vLLM Inference Performance on Qwen3-14B
3.082Model Load Time (s)Safetensors
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Safetensors2025.12 | 3.082 | 134.347 | 78.02 | 11,378.91 | 27.78 | |
| CryptoTensorsEncryption=Unencrypted2025.12 | 3.165 | 134.765 | 76.64 | 11,387.66 | 27.78 | |
| CryptoTensorsEncryption=Encrypted2025.12 | 61.879 | 136.105 | 78.61 | 11,425.65 | 27.78 |