Idefics models
3 Idefics models on Hugging Face, from 8.4B to 8.9B parameters. At the precision each one is published in, the smallest needs about 18.9 GB of VRAM (Idefics3-8B-Llama3, cheapest live fit: RTX A5000) and the largest about 39.9 GB (idefics-9b, cheapest live fit: RTX A6000). The cheapest way to run Idefics3-8B-Llama3 is $0.176/hr.
Idefics models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| idefics2-8b | 8.4B | F32 | 37.6 GB | RTX A6000 | $0.363/hr |
| Idefics3-8B-Llama3 | 8.5B | BF16 | 18.9 GB | RTX A5000 | $0.176/hr |
| idefics-9b | 8.9B | F32 | 39.9 GB | RTX A6000 | $0.363/hr |
VRAM is for the precision the model is published in, with the same overhead and an 8,192-token context assumed on every page; see the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.
