Surface Normals Estimation On Taskonomy

L1 error

Results

Performance results of various models on this benchmark

Comparison Table

Model Name	L1 error
robust-learning-through-cross-task	4.80