SQ and QM: Remove torch.cuda.empty_cache, use calibration_forward_context
#246
| Job | Run time |
|---|---|
| 2s | |
| 2s |
torch.cuda.empty_cache, use calibration_forward_context
#246
| Job | Run time |
|---|---|
| 2s | |
| 2s |