Loading…
Loading…
Models / PolDense
GPU memory needed (smallest version of each): PolDense 17M about 1.5 GB, PolDense 32M about 1.6 GB, PolDense 68M about 1.7 GB, PolDense 150M about 1.9 GB, PolDense 400M about 2.4 GB, PolDense 1B about 4 GB. The cheapest runs on RTX 3060 from $0.09/h right now.
6 models (17M to 1B) · ↓ 10.2K downloads a month · from $0.09/h on RTX 3060 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-17M | vLLM | 30 MB | 1.5 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-32M | vLLM | 60 MB | 1.6 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-68M | vLLM | 140 MB | 1.7 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-150M | vLLM | 300 MB | 1.9 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-400M | vLLM | 790 MB | 2.4 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalOPI-PIB/PolDense-1B | vLLM | 2.1 GB | 4 GB | $0.09/h RTX 3060 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").