distilbert-base-multilingual-cased
by distilbert
135M params · fill-mask · 235 likes · 1.0M downloads
distilbert-base-multilingual-cased is a 135M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (9693 tok/s), GeForce RTX 5090 (14537 tok/s), M4 Max 128GB (3547 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 9693 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 14537 tok/s | comfortable |
| M4 Max 128GB | 128GB | 3547 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 1773 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 1773 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 17760 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 35722 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 8570 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 7087 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 6879 tok/s | comfortable |