Loading…
Loading…
Models / MiniLM-L12-H384-uncased
1 model · ↓ 87.8M downloads a month · from $0.08/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 |
|---|---|---|---|---|
| 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 GTX 1660 SUPER |
| 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 GTX 1660 SUPER |
| all MiniLM L12 v2 (sentence-transformers)sentence-transformers/all-MiniLM-L12-v2 | Text Embeddings Inference | 70 MB | 1.6 GB | $0.08/h GTX 1660 SUPER |
| 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 GTX 1660 SUPER |
| 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 GTX 1660 SUPER |
| 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 GTX 1660 SUPER |
| context reranker (Badawi-a)Badawi-a/context-reranker | ONNX Runtime | 20 MB | 1 GB | $0.08/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").