Uno

icon indicating a website link https://www.linkedin.com/in/dariocazzani icon indicating an email [email protected]

icon company Cisco icon location New York icon location Technical Lead - Machine Learning at Cisco, where I'm researching Computer Vision, 3D perception, and scene understanding.

Results 7 comments of Uno

PPO implementation

The real "game changer" for me was to run `envs_reset()` at each iteration

Homoscedastic Loss Function

I think the problem arises from the fact that the implementation doesn't use ![](https://latex.codecogs.com/gif.latex?{\color{Yellow}&space;\sigma&space;^{2}}) See [Issue 3](https://github.com/Hui-Li/multi-task-learning-example-PyTorch/issues/3)

about train dataset

Has anyone successfully downloaded the SIDD dataset?

Is this macOS problem?

You might want to try to run a Docker Container or a VM running Linux on your Mac

Getting outofMemory error: CUDA

You need at least 27GB of GPU memory for the 7B parameter model (around 10GB for the 3B one)

[request] Depth estimation documentation, training code and / or model weights

@patricklabatut - any updates on the depth estimation code? I am having a hard time reproducing with the same quality you show in the paper

Code & Checkpoint ETA

Hi, great paper and very interesting work Any ETA now on the release of the code? Thanks