resnet-50
by microsoft
26M params · image-classification · 495 likes · 344.8k downloads
resnet-50 is a 26M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (51268 tok/s), GeForce RTX 5090 (76883 tok/s), M4 Max 128GB (18761 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 51268 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 76883 tok/s | comfortable |
| M4 Max 128GB | 128GB | 18761 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 9380 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 9380 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 93931 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 188925 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 45329 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 37485 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 36385 tok/s | comfortable |