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
Some ideas about TensorFlow re-implementation #179
Hi! First of all, thanks for this project. I was really impressed by the quality of depth that can be estimated using MiDaS.
I am now trying to re-reimplement the model "midas_v21_small" using TensorFlow2, but I get stuck at the point when I need to create the network backbone: EfficientNet-Lite3 architecture is not unfortunately provided in tf.keras.applications.efficientnet.
So my question is: do you by chance have something to suggest me how to create an EfficientNet-Lite3 in TensorFlow2? Thanks a lot in advance!
P.S.: I would prefer to not reimplement it myself if possible :)
Hi! First of all, thanks for this project. I was really impressed by the quality of depth that can be estimated using MiDaS.
I am now trying to re-reimplement the model "midas_v21_small" using TensorFlow2, but I get stuck at the point when I need to create the network backbone: EfficientNet-Lite3 architecture is not unfortunately provided in
tf.keras.applications.efficientnet
.So my question is: do you by chance have something to suggest me how to create an EfficientNet-Lite3 in TensorFlow2? Thanks a lot in advance!
P.S.: I would prefer to not reimplement it myself if possible :)