[CPU] Add per-token asym INT8 dynamic quantization support to QKV/MLP node #8720
Job | Run time |
---|---|
19s | |
1m 7s | |
6m 36s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
1s | |
8m 3s |
Job | Run time |
---|---|
19s | |
1m 7s | |
6m 36s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
0s | |
1s | |
8m 3s |