Dynamically set max_new_tokens
based on output feature length, GMSL and model window size
#11297
Job | Run time |
---|---|
19m 40s | |
20m 19s | |
23m 16s | |
19m 8s | |
2s | |
26m 12s | |
23m 25s | |
26m 52s | |
37m 45s | |
11m 5s | |
8m 58s | |
26m 54s | |
11m 7s | |
35m 10s | |
10m 59s | |
5h 0m 52s |