Keller Jordan @kellerjordan0 · 13 Apr 2023
Neural network trainings are nondeterministic. Repeated runs each produce a unique network, often with significantly _varying_ test-set performance. 🆕📜 I demonstrate that this variation has a simple statistical structure, and is harmless & inevitable https://t.co/1zzpNHi0Vy https://t.co/vhnjQlqOsH
273 795Views
1 328Likes
160Reposts
28Replies
17Quotes
745Bookmarks
Is that a lot?
29.3×vs this author's median9 348 views is typical
100Percentile for this authorof 5 recent posts
188×vs 10K–100K median1 456 views is typical
14.2× audienceReachviews ÷ followers
0.56%Engagement rateof viewers reacted