release v0.1.dev15

Fix some link bugs in the README.md

.gitignore

@@ -0,0 +1,118 @@
+# project
+.DS_Store
+.data/
+run_configs/
+test_tube_logs/
+test_tube_data/
+datasets/
+model_weights/
+app/models/
+pip-wheel-metadata/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+example.py
+timit_data/
+LJSpeech-1.1/
+
+# C extensions
+*.so
+
+.idea/
+
+# Distribution / packaging
+.Python
+env/
+ide_layouts/
+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/

MANIFEST.in

@@ -0,0 +1,9 @@
+graft docs
+
+include COPYING
+include AUTHORS
+
+recursive-include src/einsteinpy/tests *.py *.html
+
+prune docs/source/examples/.ipynb_checkpoints
+global-exclude *.py[cod] __pycache__ *.so *.dylib

README.md

@@ -1,66 +1,127 @@
-# Pytorch-lightning
-The Keras for ML-researchers in PyTorch.    
+<p align="center">
+  <a href="https://williamfalcon.github.io/pytorch-lightning/">
+    <img alt="" src="https://github.com/williamFalcon/pytorch-lightning/blob/master/docs/source/_static/lightning_logo.png" width="50">
+  </a>
+</p>
+<h3 align="center">
+  Pytorch Lightning
+</h3>
+<p align="center">
+  The Keras for ML researchers using PyTorch. More control. Less boilerplate.    
+</p>
+<p align="center">
+  <a href="https://badge.fury.io/py/pytorch-lightning"><img src="https://badge.fury.io/py/pytorch-lightning.svg" alt="PyPI version" height="18"></a>
+<!--   <a href="https://travis-ci.org/williamFalcon/test-tube"><img src="https://travis-ci.org/williamFalcon/pytorch-lightning.svg?branch=master"></a> -->
+  <a href="https://github.com/williamFalcon/pytorch-lightning/blob/master/COPYING"><img src="https://img.shields.io/badge/License-MIT-yellow.svg"></a>
+</p>   
+
+```bash
+pip install pytorch-lightning    
+```
+
+## Docs   
+In progress. Documenting now!  
+
+## Disclaimer
+This is a research tool I built for myself internally while doing my PhD. The API is not 100% production quality, but my hope is that by open-sourcing, we can all get it there (I don't have too much time nowadays to write production-level code).
+
+## What is it?  
+Keras is too abstract for researchers. Lightning makes it so you only have to define your model but still control all details of training if you need to.   
+
+Pytorch   
+<-- Lightning   
+Your model.   
+
+**Lightning will do the following for you:**   
+
+1. Run the training loop.   
+2. Run the validation loop.   
+3. Run the testing loop.   
+4. Early stopping.   
+5. Learning rate annealing.
+6. Can train complex models like GANs or anything with multiple optimizers.
+7. Weight checkpointing.   
+8. Model saving.   
+9. Model loading.   
+10. Log training details (through test-tube).  
+11. Run training on multiple GPUs (through test-tube).   
+12. Run training on a GPU cluster managed by SLURM (through test-tube).   
+13. Distribute memory-bound models on multiple GPUs.
+14. Give your model hyperparameters parsed from the command line OR a JSON file.   
+15. Run your model in a dev environment where nothing logs.      
 
 ## Usage
 To use lightning do 2 things:  
-1. [Define a trainer](https://github.com/williamFalcon/pytorch-lightning/blob/master/pytorch_lightning/trainer_main.py) (which will run ALL your models).   
-2. [Define a model](https://github.com/williamFalcon/pytorch-lightning/blob/master/pytorch_lightning/models/sample_model_template/model_template.py).     
+1. [Define a trainer](https://github.com/williamFalcon/pytorch-lightning/blob/master/docs/source/examples/basic_trainer.py) (which will run ALL your models).   
+2. [Define a model](https://github.com/williamFalcon/pytorch-lightning/blob/master/docs/source/examples/example_model.py).     
 
-### Example:   
-
-#### Define the trainer   
+#### Basic trainer example  
+See [this demo](https://github.com/williamFalcon/pytorch-lightning/blob/master/docs/source/examples/fully_featured_trainer.py) for a more robust trainer example.   
 
 ```python
-# trainer.py
+import os
+import sys
 
-from pytorch_lightning.models.trainer import Trainer   
+from test_tube import HyperOptArgumentParser, Experiment
+from pytorch_lightning.models.trainer import Trainer
+from pytorch_lightning.utils.arg_parse import add_default_args
 from pytorch_lightning.utils.pt_callbacks import EarlyStopping, ModelCheckpoint
-from my_project import My_Model   
-from test_tube import HyperOptArgumentParser, Experiment, SlurmCluster
+from demo.example_model import ExampleModel
 
-# --------------
-# TEST TUBE INIT
-exp = Experiment(
-    name='my_exp',
-    debug=True,
-    save_dir='/some/path',
-    autosave=False,
-    description='my desc'
-)
 
-# --------------------
-# CALLBACKS
-early_stop = EarlyStopping(
-    monitor='val_loss',
-    patience=3,
-    verbose=True,
-    mode='min'
-)
+def main(hparams):
+    """
+    Main training routine specific for this project
+    :param hparams:
+    :return:
+    """
+    # init experiment
+    exp = Experiment(
+        name=hparams.tt_name,
+        debug=hparams.debug,
+        save_dir=hparams.tt_save_path,
+        version=hparams.hpc_exp_number,
+        autosave=False,
+        description=hparams.tt_description
+    )
 
-model_save_path = 'PATH/TO/SAVE'
-checkpoint = ModelCheckpoint(
-    filepath=model_save_path,
-    save_function=None,
-    save_best_only=True,
-    verbose=True,
-    monitor='val_acc',
-    mode='min'
-)
+    exp.argparse(hparams)
+    exp.save()
 
-# configure trainer
-trainer = Trainer(
-    experiment=experiment,
-    cluster=cluster,
-    checkpoint_callback=checkpoint,
-    early_stop_callback=early_stop
-)  
+    model_save_path = '{}/{}/{}'.format(hparams.model_save_path, exp.name, exp.version)
+
+    # build model
+    model = ExampleModel(hparams)
+
+    # callbacks
+    early_stop = EarlyStopping(monitor='val_acc', patience=3, mode='min', verbose=True)
+    checkpoint = ModelCheckpoint(filepath=model_save_path, save_function=None, save_best_only=True, verbose=True, monitor='val_acc', mode='min')
+
+    # configure trainer
+    trainer = Trainer(experiment=exp, checkpoint_callback=checkpoint, early_stop_callback=early_stop)
+
+    # train model
+    trainer.fit(model)
+
+
+if __name__ == '__main__':
+
+    # use default args given by lightning
+    root_dir = os.path.split(os.path.dirname(sys.modules['__main__'].__file__))[0]
+    parent_parser = HyperOptArgumentParser(strategy='random_search', add_help=False)
+    add_default_args(parent_parser, root_dir)
+
+    # allow model to overwrite or extend args
+    parser = ExampleModel.add_model_specific_args(parent_parser)
+    hyperparams = parser.parse_args()
+
+    # train model
+    main(hyperparams)
 
-# init model and train
-model = My_Model()
-trainer.fit(model)
 ```
 
-#### Define the model   
+#### Basic model example  
+Here we only show the method signatures. It's up to you to define the content.   
 
 ```python
 from torch import nn
@@ -69,32 +130,32 @@ class My_Model(RootModule):
     def __init__(self):
         # define model
         self.l1 = nn.Linear(200, 10)
-    
+
     # ---------------
     # TRAINING
     def training_step(self, data_batch):
         x, y = data_batch
         y_hat = self.l1(x)
         loss = some_loss(y_hat)
-        
+
         return loss_val, {'train_loss': loss}
-    
+
     def validation_step(self, data_batch):
         x, y = data_batch
         y_hat = self.l1(x)
         loss = some_loss(y_hat)
-        
+
         return loss_val, {'val_loss': loss}
- 
+
      def validation_end(self, outputs):
         total_accs = []
-        
+
         for output in outputs:
             total_accs.append(output['val_acc'].item())
-        
+
         # return a dict
         return {'total_acc': np.mean(total_accs)}
-     
+
      # ---------------
      # SAVING
      def get_save_dict(self):
@@ -106,7 +167,7 @@ class My_Model(RootModule):
     def load_model_specific(self, checkpoint):
         # lightning loads for you. Here's your chance to say what you want to load
         self.load_state_dict(checkpoint['state_dict'])
-    
+
     # ---------------
     # TRAINING CONFIG
     def configure_optimizers(self):
@@ -114,7 +175,7 @@ class My_Model(RootModule):
         # lightning will call automatically
         optimizer = self.choose_optimizer('adam', self.parameters(), {'lr': self.hparams.learning_rate}, 'optimizer')
         return [optimizer]
-    
+
     @property
     def tng_dataloader(self):
         return pytorch_dataloader('train')
@@ -126,7 +187,7 @@ class My_Model(RootModule):
     @property
     def test_dataloader(self):
         return pytorch_dataloader('test')
-        
+
     # ---------------
     # MODIFY YOUR COMMAND LINE ARGS
     @staticmethod
@@ -135,6 +196,8 @@ class My_Model(RootModule):
         parser.add_argument('--out_features', default=20)
         return parser
 ```
+
+
 ### Details    
 
 #### Model definition   
@@ -143,7 +206,7 @@ class My_Model(RootModule):
 | training_step  |  Called with a batch of data during training | data from your dataloaders  | tuple: scalar, dict  |
 | validation_step  |  Called with a batch of data during validation  | data from your dataloaders  | tuple: scalar, dict |
 | validation_end  |  Collate metrics from all validation steps |  outputs: array where each item is the output of a validation step | dict: for logging |
-| get_save_dict  | called when your model needs to be saved (checkpoints, hpc save, etc...)  | None  |  dict to be saved | 
+| get_save_dict  | called when your model needs to be saved (checkpoints, hpc save, etc...)  | None  |  dict to be saved |
 
 #### Model training   
 | Name  | Description  |  Input |  Return |
@@ -159,7 +222,7 @@ class My_Model(RootModule):
 |---|---|---|---|
 | get_save_dict  | called when your model needs to be saved (checkpoints, hpc save, etc...)  | None  |  dict to be saved |
 |  load_model_specific |  called when loading a model | checkpoint: dict you created in get_save_dict  | dict: modified in whatever way you want  |
-    
+
 ## Optional model hooks.   
 Add these to the model whenever you want to configure training behavior.    
 

docs/source/examples/basic_trainer.py

@@ -0,0 +1,74 @@
+import os
+import sys
+
+from test_tube import HyperOptArgumentParser, Experiment
+from pytorch-lightning.models.trainer import Trainer
+from pytorch-lightning.utils.arg_parse import add_default_args
+from pytorch-lightning.utils.pt_callbacks import EarlyStopping, ModelCheckpoint
+from demo.example_model import ExampleModel
+
+
+def main(hparams):
+    """
+    Main training routine specific for this project
+    :param hparams:
+    :return:
+    """
+    # init experiment
+    exp = Experiment(
+        name=hparams.tt_name,
+        debug=hparams.debug,
+        save_dir=hparams.tt_save_path,
+        version=hparams.hpc_exp_number,
+        autosave=False,
+        description=hparams.tt_description
+    )
+
+    exp.argparse(hparams)
+    exp.save()
+
+    # build model
+    model = ExampleModel(hparams)
+
+    # callbacks
+    early_stop = EarlyStopping(
+        monitor='val_acc',
+        patience=3,
+        mode='min',
+        verbose=True,
+    )
+
+    model_save_path = '{}/{}/{}'.format(hparams.model_save_path, exp.name, exp.version)
+    checkpoint = ModelCheckpoint(
+        filepath=model_save_path,
+        save_function=None,
+        save_best_only=True,
+        verbose=True,
+        monitor='val_acc',
+        mode='min'
+    )
+
+    # configure trainer
+    trainer = Trainer(
+        experiment=exp,
+        checkpoint_callback=checkpoint,
+        early_stop_callback=early_stop,
+    )
+
+    # train model
+    trainer.fit(model)
+
+
+if __name__ == '__main__':
+
+    # use default args given by lightning
+    root_dir = os.path.split(os.path.dirname(sys.modules['__main__'].__file__))[0]
+    parent_parser = HyperOptArgumentParser(strategy='random_search', add_help=False)
+    add_default_args(parent_parser, root_dir)
+
+    # allow model to overwrite or extend args
+    parser = ExampleModel.add_model_specific_args(parent_parser)
+    hyperparams = parser.parse_args()
+
+    # train model
+    main(hyperparams)

docs/source/examples/example_model.py

@@ -0,0 +1,200 @@
+import torch.nn as nn
+import numpy as np
+from pytorch-lightning.root_module.root_module import RootModule
+from test_tube import HyperOptArgumentParser
+from torchvision.datasets import MNIST
+import torchvision.transforms as transforms
+import torch
+import torch.nn.functional as F
+
+
+class ExampleModel(RootModule):
+    """
+    Sample model to show how to define a template
+    """
+
+    def __init__(self, hparams):
+        # init superclass
+        super(ExampleModel, self).__init__(hparams)
+
+        self.batch_size = hparams.batch_size
+
+        # build model
+        self.__build_model()
+
+    # ---------------------
+    # MODEL SETUP
+    # ---------------------
+    def __build_model(self):
+        """
+        Layout model
+        :return:
+        """
+        self.c_d1 = nn.Linear(in_features=self.hparams.in_features, out_features=self.hparams.hidden_dim)
+        self.c_d1_bn = nn.BatchNorm1d(self.hparams.hidden_dim)
+        self.c_d1_drop = nn.Dropout(self.hparams.drop_prob)
+
+        self.c_d2 = nn.Linear(in_features=self.hparams.hidden_dim, out_features=self.hparams.out_features)
+
+    # ---------------------
+    # TRAINING
+    # ---------------------
+    def forward(self, x):
+        x = self.c_d1(x)
+        x = F.tanh(x)
+        x = self.c_d1_bn(x)
+        x = self.c_d1_drop(x)
+
+        x = self.c_d2(x)
+        logits = F.log_softmax(x, dim=1)
+
+        return logits
+
+    def loss(self, labels, logits):
+        nll = F.nll_loss(logits, labels)
+        return nll
+
+    def training_step(self, data_batch):
+        """
+        Called inside the training loop
+        :param data_batch:
+        :return:
+        """
+        # forward pass
+        x, y = data_batch
+        x = x.view(x.size(0), -1)
+        y_hat = self.forward(x)
+
+        # calculate loss
+        loss_val = self.loss(y, y_hat)
+
+        tqdm_dic = {'tng_loss': loss_val.item()}
+        return loss_val, tqdm_dic
+
+    def validation_step(self, data_batch):
+        """
+        Called inside the validation loop
+        :param data_batch:
+        :return:
+        """
+        x, y = data_batch
+        x = x.view(x.size(0), -1)
+        y_hat = self.forward(x)
+
+        loss_val = self.loss(y, y_hat)
+
+        # acc
+        labels_hat = torch.argmax(y_hat, dim=1)
+        val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)
+
+        output = {'y_hat': y_hat, 'val_loss': loss_val.item(), 'val_acc': val_acc}
+        return output
+
+    def validation_end(self, outputs):
+        """
+        Called at the end of validation to aggregate outputs
+        :param outputs: list of individual outputs of each validation step
+        :return:
+        """
+        val_loss_mean = 0
+        accs = []
+        for output in outputs:
+            val_loss_mean += output['val_loss']
+            accs.append(output['val_acc'])
+
+        val_loss_mean /= len(outputs)
+        tqdm_dic = {'val_loss': val_loss_mean, 'val_acc': np.mean(accs)}
+        return tqdm_dic
+
+    def update_tng_log_metrics(self, logs):
+        return logs
+
+    # ---------------------
+    # MODEL SAVING
+    # ---------------------
+    def get_save_dict(self):
+        checkpoint = {'state_dict': self.state_dict()}
+        return checkpoint
+
+    def load_model_specific(self, checkpoint):
+        self.load_state_dict(checkpoint['state_dict'])
+        pass
+
+    # ---------------------
+    # TRAINING SETUP
+    # ---------------------
+    def configure_optimizers(self):
+        """
+        return whatever optimizers we want here
+        :return: list of optimizers
+        """
+        optimizer = self.choose_optimizer(self.hparams.optimizer_name, self.parameters(), {'lr': self.hparams.learning_rate}, 'optimizer')
+        self.optimizers = [optimizer]
+        return self.optimizers
+
+    def __dataloader(self, train):
+        # init data generators
+        transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
+
+        dataset = MNIST(root=self.hparams.data_root, train=train, transform=transform, download=True)
+
+        loader = torch.utils.data.DataLoader(
+            dataset=dataset,
+            batch_size=self.hparams.batch_size,
+            shuffle=True
+        )
+
+        return loader
+
+    @property
+    def tng_dataloader(self):
+        if self._tng_dataloader is None:
+            try:
+                self._tng_dataloader = self.__dataloader(train=True)
+            except Exception as e:
+                print(e)
+                raise e
+        return self._tng_dataloader
+
+    @property
+    def val_dataloader(self):
+        if self._val_dataloader is None:
+            try:
+                self._val_dataloader = self.__dataloader(train=False)
+            except Exception as e:
+                print(e)
+                raise e
+        return self._val_dataloader
+
+    @property
+    def test_dataloader(self):
+        if self._test_dataloader is None:
+            try:
+                self._test_dataloader = self.__dataloader(train=False)
+            except Exception as e:
+                print(e)
+                raise e
+        return self._test_dataloader
+
+    @staticmethod
+    def add_model_specific_args(parent_parser):
+        parser = HyperOptArgumentParser(strategy=parent_parser.strategy, parents=[parent_parser])
+
+        # param overwrites
+        # parser.set_defaults(gradient_clip=5.0)
+
+        # network params
+        parser.opt_list('--drop_prob', default=0.2, options=[0.2, 0.5], type=float, tunable=False)
+        parser.add_argument('--in_features', default=28*28)
+        parser.add_argument('--hidden_dim', default=500)
+        parser.add_argument('--out_features', default=10)
+
+        # data
+        parser.add_argument('--data_root', default='/Users/williamfalcon/Developer/personal/research_lib/research_proj/datasets/mnist', type=str)
+
+        # training params (opt)
+        parser.opt_list('--learning_rate', default=0.001, type=float, options=[0.0001, 0.0005, 0.001, 0.005],
+                        tunable=False)
+        parser.opt_list('--batch_size', default=256, type=int, options=[32, 64, 128, 256], tunable=False)
+        parser.opt_list('--optimizer_name', default='adam', type=str, options=['adam'], tunable=False)
+        return parser

docs/source/examples/fully_featured_trainer.py

@@ -0,0 +1,201 @@
+import os
+import sys
+import numpy as np
+from time import sleep
+import torch
+
+from test_tube import HyperOptArgumentParser, Experiment, SlurmCluster
+from pytorch-lightning.models.trainer import Trainer
+from pytorch-lightning.utils.arg_parse import add_default_args
+
+from pytorch-lightning.utils.pt_callbacks import EarlyStopping, ModelCheckpoint
+
+SEED = 2334
+torch.manual_seed(SEED)
+np.random.seed(SEED)
+
+# ---------------------
+# DEFINE MODEL HERE
+# ---------------------
+from example_model import ExampleModel
+# ---------------------
+
+AVAILABLE_MODELS = {
+    'model_template': ExampleModel
+}
+
+
+"""
+Allows training by using command line arguments
+Run by:
+# TYPE YOUR RUN COMMAND HERE
+"""
+
+
+def main_local(hparams):
+    main(hparams, None, None)
+
+
+def main(hparams, cluster, results_dict):
+    """
+    Main training routine specific for this project
+    :param hparams:
+    :return:
+    """
+    on_gpu = torch.cuda.is_available()
+    if hparams.disable_cuda:
+        on_gpu = False
+
+    device = 'cuda' if on_gpu else 'cpu'
+    hparams.__setattr__('device', device)
+    hparams.__setattr__('on_gpu', on_gpu)
+    hparams.__setattr__('nb_gpus', torch.cuda.device_count())
+    hparams.__setattr__('inference_mode', hparams.model_load_weights_path is not None)
+
+    # delay each training start to not overwrite logs
+    process_position, current_gpu = TRAINING_MODEL.get_process_position(hparams.gpus)
+    sleep(process_position + 1)
+
+    # init experiment
+    log_dir = os.path.dirname(os.path.realpath(__file__))
+    exp = Experiment(
+        name='test_tube_exp',
+        debug=True,
+        save_dir=log_dir,
+        version=0,
+        autosave=False,
+        description='test demo'
+    )
+
+    exp.argparse(hparams)
+    exp.save()
+
+    # build model
+    print('loading model...')
+    model = TRAINING_MODEL(hparams)
+    print('model built')
+
+    # callbacks
+    early_stop = EarlyStopping(
+        monitor=hparams.early_stop_metric,
+        patience=hparams.early_stop_patience,
+        verbose=True,
+        mode=hparams.early_stop_mode
+    )
+
+    model_save_path = '{}/{}/{}'.format(hparams.model_save_path, exp.name, exp.version)
+    checkpoint = ModelCheckpoint(
+        filepath=model_save_path,
+        save_function=None,
+        save_best_only=True,
+        verbose=True,
+        monitor=hparams.model_save_monitor_value,
+        mode=hparams.model_save_monitor_mode
+    )
+
+    # configure trainer
+    trainer = Trainer(
+        experiment=exp,
+        cluster=cluster,
+        checkpoint_callback=checkpoint,
+        early_stop_callback=early_stop,
+    )
+
+    # train model
+    trainer.fit(model)
+
+
+def get_default_parser(strategy, root_dir):
+
+    possible_model_names = list(AVAILABLE_MODELS.keys())
+    parser = HyperOptArgumentParser(strategy=strategy, add_help=False)
+    add_default_args(parser, root_dir, possible_model_names=possible_model_names, rand_seed=SEED)
+    return parser
+
+
+def get_model_name(args):
+    for i, arg in enumerate(args):
+        if 'model_name' in arg:
+            return args[i+1]
+
+
+def optimize_on_cluster(hyperparams):
+    # enable cluster training
+    cluster = SlurmCluster(
+        hyperparam_optimizer=hyperparams,
+        log_path=hyperparams.tt_save_path,
+        test_tube_exp_name=hyperparams.tt_name
+    )
+
+    # email for cluster coms
+    cluster.notify_job_status(email='add_email_here', on_done=True, on_fail=True)
+
+    # configure cluster
+    cluster.per_experiment_nb_gpus = hyperparams.per_experiment_nb_gpus
+    cluster.job_time = '48:00:00'
+    cluster.gpu_type = '1080ti'
+    cluster.memory_mb_per_node = 48000
+
+    # any modules for code to run in env
+    cluster.add_command('source activate pytorch_lightning')
+
+    # name of exp
+    job_display_name = hyperparams.tt_name.split('_')[0]
+    job_display_name = job_display_name[0:3]
+
+    # run hopt
+    print('submitting jobs...')
+    cluster.optimize_parallel_cluster_gpu(
+        main,
+        nb_trials=hyperparams.nb_hopt_trials,
+        job_name=job_display_name
+    )
+
+
+if __name__ == '__main__':
+
+    model_name = get_model_name(sys.argv)
+    if model_name is None:
+        model_name = 'model_template'
+
+    # use default args
+    root_dir = os.path.split(os.path.dirname(sys.modules['__main__'].__file__))[0]
+    parent_parser = get_default_parser(strategy='random_search', root_dir=root_dir)
+
+    # allow model to overwrite or extend args
+    TRAINING_MODEL = AVAILABLE_MODELS[model_name]
+    parser = TRAINING_MODEL.add_model_specific_args(parent_parser)
+    hyperparams = parser.parse_args()
+
+    # format GPU layout
+    os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
+    gpu_ids = hyperparams.gpus.split(';')
+
+    # RUN TRAINING
+    if hyperparams.on_cluster:
+        # Gets called when running via HPC cluster
+        print('RUNNING ON SLURM CLUSTER')
+        os.environ["CUDA_VISIBLE_DEVICES"] = ','.join(gpu_ids)
+        optimize_on_cluster(hyperparams)
+
+    elif hyperparams.single_run_gpu:
+        # run on 1 gpu
+        print(f'RUNNING 1 TRIAL ON GPU. gpu: {gpu_ids[0]}')
+        os.environ["CUDA_VISIBLE_DEVICES"] = gpu_ids[0]
+        main(hyperparams, None, None)
+
+    elif hyperparams.local or hyperparams.single_run:
+        # run 1 trial but on CPU
+        os.environ["CUDA_VISIBLE_DEVICES"] = '0'
+        print('RUNNING LOCALLY')
+        main(hyperparams, None, None)
+
+    else:
+        # multiple GPUs on same machine
+        print(f'RUNNING MULTI GPU. GPU ids: {gpu_ids}')
+        hyperparams.optimize_parallel_gpu(
+            main_local,
+            gpu_ids=gpu_ids,
+            nb_trials=hyperparams.nb_hopt_trials,
+            nb_workers=len(gpu_ids)
+        )

pyproject.toml

@@ -0,0 +1,5 @@
+[build-system]
+requires = [
+    "setuptools",
+    "wheel",
+]

pytorch_lightning/models/model_examples/bilstm.py

@@ -1,167 +0,0 @@
-import torch.nn as nn
-import numpy as np
-
-from test_tube import HyperOptArgumentParser
-import torch
-from torch.autograd import Variable
-from sklearn.metrics import confusion_matrix, f1_score
-from torch.nn import functional as F
-
-
-class BiLSTMPack(nn.Module):
-    """
-    Sample model to show how to define a template
-    """
-    def __init__(self, hparams):
-        # init superclass
-        super(BiLSTMPack, self).__init__(hparams)
-
-        self.hidden = None
-
-        # trigger tag building
-        self.ner_tagset = {'O': 0, 'I-Bio': 1}
-        self.nb_tags = len(self.ner_tagset)
-
-        # build model
-        print('building model...')
-        if hparams.model_load_weights_path is None:
-            self.__build_model()
-            print('model built')
-        else:
-            self = BiLSTMPack.load(hparams.model_load_weights_path, hparams.on_gpu, hparams)
-            print('model loaded from: {}'.format(hparams.model_load_weights_path))
-
-    def __build_model(self):
-        """
-        Layout model
-        :return:
-        """
-        # design the number of final units
-        self.output_dim = self.hparams.nb_lstm_units
-
-        # when it's bidirectional our weights double
-        if self.hparams.bidirectional:
-            self.output_dim *= 2
-
-        # total number of words
-        total_words = len(self.tng_dataloader.dataset.words_token_to_idx)
-
-        # word embeddings
-        self.word_embedding = nn.Embedding(
-            num_embeddings=total_words + 1,
-            embedding_dim=self.hparams.embedding_dim,
-            padding_idx=0
-        )
-
-        # design the LSTM
-        self.lstm = nn.LSTM(
-            self.hparams.embedding_dim,
-            self.hparams.nb_lstm_units,
-            num_layers=self.hparams.nb_lstm_layers,
-            bidirectional=self.hparams.bidirectional,
-            dropout=self.hparams.drop_prob,
-            batch_first=True,
-        )
-
-        # map to tag space
-        self.fc_out = nn.Linear(self.output_dim, self.out_dim)
-        self.hidden_to_tag = nn.Linear(self.output_dim, self.nb_tags)
-
-
-    def init_hidden(self, batch_size):
-
-        # the weights are of the form (nb_layers * 2 if bidirectional, batch_size, nb_lstm_units)
-        mult = 2 if self.hparams.bidirectional else 1
-        hidden_a = torch.randn(self.hparams.nb_layers * mult, batch_size, self.nb_rnn_units)
-        hidden_b = torch.randn(self.hparams.nb_layers * mult, batch_size, self.nb_rnn_units)
-
-        if self.hparams.on_gpu:
-            hidden_a = hidden_a.cuda()
-            hidden_b = hidden_b.cuda()
-
-        hidden_a = Variable(hidden_a)
-        hidden_b = Variable(hidden_b)
-
-        return (hidden_a, hidden_b)
-
-    def forward(self, model_in):
-        # layout data (expand it, etc...)
-        # x = sequences
-        x, seq_lengths = model_in
-        batch_size, seq_len = x.size()
-
-        # reset RNN hidden state
-        self.hidden = self.init_hidden(batch_size)
-
-        # embed
-        x = self.word_embedding(x)
-
-        # run through rnn using packed sequences
-        x = torch.nn.utils.rnn.pack_padded_sequence(x, seq_lengths, batch_first=True)
-        x, self.hidden = self.lstm(x, self.hidden)
-        x, _ = torch.nn.utils.rnn.pad_packed_sequence(x, batch_first=True)
-
-        # if asked for only last state, use the h_n which is the same as out(t=n)
-        if not self.return_sequence:
-            # pull out hidden states
-            # h_n = (nb_directions * nb_layers, batch_size, emb_size)
-            nb_directions = 2 if self.bidirectional else 1
-            (h_n, _) = self.hidden
-
-            # reshape to make indexing easier
-            # forward = 0, backward = 1 (of nb_directions)
-            h_n = h_n.view(self.nb_layers, nb_directions, batch_size, self.nb_rnn_units)
-
-            # pull out last forward
-            forward_h_n = h_n[-1, 0, :, :]
-            x = forward_h_n
-
-            # if bidirectional, also pull out the last hidden of backward network
-            if self.bidirectional:
-                backward_h_n = h_n[-1, 1, :, :]
-                x = torch.cat([forward_h_n, backward_h_n], dim=1)
-
-        # project to tag space
-        x = x.contiguous()
-        x = x.view(-1, self.output_dim)
-        x = self.hidden_to_tag(x)
-
-        return x
-
-    def loss(self, model_out):
-        # cross entropy loss
-        logits, y = model_out
-        y, y_lens = y
-
-        # flatten y and logits
-        y = y.view(-1)
-        logits = logits.view(-1, self.nb_tags)
-
-        # calculate a mask to remove padding tokens
-        mask = (y >= 0).float()
-
-        # count how many tokens we have
-        num_tokens = int(torch.sum(mask).data[0])
-
-        # pick the correct values and mask out
-        logits = logits[range(logits.shape[0]), y] * mask
-
-        # compute the ce loss
-        ce_loss = -torch.sum(logits)/num_tokens
-
-        return ce_loss
-
-    def pull_out_last_embedding(self, x, seq_lengths, batch_size, on_gpu):
-        # grab only the last activations from the non-padded ouput
-        x_last = torch.zeros([batch_size, 1, x.size(-1)])
-        for i, seq_len in enumerate(seq_lengths):
-            x_last[i, :, :] = x[i, seq_len-1, :]
-
-        # put on gpu when requested
-        if on_gpu:
-            x_last = x_last.cuda()
-
-        # turn into torch var
-        x_last = Variable(x_last)
-
-        return x_last

pytorch_lightning/models/trainer.py

@@ -1,9 +1,9 @@
 import torch
 import tqdm
 import numpy as np
-from research_lib.root_module.memory import get_gpu_memory_map
+from pytorch_lightning.root_module.memory import get_gpu_memory_map
 import traceback
-from research_lib.root_module.model_saving import TrainerIO
+from pytorch_lightning.root_module.model_saving import TrainerIO
 from torch.optim.lr_scheduler import MultiStepLR
 
 
@@ -11,17 +11,17 @@ class Trainer(TrainerIO):
 
     def __init__(self,
                  experiment,
-                 cluster,
                  checkpoint_callback, early_stop_callback,
+                 cluster=None,
                  process_position=0,
                  current_gpu_name=0,
                  on_gpu=False,
                  enable_tqdm=True,
-                 overfit_pct=None,
+                 overfit_pct=0.0,
                  track_grad_norm=-1,
                  check_val_every_n_epoch=1,
                  fast_dev_run=False,
-                 accumulate_grad_batches=False,
+                 accumulate_grad_batches=1,
                  enable_early_stop=True, max_nb_epochs=5, min_nb_epochs=1,
                  train_percent_check=1.0, val_percent_check=1.0, test_percent_check=1.0, val_check_interval=0.95,
                  log_save_interval=1, add_log_row_interval=1,
@@ -86,7 +86,7 @@ class Trainer(TrainerIO):
             self.test_percent_check = overfit_pct
 
     def __is_function_implemented(self, f_name):
-        f_op = getattr(self, f_name, None)
+        f_op = getattr(self.model, f_name, None)
         return callable(f_op)
 
     @property
@@ -195,6 +195,7 @@ class Trainer(TrainerIO):
     # -----------------------------
     def fit(self, model):
         self.model = model
+        model.trainer = self
 
         # transfer data loaders from model
         self.__get_dataloaders(model)
@@ -226,7 +227,8 @@ class Trainer(TrainerIO):
         self.experiment.save()
 
         # enable cluster checkpointing
-        self.enable_auto_hpc_walltime_manager()
+        if self.cluster is not None:
+            self.enable_auto_hpc_walltime_manager()
 
         # ---------------------------
         # CORE TRAINING LOOP
@@ -268,13 +270,14 @@ class Trainer(TrainerIO):
                 # ---------------
                 # RUN TRAIN STEP
                 # ---------------
-                self.__run_tng_batch(data_batch)
+                batch_result = self.__run_tng_batch(data_batch)
+                early_stop_epoch = batch_result == -1
 
                 # ---------------
                 # RUN VAL STEP
                 # ---------------
                 is_val_check_batch = (batch_nb + 1) % self.val_check_batch == 0
-                if self.fast_dev_run or is_val_check_batch:
+                if self.fast_dev_run or is_val_check_batch or early_stop_epoch:
                     self.__run_validation()
 
                 # when batch should be saved
@@ -306,6 +309,10 @@ class Trainer(TrainerIO):
                 if self.__is_function_implemented('on_batch_end'):
                     self.model.on_batch_end()
 
+                # end epoch early
+                if early_stop_epoch:
+                    break
+
             # hook
             if self.__is_function_implemented('on_epoch_end'):
                 self.model.on_epoch_end()
@@ -320,13 +327,16 @@ class Trainer(TrainerIO):
                 if stop:
                     return
 
+
     def __run_tng_batch(self, data_batch):
         if data_batch is None:
-            return
+            return 0
 
         # hook
         if self.__is_function_implemented('on_batch_start'):
-            self.model.on_batch_start()
+            response = self.model.on_batch_start(data_batch)
+            if response == -1:
+                return -1
 
         if self.enable_tqdm:
             self.prog_bar.update(1)
@@ -368,6 +378,8 @@ class Trainer(TrainerIO):
         if self.__is_function_implemented('on_batch_end'):
             self.model.on_batch_end()
 
+        return 0
+
     def __run_validation(self):
         # decide if can check epochs
         can_check_epoch = (self.current_epoch + 1) % self.check_val_every_n_epoch == 0

pytorch_lightning/root_module/hooks.py

@@ -1,7 +1,7 @@
 import torch
 
 class ModelHooks(torch.nn.Module):
-    def on_batch_start(self):
+    def on_batch_start(self, data_batch):
         pass
 
     def on_batch_end(self):
@@ -18,3 +18,4 @@ class ModelHooks(torch.nn.Module):
 
     def on_post_performance_check(self):
         pass
+

pytorch_lightning/root_module/root_module.py

@@ -14,7 +14,7 @@ class RootModule(GradInformation, ModelIO, OptimizerConfig, ModelHooks):
     def __init__(self, hparams):
         super(RootModule, self).__init__()
         self.hparams = hparams
-        self.on_gpu = hparams.on_gpu
+
         self.dtype = torch.FloatTensor
         self.exp_save_path = None
         self.current_epoch = 0
@@ -24,6 +24,14 @@ class RootModule(GradInformation, ModelIO, OptimizerConfig, ModelHooks):
         self.overfit = hparams.overfit
         self.gradient_clip = hparams.gradient_clip
         self.num = 2
+        self.trainer = None
+
+        # track if gpu was requested for checkpointing
+        self.on_gpu = False
+        try:
+            self.on_gpu = hparams.on_gpu
+        except Exception as e:
+            pass
 
         # computed vars for the dataloaders
         self._tng_dataloader = None

pytorch_lightning/trainer_main.py

@@ -9,7 +9,6 @@ from pytorch_lightning.utils.arg_parse import add_default_args
 from time import sleep
 
 from pytorch_lightning.utils.pt_callbacks import EarlyStopping, ModelCheckpoint
-
 SEED = 2334
 torch.manual_seed(SEED)
 np.random.seed(SEED)

pytorch_lightning/utils/arg_parse.py

@@ -1,4 +1,4 @@
-def add_default_args(parser, root_dir, possible_model_names, rand_seed):
+def add_default_args(parser, root_dir, rand_seed=None, possible_model_names=None):
 
     # tng, test, val check intervals
     parser.add_argument('--eval_test_set', dest='eval_test_set', action='store_true', help='true = run test set also')
@@ -32,7 +32,9 @@ def add_default_args(parser, root_dir, possible_model_names, rand_seed):
 
     # model paths
     parser.add_argument('--model_load_weights_path', default=None, type=str)
-    parser.add_argument('--model_name', default='', help=','.join(possible_model_names))
+
+    if possible_model_names is not None:
+        parser.add_argument('--model_name', default='', help=','.join(possible_model_names))
 
     # test_tube settings
     parser.add_argument('-en', '--tt_name', default='r_lib_')
@@ -58,7 +60,9 @@ def add_default_args(parser, root_dir, possible_model_names, rand_seed):
     parser.add_argument('--overfit', default=-1, type=float, help='% of dataset to use with this option. float, or -1 for none')
 
     # debug args
-    parser.add_argument('--random_seed', default=rand_seed, type=int)
+    if rand_seed is not None:
+        parser.add_argument('--random_seed', default=rand_seed, type=int)
+
     parser.add_argument('--live', dest='live', action='store_true', help='runs on gpu without cluster')
     parser.add_argument('--enable_debug', dest='debug', action='store_true', help='enables/disables test tube')
     parser.add_argument('--enable_local', dest='local', action='store_true', help='enables local tng')

pytorch_lightning/utils/embeddings.py

@@ -13,11 +13,7 @@ class PretrainedEmbedding(torch.nn.Embedding):
         >>> emb = PretrainedEmbedding(embedding_path='glove.840B.300d.txt',embedding_dim=300, task_vocab={'hello': 1, 'world': 2})
         >>> data = torch.Tensor([[0, 1], [0, 2]]).long()
         >>> embedded = emb(data)
-        tensor([[[ 0.0000,  0.0000,  0.0000,  ...,  0.0000,  0.0000,  0.0000],
-         [ 0.2523,  0.1018, -0.6748,  ...,  0.1787, -0.5192,  0.3359]],
-
-        [[ 0.0000,  0.0000,  0.0000,  ...,  0.0000,  0.0000,  0.0000],
-         [-0.0067,  0.2224,  0.2771,  ...,  0.0594,  0.0014,  0.0987]]])
+        
 
 
         :param embedding_path:
@@ -37,7 +33,8 @@ class PretrainedEmbedding(torch.nn.Embedding):
         self.weight = new_emb.weight
 
         # apply freeze
-        self.weight.requires_grad = not freeze
+        should_freeze = not freeze
+        self.weight.requires_grad = should_freeze
 
     def __load_task_specific_embeddings(self, vocab_words, embedding_path, emb_dim, freeze):
         """
@@ -97,11 +94,11 @@ class PretrainedEmbedding(torch.nn.Embedding):
 
 if __name__ == '__main__':
     emb = PretrainedEmbedding(
-        embedding_path='/Users/waf/Developer/NGV/research-fermat/fermat/.vector_cache/glove.840B.300d.txt',
+        embedding_path='/Users/waf/Developer',
         embedding_dim=300,
         task_vocab={'hello': 1, 'world': 2}
     )
 
     data = torch.Tensor([[0, 1], [0, 2]]).long()
     embedded = emb(data)
-    print(embedded)
+    print(embedded)

setup.cfg

@@ -0,0 +1,21 @@
+[tool:pytest]
+norecursedirs =
+    .git
+    dist
+    build
+python_files =
+    test_*.py
+doctest_plus = disabled
+addopts = --strict
+markers =
+    slow
+    remote_data
+    filterwarnings
+
+[pycodestyle]
+ignore = E731,W504
+max-line-length = 120
+
+[flake8]
+ignore = E731,W504,F401,F841
+max-line-length = 120

setup.py

@@ -2,12 +2,27 @@
 
 from setuptools import setup, find_packages
 
-setup(name='pytorch-lightning',
-      version='0.0.1',
-      description='Rapid research framework',
-      author='',
-      author_email='',
-      url='https://github.com/williamFalcon/pytorch-lightning',
-      install_requires=[],
-      packages=find_packages()
-      )
+# https://packaging.python.org/guides/single-sourcing-package-version/
+
+# http://blog.ionelmc.ro/2014/05/25/python-packaging/
+setup(
+    name="pytorch-lightning",
+    version='0.1.dev15',
+    description="The Keras for ML researchers using PyTorch",
+    author="William Falcon",
+    author_email="waf2107@columbia.edu",
+    url="https://github.com/williamFalcon/pytorch-lightning",
+    download_url="https://github.com/williamFalcon/pytorch-lightning",
+    license="MIT",
+    keywords=["deep learning", "pytorch", "AI"],
+    python_requires=">=3.5",
+    install_requires=[
+        "torch",
+        "tqdm",
+        "test-tube",
+    ],
+    packages=find_packages(),
+    long_description=open("README.md", encoding="utf-8").read(),
+    include_package_data=True,
+    zip_safe=False,
+)

tests/README.md

@@ -1,65 +0,0 @@
-# Testing setup
-
-## A. Enable CircleCI for your project
-1. Integrate CircleCI by clicking "Set up Project" at [this link](https://circleci.com/add-projects/gh/NextGenVest).
-
-## B. Add your own tests
-1. In the /tests, emulate exactly the folder structure for your module found under /bot_seed
-2. To create a test for file ```/bot_seed/folder/example.py```:
-    - create the file ```/tests/folder/example_test.py```
-    - notice the **_test**
-    - notice the mirror path under **/tests**
-
-3. Your ```example_test.py``` file should have these main components
-
-```python
-# example.py
-
-def function_i_want_to_test(x):
-    return x*2
-
-def square(x):
-    return x*x
-
-```
-
-```python
-# example_test.py
-
-import pytest
-
-# do whatever imports you need
-from app.bot_seed.folder.example import function_i_want_to_test, square
-
-def test_function_i_want_to_test():
-    answer = function_i_want_to_test(4)
-    assert answer == 8
-
-# -----------------------------------
-# Your function must start with test_
-# -----------------------------------
-def test_square():
-    answer = square(3)
-    assert answer == 9
-
-# -----------------------------------
-# boilerplate (link this file to pytest)
-# -----------------------------------
-if __name__ == '__main__':
-    pytest.main([__file__])
-```
-
-## C. Add build passing badge
-1. Create a CircleCI status token:
-    - Go here: https://circleci.com/gh/NextGenVest/your-project-name/edit#api
-    - Click create token
-    - Select status
-    - Type "badge status"
-
-2. Get a copy of the markdown code:
-    - Go here: https://circleci.com/gh/NextGenVest/your-project-name/edit#badges
-    - Select master
-    - Select "badge status" token
-    - Select image URL
-    - Copy the image url link and change the html at the top of the root README.md file for your project
-    

tests/research_proj/sample_model_template/model_template_test.py

@@ -1,13 +0,0 @@
-import pytest
-
-"""
-Example test to show how to add a test for anything in the project.  
-Look at the README for more instructions   
-"""
-
-
-def test_cube():
-    assert 27 == 27
-
-if __name__ == '__main__':
-    pytest.main([__file__])