llama : improve BPE pre-processing + LLaMA 3 and Deepseek support #11280
Job | Run time |
---|---|
2m 28s | |
1m 49s | |
1m 46s | |
1m 25s | |
1m 23s | |
1m 50s | |
3m 18s | |
5m 48s | |
3m 45s | |
5m 47s | |
2m 20s | |
2m 3s | |
3m 33s | |
1m 32s | |
2m 24s | |
2m 48s | |
31m 28s | |
3m 9s | |
27m 54s | |
6m 4s | |
5m 6s | |
5m 24s | |
21m 21s | |
6m 35s | |
4m 53s | |
5m 32s | |
5m 45s | |
1m 43s | |
4m 19s | |
5m 8s | |
15m 29s | |
0s | |
3h 13m 49s |