speaker_follower

speaker_follower copied to clipboard

Visualization Code

13

Hi In https://arxiv.org/pdf/1806.02724.pdf, there are some visualizations of the predicted trajectories in Fig 4. Is the code to generate these available?

Mrs-Hudson

Bumps [pyyaml](https://github.com/yaml/pyyaml) from 5.1 to 5.4. Changelog Sourced from pyyaml's changelog. 5.4 (2021-01-19) yaml/pyyaml#407 -- Build modernization, remove distutils, fix metadata, build wheels, CI to GHA yaml/pyyaml#472 -- Fix for...

dependabot[bot]

Bumps [numpy](https://github.com/numpy/numpy) from 1.13.3 to 1.22.0. Release notes Sourced from numpy's releases. v1.22.0 NumPy 1.22.0 Release Notes NumPy 1.22.0 is a big release featuring the work of 153 contributors spread...

dependabot[bot]

Bumps [pillow](https://github.com/python-pillow/Pillow) from 4.3.0 to 9.3.0. Release notes Sourced from pillow's releases. 9.3.0 https://pillow.readthedocs.io/en/stable/releasenotes/9.3.0.html Changes Initialize libtiff buffer when saving #6699 [@​radarhere] Limit SAMPLESPERPIXEL to avoid runtime DOS #6700 [@​wiredfool]...

dependabot[bot]

Download: data augmentation files: 404 Not Found

2

Hi, while executing ./tasks/R2R/data/download.sh I get the following result: ``` silversurfer42@ideas:~/fun/grounded/speaker_follower/tasks/R2R/data$ sudo bash download.sh [sudo] password for silversurfer42: --2021-09-25 15:00:49-- https://www.dropbox.com/s/lztjsji51pr5ig2/R2R_train.json Resolving www.dropbox.com (www.dropbox.com)... 2620:100:6022:18::a27d:4212, 162.125.66.18 Connecting to www.dropbox.com (www.dropbox.com)|2620:100:6022:18::a27d:4212|:443......

marvosyntactical

Second step of training is unclear

1

Hi, when I try to do the second step of the training `export SPEAKER_PATH_PREFIX=tasks/R2R/speaker/snapshots/speaker_teacher_imagenet_mean_pooled_train_iter_20000` it fails because the files available after training are speaker_teacher_imagenet_mean_pooled_train_iter_20000_enc and speaker_teacher_imagenet_mean_pooled_train_iter_20000_dec, which one should we...

mees

I changed the dataset from R2R to R4R which contains over 45k instructions in the val_unseen dataset. The training is killed when about 15k of them are evaluated. The machine...

ZhangTianrong

How to use non-panoramic space

1

Hi, could you please describe what changes ones need to make in order to train your model with data augmentation and pragmatic inference, but without panoramic space, so normal camera...

mees

Hi, if we want to use our own image feature extractor, how do we get the images for a given viewpoint?

balaramb

Hi, have you tried using beam search to synthesize instructions? I saw your beam search code in speaker but you simply use greedy search to synthesize instructions. https://github.com/ronghanghu/speaker_follower/blob/389ee0229b6fabe6ab93b00bdcaf1131912a8474/tasks/R2R/speaker.py#L342

zwx8981