| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Inference Latency | OPT model family | Latency (ms)6.2 | 79 | |
| Quantization | OPT | Processing Time (s)4.8 | 46 | |
| Quantization | OPT v1 (train) | Processing Time (min)0.08 | 23 | |
| INT2 Quantization | OPT-1.3B | Memory Footprint (GB)4.49 | 12 | |
| INT2 Quantization | OPT-125M | Memory Cost (GB)1.39 | 12 | |
| Block-wise quantization complexity analysis | OPT Models | GFLOPS0.24 | 12 | |
| INT2 Quantization | OPT-6.7B | Memory Footprint (GB)11.68 | 10 | |
| Open domain dialogue | OPT-66B ZS | RSR74.7 | 9 | |
| Energy Consumption Estimation | OPT-66B | Energy (J)1,786.62 | 8 | |
| Energy Consumption Estimation | OPT-30B | Energy (J)846.54 | 8 | |
| Energy Consumption Estimation | OPT-13B | Energy (J)385.31 | 8 | |
| Energy Consumption Estimation | OPT 6.7B | Energy (J)215.27 | 8 | |
| Energy Consumption Estimation | OPT-2.7B | Energy (J)95.67 | 8 | |
| Energy Consumption Estimation | OPT 1.3B | Energy (J)55.8 | 8 | |
| Energy Consumption Estimation | OPT-350M | Energy (J)20.17 | 8 | |
| Energy Consumption Estimation | OPT-125M | Energy Consumption (J)11.41 | 8 | |
| Text-to-Image Generation | OPT-66B ZS user input pool v1.4 (test) | RSR71.34 | 5 | |
| Inference Latency | OPT-30B | Latency (ms)15.7 | 5 | |
| Inference Latency | OPT-175B first FFN layer | Latency (ms)0.225 | 5 | |
| Machine Unlearning | OPT-125 Personal Set | Accuracy77.4 | 4 | |
| Machine Unlearning | OPT-125 Forget Set | Accuracy65.3 | 4 | |
| Text Generation | OPT 512 prompt + 512 generation tokens 6.7B | Throughput (tokens/sec)72.5 | 4 | |
| Text Generation | OPT 512 prompt + 32 generation tokens 6.7B | Throughput (token/s)50.5 | 4 | |
| Text Generation | OPT 512 prompt + 1024 generation tokens 6.7B | Throughput (token/s)62.3 | 3 |