Loading…
Loading…
Models / LaBSE-ru-turbo / LaBSE-ru-turbo ~128M
Embeddings · served with Text Embeddings Inference · ↓ 169 downloads a month · View on Hugging Face
sergeyzh/BERTA-uncased needs about 1.8 GB of GPU memory. It runs on one GTX 1650 from $0.06/h right now; best value: RTX 5090 at $0.56/h.
The same model packaged differently: a quantized version needs less GPU memory and runs on a cheaper machine.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalsergeyzh/LaBSE-ru-turbo | vLLM | 260 MB | 1.8 GB | $0.09/h RTX 3060 |
| BERTAsergeyzh/BERTA | vLLM | 260 MB | 1.8 GB | $0.09/h RTX 3060 |
| rubert base retrieversergeyzh/rubert-base-retriever | vLLM | 260 MB | 1.8 GB | $0.09/h RTX 3060 |
| BERTA uncasedsergeyzh/BERTA-uncased | Text Embeddings Inference | 260 MB | 1.8 GB | $0.06/h GTX 1650 |
Deploying ONNX models from the catalog is coming soon. Today, download the files from Hugging Face and upload them in Inference.
Prices are per hour for the whole machine and its disk, right now. Speed is Vast's DLPerf score of the machine, relative to the cheapest. Best value: the most speed per dollar. Only machines with a Vast reliability above 95% that download at 400 Mbit/s or more are offered: the machine fetches the model itself.