vLLM Inference Performance on Qwen3-1.7B
0.54Model Load Time (s)Safetensors
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Safetensors2025.12 | 0.54 | 101.911 | 198.45 | 10,547.66 | 3.25 | |
| CryptoTensorsEncryption=Unencrypted2025.12 | 0.594 | 101.727 | 190.9 | 10,560.64 | 3.25 | |
| CryptoTensorsEncryption=Encrypted2025.12 | 5.65 | 102.108 | 195.19 | 10,666.21 | 3.25 |