[torchlib] Fix and improve quantization support #7425
| Job | Run time |
|---|---|
| 2m 56s | |
| 10m 49s | |
| 3s | |
| 5m 57s | |
| 8m 29s | |
| 2m 25s | |
| 10m 2s | |
| 17m 59s | |
| 8m 33s | |
| 10m 30s | |
| 2m 8s | |
| 8m 56s | |
| 5m 17s | |
| 17m 36s | |
| 14m 13s | |
| 19m 37s | |
| 18m 41s | |
| 3m 3s | |
| 3m 12s | |
| 2m 47s | |
| 2m 49s | |
| 9m 45s | |
| 7m 27s | |
| 2m 45s | |
| 9m 12s | |
| 1m 48s | |
| 5m 7s | |
| 1m 43s | |
| 1m 51s | |
| 4m 5s | |
| 3m 41s | |
| 6m 18s | |
| 5m 32s | |
| 32s | |
| 3h 55m 48s |