GPT-NeoX models
2 GPT-NeoX models on Hugging Face, from 204M to 20.7B parameters. At the precision each one is published in, the smallest needs about 0.9 GB of VRAM (japanese-gpt-neox-small, cheapest live fit: V100) and the largest about 46.4 GB (gpt-neox-20b, cheapest live fit: RTX A6000). The cheapest way to run japanese-gpt-neox-small is $0.088/hr.
GPT-NeoX models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| japanese-gpt-neox-small | 204M | F32 | 0.9 GB | V100 | $0.088/hr |
| gpt-neox-20b | 20.7B | F16 | 46.4 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.
