clip-ViT-B-32-multilingual-v1
by sentence-transformers
135M params · sentence-similarity · 188 likes · 100.5k downloads
clip-ViT-B-32-multilingual-v1 is a 135M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (9745 tok/s), GeForce RTX 5090 (14613 tok/s), M4 Max 128GB (3566 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 9745 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 14613 tok/s | comfortable |
| M4 Max 128GB | 128GB | 3566 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 1783 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 1783 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 17854 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 35910 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 8616 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 7125 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 6916 tok/s | comfortable |