Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

WIP: TCVC implementation #252

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Conversation

TE-andrewshin
Copy link
Contributor

No description provided.

@@ -0,0 +1,59 @@
# TCVC Example by NNabla
This is a NNabla implementation of the "Automatic Temporally Coherent Video Colorization" (TCVC).
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please add a link to the paper here as well.

For more information, see [NNabla CUDA extension package installation guild](https://nnabla.readthedocs.io/en/latest/python/pip_installation_cuda.html).

## Dataset
Please prepare the dataset by extracting frames (require `ffmpeg`, currently won't work when installed with pip but works with conda) from anime episodes with
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You should mention that users should prepare their own arbitrary anime clip.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How should user specify frame rate? Is it supposed to be 24? Or Maybe kindly add one-liner ffmpeg command :)

### Train with the original tcvc (or reference-based version)
```
# by a single GPU
./scripts/generate_tcvc.sh
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This script is for inference once you already have a trained model. So should be changed to train_tcvc.sh, etc. and the likewise for the examples below.

# by a single GPU
./scripts/generate_tcvc.sh
or
./scripts/generate_tcvc_feat.sh
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please describe clearly what is different when you have _feat in the filename.

@@ -0,0 +1 @@
mpirun -N 4 python train.py --fix-global-epoch 0 --d-n-scales 1 --g-n-scales 1 --save-path ./results/tcvc_01 --gen_feat
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Get rid of GUNDAM from all filenames.

@@ -0,0 +1,15 @@
# Copyright (c) 2019 Sony Corporation. All Rights Reserved.
Copy link
Contributor Author

@TE-andrewshin TE-andrewshin Dec 27, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can't remember our discussion on neu, but I don't think you made changes on everything, so you can probably load neu of nnabla-examples as is and keep the separate files here for the ones you did make changes only. (Maybe I said I would do it? I can't really remember)

@@ -0,0 +1,73 @@
import os
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Add the following at the header of every file (except README)

# Copyright 2021 Sony Group Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants