xlm-roberta-base
by FacebookAI
279M params · fill-mask · 795 likes · 20.2M downloads
xlm-roberta-base is a 279M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (4708 tok/s), GeForce RTX 5090 (7060 tok/s), M4 Max 128GB (1722 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 4708 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 7060 tok/s | comfortable |
| M4 Max 128GB | 128GB | 1722 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 861 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 861 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 8625 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 17349 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 4162 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 3442 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 3341 tok/s | comfortable |