Loading…
Loading…
Models / MiniLM-L12-H384-uncased / MiniLM-L12-H384-uncased
Embeddings · served with Text Embeddings Inference · ↓ 144 downloads a month · View on Hugging Face
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 |
|---|---|---|---|---|
| ms marco MiniLM L6 v2 (cross-encoder)cross-encoder/ms-marco-MiniLM-L6-v2 | Text Embeddings Inference | 50 MB | 1.6 GB | $0.08/h RTX 3060 |
| ms marco MiniLM L4 v2 (cross-encoder)cross-encoder/ms-marco-MiniLM-L4-v2 | Text Embeddings Inference | 40 MB | 1.5 GB | $0.08/h RTX 3060 |
| all MiniLM L12 v2 (sentence-transformers)sentence-transformers/all-MiniLM-L12-v2 | Text Embeddings Inference | 70 MB | 1.6 GB | $0.08/h RTX 3060 |
| ms marco MiniLM L12 v2 (cross-encoder)cross-encoder/ms-marco-MiniLM-L12-v2 | Text Embeddings Inference | 70 MB | 1.6 GB | $0.08/h RTX 3060 |
| ms marco MiniLM L 6 v2 (Xenova)Xenova/ms-marco-MiniLM-L-6-v2 | ONNX Runtime | 20 MB | 1 GB | $0.08/h GTX 1660 SUPER |
| ms marco MiniLM L2 v2 (cross-encoder)cross-encoder/ms-marco-MiniLM-L2-v2 | Text Embeddings Inference | 30 MB | 1.5 GB | $0.08/h RTX 3060 |
| all MiniLM L12 v2 (Xenova)Xenova/all-MiniLM-L12-v2 | ONNX Runtime | 30 MB | 1.1 GB | $0.08/h GTX 1660 SUPER |
| ms marco MiniLM L2 v2 (KnightsAnalytics)KnightsAnalytics/ms-marco-MiniLM-L2-v2 | ONNX Runtime | 20 MB | 1 GB | $0.08/h GTX 1660 SUPER |
| all MiniLM L12 v2 code search 512 (isuruwijesiri)isuruwijesiri/all-MiniLM-L12-v2-code-search-512 | Text Embeddings Inference | 70 MB | 1.6 GB | $0.08/h RTX 3060 |
| context reranker (Badawi-a)Badawi-a/context-reranker | ONNX Runtime | 20 MB | 1 GB | $0.08/h GTX 1660 SUPER |
Deploying embedding and reranking models (Text Embeddings Inference) is coming soon.
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.