Did you use .half() precision training?

isl-org / MiDaS

Code for robust monocular depth estimation described in "Ranftl et. al., Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer, TPAMI 2022"

MIT License

4.27k stars 599 forks source link

Did you use .half() precision training? #221

Open andafonin opened 1 year ago

andafonin commented 1 year ago

Hi, I wonder how you managed to fit batch size 24 of your 48 GB GPU. Did you use .half() conversion for model and training data during training?

JamesMcCullochDickens commented 9 months ago

Also to add to this would running 16 bit precision training speed up inference?