Bing

Dear Ogroth, I checked the gqn dataset from https://github.com/deepmind/gqn-datasets. Shepard_metzler_7_parts contains 900 tfrecords for training. Each tfrecord has 20 scenes. So there are only 18000 scenes. For Mazes, there are...

build_pointcloud.py and interpolate_poses.py errors

1

Hello, In line.84 of build_pointcloud.py, the timestamps[i] should be timestamps[i+1] as the len(poses) isn't equal to len(timestamps). The timestamps' value has been changed by function line.72 (interpolate_ins_poses) or line.75 (interpolate_vo_poses)...

Loss

1

Hi, Many thanks for your contribution on implementing GQN. Recently, I implemented GQN. After the training on rooms_ring_camera, I found that the loss will converge to around 6950 after 200K...

Loss

2

Hi, Many thanks for your contribution on implementing GQN. Recently, I implemented GQN. After the training on rooms_ring_camera, I found that the loss will converge to around 6950 after 200K...

GQN Loss and results

4

Hi, Many thanks for your contribution on implementing GQN. Recently, I implemented GQN. After the training on rooms_ring_camera, I found that the loss will converge to around 6950 after 200K...

Dear Fabio Viola, I checked the gqn dataset from https://github.com/deepmind/gqn-datasets. Shepard_metzler_7_parts contains 900 tfrecords for training. Each tfrecord has 20 scenes. So there are only 18000 scenes. For Mazes, there...