Loading…
Loading…
Models / Arabic-STS-Matryoshka
GPU memory needed (smallest version of each): Arabic-STS-Matryoshka ~135M about 1.8 GB, Arabic-STS-Matryoshka ~560M about 3 GB. The cheapest runs on GTX 1660 SUPER from $0.07/h right now.
2 models (~135M to ~560M) · ↓ 5.9K downloads a month · from $0.07/h on GTX 1660 SUPER 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 |
|---|---|---|---|---|
| Originalomarelshehy/Arabic-STS-Matryoshka-V2 | vLLM | 270 MB | 1.8 GB | $0.08/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalomarelshehy/Arabic-STS-Matryoshka | Text Embeddings Inference | 1.1 GB | 3 GB | $0.07/h GTX 1660 SUPER |
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").