Dynamically set max_new_tokens
based on output feature length, GMSL and model window size
#11306
Job | Run time |
---|---|
18m 39s | |
5s | |
20m 14s | |
21m 58s | |
17m 10s | |
32m 27s | |
6m 58s | |
13m 38s | |
28m 27s | |
25m 41s | |
26m 26s | |
9m 26s | |
30m 50s | |
29m 40s | |
12m 2s | |
4h 53m 41s |