
12 GB card- Parameters
- 12,2B
- Q4_K_M file
- 6,8 Go
- Recommended VRAM
- ≥ 8,3 Go
- Max context
- 131 072
Not measured yet
27B
24 GB card- Parameters
- 27,4B
- Q4_K_M file
- 15,4 Go
- Recommended VRAM
- ≥ 16,9 Go
- Max context
- 131 072
Not measured yet
MoE (“30B-A3B”): all the memory of a 30B, but the speed of a model with 3B active parameters. “> 32 GB”: several cards, or part of the model in system memory (much slower).