initial upload

Author: YongWookHa

.gitignore

@@ -0,0 +1,116 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+env/
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+.hypothesis/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# pyenv
+.python-version
+
+# celery beat schedule file
+celerybeat-schedule
+
+# SageMath parsed files
+*.sage.py
+
+# dotenv
+.env
+
+# virtualenv
+.venv
+venv/
+ENV/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+
+# Pycharm
+.idea/
+
+# Experiments
+experiments/*
+!experiments/.keep
+
+# Pretrained Weights
+pretrained_weights/*
+!pretrained_weights/.keep
+
+# Data
+data/*
+!data/.keep

__init__.py

@@ -0,0 +1 @@
+__authors__ = ["Mo'men", "Hager"]

agents/__init__.py

@@ -0,0 +1,10 @@
+import os
+import sys
+
+path = os.path.dirname(os.path.abspath(__file__))
+
+for py in [f[:-3] for f in os.listdir(path) if f.endswith('.py') and f != '__init__.py']:
+    mod = __import__('.'.join([__name__, py]), fromlist=[py])
+    classes = [getattr(mod, x) for x in dir(mod) if isinstance(getattr(mod, x), type)]
+    for cls in classes:
+        setattr(sys.modules[__name__], cls.__name__, cls)

agents/base.py

@@ -0,0 +1,66 @@
+"""
+The Base Agent class, where all other agents inherit from, that contains definitions for all the necessary functions
+"""
+import logging
+
+
+class BaseAgent:
+    """
+    This base class will contain the base functions to be overloaded by any agent you will implement.
+    """
+
+    def __init__(self, config):
+        self.config = config
+        self.logger = logging.getLogger("Agent")
+
+    def load_checkpoint(self, file_name):
+        """
+        Latest checkpoint loader
+        :param file_name: name of the checkpoint file
+        :return:
+        """
+        raise NotImplementedError
+
+    def save_checkpoint(self, file_name="checkpoint.pth.tar", is_best=0):
+        """
+        Checkpoint saver
+        :param file_name: name of the checkpoint file
+        :param is_best: boolean flag to indicate whether current checkpoint's metric is the best so far
+        :return:
+        """
+        raise NotImplementedError
+
+    def run(self):
+        """
+        The main operator
+        :return:
+        """
+        raise NotImplementedError
+
+    def train(self):
+        """
+        Main training loop
+        :return:
+        """
+        raise NotImplementedError
+
+    def train_one_epoch(self):
+        """
+        One epoch of training
+        :return:
+        """
+        raise NotImplementedError
+
+    def validate(self):
+        """
+        One cycle of model validation
+        :return:
+        """
+        raise NotImplementedError
+
+    def finalize(self):
+        """
+        Finalizes all the operations of the 2 Main classes of the process, the operator and the data loader
+        :return:
+        """
+        raise NotImplementedError

agents/bert.py

@@ -0,0 +1,71 @@
+import numpy as np
+from tqdm import tqdm
+
+import random
+
+import torch
+from torch import nn
+from torch.backends import cudnn
+from torch.autograd import Variable
+import torchvision.utils as vutils
+
+from agent.base import BaseAgent
+
+cudnn.benchmark = True
+
+class BERTAgent(BaseAgent):
+
+    def __init__(self, config):
+        super().__init__(config)
+
+    def load_checkpoint(self, file_name):
+        """
+        Latest checkpoint loader
+        :param file_name: name of the checkpoint file
+        :return:
+        """
+        raise NotImplementedError
+
+    def save_checkpoint(self, file_name="checkpoint.pth.tar", is_best=0):
+        """
+        Checkpoint saver
+        :param file_name: name of the checkpoint file
+        :param is_best: boolean flag to indicate whether current checkpoint's metric is the best so far
+        :return:
+        """
+        raise NotImplementedError
+
+    def run(self):
+        """
+        The main operator
+        :return:
+        """
+        raise NotImplementedError
+
+    def train(self):
+        """
+        Main training loop
+        :return:
+        """
+        raise NotImplementedError
+
+    def train_one_epoch(self):
+        """
+        One epoch of training
+        :return:
+        """
+        raise NotImplementedError
+
+    def validate(self):
+        """
+        One cycle of model validation
+        :return:
+        """
+        raise NotImplementedError
+
+    def finalize(self):
+        """
+        Finalizes all the operations of the 2 Main classes of the process, the operator and the data loader
+        :return:
+        """
+        raise NotImplementedError

data/.keep

@@ -0,0 +1,35 @@
+import numpy as numpy
+import sentencepiece as spm
+
+import torch
+
+from torch.utils.data import DataLoader, TensorDataset, Dataset
+
+class SentencePairDataLoader:
+    def __init__(self, config):
+        self.config = config
+
+        if config.data_mode == "corpus":
+            self.f = open(self.config.data_folder, "r", encoding="utf-8", errors='ignore')
+            self.tokenizer = self.config.tokenizer
+            if self.tokenizer is 'bpe':
+                prefix = bpe_model
+                cmd = '--input={} --vocab_size={} --model_prefix={}'
+                cmd = cmd.format(self.config.data_folder, vocab_size, prefix)
+                try:
+                    spm.SentencePieceTrainer.Train(cmd)
+                    self.sp = spm.SenetencePieceProcesseor()
+                    sp.Load('{}.model'.format(prefix))
+                except Exception:
+                    raise
+                # bpe = self.sp.EncodeAsPieces(line) # to list
+            else:
+                raise NotImplementedError
+            self.max_len = max_len 
+            self.batch_size = self.config.batch_size
+        else:
+            raise NotImplementedError
+
+        
+
+            

datasets/SentencePair.py

@@ -0,0 +1,10 @@
+import os
+import sys
+
+path = os.path.dirname(os.path.abspath(__file__))
+
+for py in [f[:-3] for f in os.listdir(path) if f.endswith('.py') and f != '__init__.py']:
+    mod = __import__('.'.join([__name__, py]), fromlist=[py])
+    classes = [getattr(mod, x) for x in dir(mod) if isinstance(getattr(mod, x), type)]
+    for cls in classes:
+        setattr(sys.modules[__name__], cls.__name__, cls)

experiments/.keep

@@ -0,0 +1,10 @@
+import os
+import sys
+
+path = os.path.dirname(os.path.abspath(__file__))
+
+for py in [f[:-3] for f in os.listdir(path) if f.endswith('.py') and f != '__init__.py']:
+    mod = __import__('.'.join([__name__, py]), fromlist=[py])
+    classes = [getattr(mod, x) for x in dir(mod) if isinstance(getattr(mod, x), type)]
+    for cls in classes:
+        setattr(sys.modules[__name__], cls.__name__, cls)

graphs/__init__.py

@@ -0,0 +1,16 @@
+"""
+Binary Cross Entropy for DCGAN
+"""
+
+import torch
+import torch.nn as nn
+
+
+class BinaryCrossEntropy(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.loss = nn.BCELoss()
+
+    def forward(self, logits, labels):
+        loss = self.loss(logits, labels)
+        return loss

graphs/losses/__init__.py

@@ -0,0 +1,24 @@
+"""
+Cross Entropy 2D for CondenseNet
+"""
+
+import torch
+import torch.nn.functional as F
+import torch.nn as nn
+
+import numpy as np
+
+
+class CrossEntropyLoss(nn.Module):
+    def __init__(self, config=None):
+        super(CrossEntropyLoss, self).__init__()
+        if config == None:
+            self.loss = nn.CrossEntropyLoss()
+        else:
+            class_weights = np.load(config.class_weights)
+            self.loss = nn.CrossEntropyLoss(ignore_index=config.ignore_index,
+                                      weight=torch.from_numpy(class_weights.astype(np.float32)),
+                                      size_average=True, reduce=True)
+
+    def forward(self, inputs, targets):
+        return self.loss(inputs, targets)