-
Notifications
You must be signed in to change notification settings - Fork 93
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
base: master
Are you sure you want to change the base?
Conversation
@@ -0,0 +1,59 @@ | |||
# TCVC Example by NNabla | |||
This is a NNabla implementation of the "Automatic Temporally Coherent Video Colorization" (TCVC). |
There was a problem hiding this comment.
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 |
There was a problem hiding this comment.
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.
There was a problem hiding this comment.
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 |
There was a problem hiding this comment.
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 |
There was a problem hiding this comment.
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 |
There was a problem hiding this comment.
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. |
There was a problem hiding this comment.
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 |
There was a problem hiding this comment.
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.
No description provided.