Commits · 4211abb142588f93af124b1d65dca45b5f8c6df4 · esde / machine-learning / AMBS · GitLab

Snippets Groups Projects

Sep 25, 2020
- Rename top-level directory 'video_prediction_savp' to 'video_prediction_tools'. · 4211abb1
  Michael Langguth authored Sep 25, 2020
  
  4211abb1
Sep 15, 2020
- remove useless files and move files to new structure · e75c7ab8
  gong1 authored Sep 15, 2020
  
  e75c7ab8
Jun 18, 2020
- move video_prediction_savp into subdir · 937846d8
  b.gong authored Jun 18, 2020
  
  937846d8
Dec 27, 2019
- update script · 8e3a205d
  Bing Gong authored Dec 27, 2019
  
  8e3a205d
Dec 07, 2018
- Use smaller batch size for longer model, and automatically infer the number of... · 99efa48c
  Alex Lee authored Dec 7, 2018
  99efa48c
Dec 06, 2018
- Fix order of checking checkpoint_dir in training script. · 7c882876
  Alex Lee authored Dec 6, 2018
  
  7c882876
- Decrease the eval_summary_freq default. · 508c7b7f
  Alex Lee authored Dec 6, 2018
  
  508c7b7f
Dec 05, 2018
- Adjust training loop logging outputs. · 39d5bee6
  Alex Lee authored Dec 5, 2018
  
  39d5bee6
- Change long_model's mode to test and change to consistent behavior in capturing saveable variables. · bc05779b
  Alex Lee authored Dec 5, 2018
  
  bc05779b
- Refactor savp models and networks. Add learned prior that does it's own... · 6712a2d3
  Alex Lee authored Dec 5, 2018
  6712a2d3
- Add validation with longer sequences in the training loop. · 1254cf36
  Alex Lee authored Dec 5, 2018
  
  1254cf36
Dec 04, 2018
- Allow accum_eval_summary at the begining when it's also the last step. · e4fd154e
  Alex Lee authored Dec 4, 2018
  
  e4fd154e
- Fix changed lr backed to 0.001. Use tf.accumulate_n instead of sum for total... · 9f8ce2c1
  Alex Lee authored Dec 4, 2018
  9f8ce2c1
- Fix regression where model_hparams_dict not being used in train script. · bef614a3
  Alex Lee authored Dec 4, 2018
  
  bef614a3
- Add script-level flag for aggregate_nccl. · 2a6e7697
  Alex Lee authored Dec 4, 2018
  
  2a6e7697
- Use only one model for both datasets and other improvement in training script. · f327666d
  Alex Lee authored Dec 4, 2018
  f327666d
Dec 03, 2018
- Refactor the base model and add multi-gpu training with nccl. · 5b00c1f9
  Alex Lee authored Dec 3, 2018
  5b00c1f9
Nov 30, 2018
- Add diversity metric and summary for accumulated metrics. · 0a884dac
  Alex Lee authored Nov 30, 2018
  
  0a884dac
Nov 15, 2018
- Add gif_summary and remove tensor-to-gif conversion after graph evaluation.... · 78d821f5
  Alex Lee authored Nov 9, 2018
  78d821f5
May 27, 2018
- Minor. · 441538cb
  Alex Lee authored May 27, 2018
  
  441538cb
Apr 12, 2018
- Add imports from future in main scripts. · 28e9c79f
  Alex Lee authored Apr 12, 2018
  
  28e9c79f
Apr 08, 2018
- Rename softmotion30_44k to bair. Fix train_all script. Add hparams_dict argparse options. · 6d0e1d93
  Alex Lee authored Apr 8, 2018
  
  6d0e1d93
- Infer num_gpus automatically. Clean up evaluate script. Changed defaults:... · e28808c6
  Alex Lee authored Apr 8, 2018
  e28808c6
- Move max_steps to model hparams. · fb4ac556
  Alex Lee authored Apr 8, 2018
  
  fb4ac556
Mar 28, 2018
- Remove force_gpu_compatible option. · 9e0c3ad2
  Alex Lee authored Mar 28, 2018
  
  9e0c3ad2
- Clean up train script. · 5263dda2
  Alex Lee authored Mar 28, 2018
  
  5263dda2
Mar 24, 2018
- Minor changes in train.py. · cb57fae4
  Alex Lee authored Mar 24, 2018
  
  cb57fae4
Mar 23, 2018
- Add train script. · 82d1e21d
  Alex Lee authored Mar 23, 2018
  
  82d1e21d