Knowledge Distillation. Intial Commit.

E3_Distill_ESRGAN/config/config.yaml

@@ -0,0 +1,16 @@
+# Config File for Distillation
+
+student_network: "name1"
+student_networks_config:
+  name1:
+    param1: 1
+    param2: "two"
+    paramN: "N"
+  name2:
+    param1: 1
+    param2: "two"
+    paramN: "N"
+  name3:
+    param1: 1
+    param2: "two"
+    paramN: "N"

E3_Distill_ESRGAN/libs/model.py

@@ -0,0 +1,2 @@
+from libs import models
+from libs.models.abstract import Registry

E3_Distill_ESRGAN/libs/models/__init__.py

@@ -0,0 +1 @@
+

E3_Distill_ESRGAN/libs/models/abstract.py

@@ -0,0 +1,14 @@
+from tensorflow.python import keras
+class Registry(type):
+  models = {}
+  def __init__(cls, name, bases, attrs):
+    if name.lower() != "models":
+      Registry.models[cls.__name__.lower()] = cls
+
+# Abstract class for Auto Registration of Kernels
+class Models(keras.models.Model, metaclass=Registry):
+  def __init__(self, *args, **kwargs):
+    super(Models, self).__init__()
+    self.init(*args, **kwargs)
+  def init(self, *args, **kwargs):
+    pass

E3_Distill_ESRGAN/libs/models/teacher.py

@@ -0,0 +1,5 @@
+import os
+import sys
+# Fetching Generator from ESRGAN
+sys.path.insert(0, "../E2_ESRGAN")
+from lib.model import RRDBNet

E3_Distill_ESRGAN/libs/settings.py

@@ -0,0 +1,52 @@
+import os
+import yaml
+
+def singleton(cls):
+  instances = {}
+  def getinstance(*args, **kwargs):
+    distill_config = kwargs.get("student", "")
+    key = cls.__name__
+    if distill_config:
+      key = "%s_student" % (cls.__name__)
+    if key not in instances:
+      instances[key] = cls(*args, **kwargs)
+    return instances[key]
+  return getinstance
+
+@singleton
+class Settings(object):
+  def __init__(self, filename="config.yaml", student=False):
+    self.__path = os.path.abspath(filename)
+
+  @property
+  def path(self):
+    return os.path.dirname(self.__path)
+
+  def __getitem__(self, index):
+    with open(self.__path, "r") as file_:
+      return yaml.load(file_.read(), Loader=yaml.FullLoader)[index]
+
+  def get(self, index, default=None):
+    with open(self.__path, "r") as file_:
+      return yaml.load(file_.read(), Loader=yaml.FullLoader).get(index, default)
+
+
+class Stats(object):
+  def __init__(self, filename="stats.yaml"):
+    if os.path.exists(filename):
+      with open(filename, "r") as file_:
+        self.__data = yaml.load(file_.read(), Loader=yaml.FullLoader)
+    else:
+      self.__data = {}
+    self.file = filename
+
+  def get(self, index, default=None):
+    self.__data.get(index, default)
+
+  def __getitem__(self, index):
+    return self.__data[index]
+
+  def __setitem__(self, index, data):
+    self.__data[index] = data
+    with open(self.file, "w") as file_:
+      yaml.dump(self.__data, file_, default_flow_style=False)

E3_Distill_ESRGAN/libs/train.py

@@ -0,0 +1 @@
+import tensorflow as tf

E3_Distill_ESRGAN/libs/utils.py

@@ -0,0 +1,5 @@
+import os
+import sys
+# Fetching Ra Loss from project ESRGAN
+sys.path.insert(0, os.path.abspath(".."))
+from E2_ESRGAN.lib.utils import RelativisticAverageLoss

E3_Distill_ESRGAN/main.py

@@ -0,0 +1,45 @@
+from absl import logging
+import argparse
+from libs.models import teacher
+from libs import model
+from libs import utils
+from libs import settings
+import tensorflow as tf
+"""
+  Compressing GANs using Knowledge Distillation.
+  Teacher GAN: ESRGAN (https://github.com/captain-pool/E2_ESRGAN)
+
+Citation:
+  @article{DBLP:journals/corr/abs-1902-00159,
+    author    = {Angeline Aguinaldo and
+    Ping{-}Yeh Chiang and
+    Alexander Gain and
+    Ameya Patil and
+    Kolten Pearson and
+    Soheil Feizi},
+    title     = {Compressing GANs using Knowledge Distillation},
+    journal   = {CoRR},
+    volume    = {abs/1902.00159},
+    year      = {2019},
+    url       = {http://arxiv.org/abs/1902.00159},
+    archivePrefix = {arXiv},
+    eprint    = {1902.00159},
+    timestamp = {Tue, 21 May 2019 18:03:39 +0200},
+    biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1902-00159},
+    bibsource = {dblp computer science bibliography, https://dblp.org}
+  }
+"""
+def main(**kwargs):
+  student_settings = settings.Settings("../E2_ESRGAN/config.yaml", student=True)
+  teacher_settings = settings.Settings("config/config.yaml", student=False)
+
+if __name__ == "__main__":
+  parser = argparse.ArgumentParser()
+  parser.add_argument("--logdir", default=None, help="Path to log directory")
+  parser.add_argument("--modeldir", default=None, help="directory to store checkpoints and SavedModel")
+  parser.add_argument("--verbose", "-v", action="count", default=0, help="Increases Verbosity. Repeat to increase more")
+  FLAGS, unparsed = parser.parse_known_args()
+  log_levels = [logging.WARNING, logging.INFO, logging.DEBUG]
+  log_level = log_levels[min(FLAGS.verbose, len(log_levels)-1)]
+  logging.set_verbosity(log_level)
+  main(**vars(FLAGS))

E3_Distill_ESRGAN/train.sh

@@ -0,0 +1 @@
+#! /bin/bash