wassname/pytorch-lightning
1
0
Fork 0
mirror of https://github.com/wassname/pytorch-lightning.git synced 2026-09-09 11:32:07 +08:00
Code Issues Packages Projects Releases Wiki Activity

Deployed 28618a3 with MkDocs version: 1.0.4

Browse Source
This commit is contained in:
William Falcon
2019-06-28 16:46:09 -05:00
parent f1d00ca90d
commit ee71cdbcfa
17 changed files with 560 additions and 101 deletions
Download Patch File Download Diff File Expand all files Collapse all files
LightningModule/RequiredTrainerInterface/index.html
+640
View File
@@ -0,0 +1,640 @@
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="shortcut icon" href="../../img/favicon.ico">
<title>Lightning Module interface - Pytorch lightning Documentation</title>
<link href='https://fonts.googleapis.com/css?family=Lato:400,700|Roboto+Slab:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>
<link rel="stylesheet" href="../../css/theme.css" type="text/css" />
<link rel="stylesheet" href="../../css/theme_extra.css" type="text/css" />
<link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
<script>
// Current page data
var mkdocs_page_name = "Lightning Module interface";
var mkdocs_page_input_path = "LightningModule/RequiredTrainerInterface.md";
var mkdocs_page_url = null;
</script>
<script src="../../js/jquery-2.1.1.min.js" defer></script>
<script src="../../js/modernizr-2.8.3.min.js" defer></script>
<script src="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
</head>
<body class="wy-body-for-nav" role="document">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
<div class="wy-side-nav-search">
<a href="../.." class="icon icon-home"> Pytorch lightning Documentation</a>
<div role="search">
<form id ="rtd-search-form" class="wy-form" action="../../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" title="Type search term here" />
</form>
</div>
</div>
<div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<ul class="current">
<li class="toctree-l1">
<a class="" href="../..">PYTORCH-LIGHTNING DOCUMENTATION</a>
</li>
<li class="toctree-l1">
<span class="caption-text">LightningModule</span>
<ul class="subnav">
<li class=" current">
<a class="current" href="./">Lightning Module interface</a>
<ul class="subnav">
<li class="toctree-l3"><a href="#lightning-module-interface">Lightning Module interface</a></li>
<ul>
<li><a class="toctree-l4" href="#training_step">training_step</a></li>
<li><a class="toctree-l4" href="#validation_step">validation_step</a></li>
<li><a class="toctree-l4" href="#validation_end">validation_end</a></li>
<li><a class="toctree-l4" href="#configure_optimizers">configure_optimizers</a></li>
<li><a class="toctree-l4" href="#get_save_dict">get_save_dict</a></li>
<li><a class="toctree-l4" href="#load_model_specific">load_model_specific</a></li>
<li><a class="toctree-l4" href="#tng_dataloader">tng_dataloader</a></li>
<li><a class="toctree-l4" href="#val_dataloader">val_dataloader</a></li>
<li><a class="toctree-l4" href="#test_dataloader">test_dataloader</a></li>
<li><a class="toctree-l4" href="#update_tng_log_metrics">update_tng_log_metrics</a></li>
<li><a class="toctree-l4" href="#add_model_specific_args">add_model_specific_args</a></li>
</ul>
</ul>
</li>
<li class="">
<a class="" href="../methods/">Methods</a>
</li>
</ul>
</li>
<li class="toctree-l1">
<span class="caption-text">Trainer</span>
<ul class="subnav">
<li class="">
<a class="" href="../../Trainer/">Trainer</a>
</li>
<li class="">
<a class="" href="../../Trainer/Checkpointing/">Checkpointing</a>
</li>
<li class="">
<a class="" href="../../Trainer/Distributed training/">Distributed training</a>
</li>
<li class="">
<a class="" href="../../Trainer/Logging/">Logging</a>
</li>
<li class="">
<a class="" href="../../Trainer/SLURM Managed Cluster/">SLURM Managed Cluster</a>
</li>
<li class="">
<a class="" href="../../Trainer/Training Loop/">Training Loop</a>
</li>
<li class="">
<a class="" href="../../Trainer/Validation loop/">Validation loop</a>
</li>
<li class="">
<a class="" href="../../Trainer/debugging/">Debugging</a>
</li>
<li class="">
<a class="" href="../../Trainer/hooks/">Hooks</a>
</li>
</ul>
</li>
</ul>
</div>
&nbsp;
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
<nav class="wy-nav-top" role="navigation" aria-label="top navigation">
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../..">Pytorch lightning Documentation</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="wy-breadcrumbs">
<li><a href="../..">Docs</a> &raquo;</li>
<li>LightningModule &raquo;</li>
<li>Lightning Module interface</li>
<li class="wy-breadcrumbs-aside">
<a href="https://github.com/williamFalcon/pytorch-lightning/edit/master/docs/LightningModule/RequiredTrainerInterface.md"
class="icon icon-github"> Edit on GitHub</a>
</li>
</ul>
<hr/>
</div>
<div role="main">
<div class="section">
<h1 id="lightning-module-interface">Lightning Module interface</h1>
<p>[<a href="https://github.com/williamFalcon/pytorch-lightning/blob/master/pytorch_lightning/root_module/root_module.py">Github Code</a>]</p>
<p>A lightning module is a strict superclass of nn.Module, it provides a standard interface for the trainer to interact with the model.</p>
<p>The easiest thing to do is copy <a href="../../examples/new_project_templates/lightning_module_template.py">this template</a> and modify accordingly. </p>
<p>Otherwise, to Define a Lightning Module, implement the following methods:</p>
<p><strong>Required</strong>: </p>
<ul>
<li><a href="./#training_step">training_step</a> </li>
<li><a href="./#validation_step">validation_step</a></li>
<li>
<p><a href="./#validation_end">validation_end</a></p>
</li>
<li>
<p><a href="./#configure_optimizers">configure_optimizers</a></p>
</li>
<li><a href="./#get_save_dict">get_save_dict</a></li>
<li>
<p><a href="./#load_model_specific">load_model_specific</a></p>
</li>
<li>
<p><a href="./#tng_dataloader">tng_dataloader</a></p>
</li>
<li><a href="./#tng_dataloader">tng_dataloader</a></li>
<li><a href="./#test_dataloader">test_dataloader</a></li>
</ul>
<p><strong>Optional</strong>: </p>
<ul>
<li><a href="./#update_tng_log_metrics">update_tng_log_metrics</a></li>
<li><a href="./#add_model_specific_args">add_model_specific_args</a></li>
</ul>
<hr />
<h3 id="training_step">training_step</h3>
<pre><code class="python">def training_step(self, data_batch, batch_nb)
</code></pre>
<p>In this step you'd normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something specific to your model.</p>
<p><strong>Params</strong> </p>
<table>
<thead>
<tr>
<th>Param</th>
<th>description</th>
</tr>
</thead>
<tbody>
<tr>
<td>data_batch</td>
<td>The output of your dataloader. A tensor, tuple or list</td>
</tr>
<tr>
<td>batch_nb</td>
<td>Integer displaying which batch this is</td>
</tr>
</tbody>
</table>
<p><strong>Return</strong> </p>
<p>Dictionary or OrderedDict </p>
<table>
<thead>
<tr>
<th>key</th>
<th>value</th>
<th>is required</th>
</tr>
</thead>
<tbody>
<tr>
<td>loss</td>
<td>tensor scalar</td>
<td>Y</td>
</tr>
<tr>
<td>prog</td>
<td>Dict for progress bar display. Must have only tensors</td>
<td>N</td>
</tr>
</tbody>
</table>
<p><strong>Example</strong></p>
<pre><code class="python">def training_step(self, data_batch, batch_nb):
x, y, z = data_batch
# implement your own
out = self.forward(x)
loss = self.loss(out, x)
output = {
'loss': loss, # required
'prog': {'tng_loss': loss, 'batch_nb': batch_nb} # optional
}
# return a dict
return output
</code></pre>
<hr />
<h3 id="validation_step">validation_step</h3>
<pre><code class="python">def validation_step(self, data_batch, batch_nb)
</code></pre>
<p>In this step you'd normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something specific to your model.
This is most likely the same as your training_step. But unlike training step, the outputs from here will go to validation_end for collation.</p>
<p><strong>Params</strong> </p>
<table>
<thead>
<tr>
<th>Param</th>
<th>description</th>
</tr>
</thead>
<tbody>
<tr>
<td>data_batch</td>
<td>The output of your dataloader. A tensor, tuple or list</td>
</tr>
<tr>
<td>batch_nb</td>
<td>Integer displaying which batch this is</td>
</tr>
</tbody>
</table>
<p><strong>Return</strong> </p>
<table>
<thead>
<tr>
<th>Return</th>
<th>description</th>
<th>optional</th>
</tr>
</thead>
<tbody>
<tr>
<td>dict</td>
<td>Dict of OrderedDict with metrics to display in progress bar. All keys must be tensors.</td>
<td>Y</td>
</tr>
</tbody>
</table>
<p><strong>Example</strong></p>
<pre><code class="python">def validation_step(self, data_batch, batch_nb):
x, y, z = data_batch
# implement your own
out = self.forward(x)
loss = self.loss(out, x)
# calculate acc
labels_hat = torch.argmax(out, dim=1)
val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)
# all optional...
# return whatever you need for the collation function validation_end
output = OrderedDict({
'val_loss': loss_val,
'val_acc': torch.tensor(val_acc), # everything must be a tensor
})
# return an optional dict
return output
</code></pre>
<hr />
<h3 id="validation_end">validation_end</h3>
<pre><code class="python">def validation_end(self, outputs)
</code></pre>
<p>Called at the end of the validation loop with the output of each validation_step.</p>
<p><strong>Params</strong> </p>
<table>
<thead>
<tr>
<th>Param</th>
<th>description</th>
</tr>
</thead>
<tbody>
<tr>
<td>outputs</td>
<td>List of outputs you defined in validation_step</td>
</tr>
</tbody>
</table>
<p><strong>Return</strong> </p>
<table>
<thead>
<tr>
<th>Return</th>
<th>description</th>
<th>optional</th>
</tr>
</thead>
<tbody>
<tr>
<td>dict</td>
<td>Dict of OrderedDict with metrics to display in progress bar</td>
<td>Y</td>
</tr>
</tbody>
</table>
<p><strong>Example</strong></p>
<pre><code class="python">def validation_end(self, outputs):
&quot;&quot;&quot;
Called at the end of validation to aggregate outputs
:param outputs: list of individual outputs of each validation step
:return:
&quot;&quot;&quot;
val_loss_mean = 0
val_acc_mean = 0
for output in outputs:
val_loss_mean += output['val_loss']
val_acc_mean += output['val_acc']
val_loss_mean /= len(outputs)
val_acc_mean /= len(outputs)
tqdm_dic = {'val_loss': val_loss_mean.item(), 'val_acc': val_acc_mean.item()}
return tqdm_dic
</code></pre>
<hr />
<h3 id="configure_optimizers">configure_optimizers</h3>
<pre><code class="python">def configure_optimizers(self)
</code></pre>
<p>Set up as many optimizers as you need. Normally you'd need one. But in the case of GANs or something more esoteric you might have multiple.
Lightning will call .backward() and .step() on each one. If you use 16 bit precision it will also handle that.</p>
<h5 id="return">Return</h5>
<p>List - List of optimizers</p>
<p><strong>Example</strong></p>
<pre><code class="python"># most cases
def configure_optimizers(self):
opt = Adam(lr=0.01)
return [opt]
# gan example
def configure_optimizers(self):
generator_opt = Adam(lr=0.01)
disriminator_opt = Adam(lr=0.02)
return [generator_opt, disriminator_opt]
</code></pre>
<hr />
<h3 id="get_save_dict">get_save_dict</h3>
<pre><code class="python">def get_save_dict(self)
</code></pre>
<p>Called by lightning to checkpoint your model. Lightning saves current epoch, current batch nb, etc...
All you have to return is what specifically about your lightning model you want to checkpoint.</p>
<h5 id="return_1">Return</h5>
<p>Dictionary - No required keys. Most of the time as described in this example. </p>
<p><strong>Example</strong></p>
<pre><code class="python">def get_save_dict(self):
# 99% of use cases this is all you need to return
checkpoint = {'state_dict': self.state_dict()}
return checkpoint
</code></pre>
<hr />
<h3 id="load_model_specific">load_model_specific</h3>
<pre><code class="python">def load_model_specific(self, checkpoint)
</code></pre>
<p>Called by lightning to restore your model. This is your chance to restore your model using the keys you added in get_save_dict.
Lightning will automatically restore current epoch, batch nb, etc. </p>
<h5 id="return_2">Return</h5>
<p>Nothing </p>
<p><strong>Example</strong></p>
<pre><code class="python">def load_model_specific(self, checkpoint):
# you defined 'state_dict' in get_save_dict()
self.load_state_dict(checkpoint['state_dict'])
</code></pre>
<hr />
<h3 id="tng_dataloader">tng_dataloader</h3>
<pre><code class="python">@property
def tng_dataloader(self)
</code></pre>
<p>Called by lightning during training loop. Define it as a property.</p>
<h5 id="return_3">Return</h5>
<p>Pytorch DataLoader</p>
<p><strong>Example</strong></p>
<pre><code class="python">@property
def tng_dataloader(self):
if self._tng_dataloader is None:
try:
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
dataset = MNIST(root='/path/to/mnist/', train=True, transform=transform, download=True)
loader = torch.utils.data.DataLoader(
dataset=dataset,
batch_size=self.hparams.batch_size,
shuffle=True
)
self._tng_dataloader = loader
except Exception as e:
raise e
return self._tng_dataloader
</code></pre>
<hr />
<h3 id="val_dataloader">val_dataloader</h3>
<pre><code class="python">@property
def tng_dataloader(self)
</code></pre>
<p>Called by lightning during validation loop. Define it as a property.</p>
<h5 id="return_4">Return</h5>
<p>Pytorch DataLoader</p>
<p><strong>Example</strong></p>
<pre><code class="python">@property
def val_dataloader(self):
if self._val_dataloader is None:
try:
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
dataset = MNIST(root='/path/to/mnist/', train=False, transform=transform, download=True)
loader = torch.utils.data.DataLoader(
dataset=dataset,
batch_size=self.hparams.batch_size,
shuffle=True
)
self._val_dataloader = loader
except Exception as e:
raise e
return self._val_dataloader
</code></pre>
<hr />
<h3 id="test_dataloader">test_dataloader</h3>
<pre><code class="python">@property
def test_dataloader(self)
</code></pre>
<p>Called by lightning during test loop. Define it as a property.</p>
<h5 id="return_5">Return</h5>
<p>Pytorch DataLoader</p>
<p><strong>Example</strong></p>
<pre><code class="python">@property
def test_dataloader(self):
if self._test_dataloader is None:
try:
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (1.0,))])
dataset = MNIST(root='/path/to/mnist/', train=False, transform=transform, download=True)
loader = torch.utils.data.DataLoader(
dataset=dataset,
batch_size=self.hparams.batch_size,
shuffle=True
)
self._test_dataloader = loader
except Exception as e:
raise e
return self._test_dataloader
</code></pre>
<hr />
<h3 id="update_tng_log_metrics">update_tng_log_metrics</h3>
<pre><code class="python">def update_tng_log_metrics(self, logs)
</code></pre>
<p>Called by lightning right before it logs metrics for this batch.
This is a chance to ammend or add to the metrics about to be logged.</p>
<h5 id="return_6">Return</h5>
<p>Dict </p>
<p><strong>Example</strong></p>
<pre><code class="python">def update_tng_log_metrics(self, logs):
# modify or add to logs
return logs
</code></pre>
<hr />
<h3 id="add_model_specific_args">add_model_specific_args</h3>
<pre><code class="python">@staticmethod
def add_model_specific_args(parent_parser, root_dir)
</code></pre>
<p>Lightning has a list of default argparse commands.
This method is your chance to add or modify commands specific to your model.
The argument parser is available anywhere in your model by calling self.hparams</p>
<h5 id="return_7">Return</h5>
<p>An argument parser</p>
<p><strong>Example</strong></p>
<pre><code class="python">@staticmethod
def add_model_specific_args(parent_parser, root_dir):
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('--out_features', default=10)
parser.add_argument('--hidden_dim', default=50000) # use 500 for CPU, 50000 for GPU to see speed difference
# data
parser.add_argument('--data_root', default=os.path.join(root_dir, '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
</code></pre>
</div>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="../methods/" class="btn btn-neutral float-right" title="Methods">Next <span class="icon icon-circle-arrow-right"></span></a>
<a href="../.." class="btn btn-neutral" title="PYTORCH-LIGHTNING DOCUMENTATION"><span class="icon icon-circle-arrow-left"></span> Previous</a>
</div>
<hr/>
<div role="contentinfo">
<!-- Copyright etc -->
</div>
Built with <a href="http://www.mkdocs.org">MkDocs</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<div class="rst-versions" role="note" style="cursor: pointer">
<span class="rst-current-version" data-toggle="rst-current-version">
<a href="https://github.com/williamFalcon/pytorch-lightning/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
<span><a href="../.." style="color: #fcfcfc;">&laquo; Previous</a></span>
<span style="margin-left: 15px"><a href="../methods/" style="color: #fcfcfc">Next &raquo;</a></span>
</span>
</div>
<script>var base_url = '../..';</script>
<script src="../../js/theme.js" defer></script>
<script src="../../search/main.js" defer></script>
</body>
</html>
LightningModule/methods/index.html
+274
View File
@@ -0,0 +1,274 @@
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="shortcut icon" href="../../img/favicon.ico">
<title>Methods - Pytorch lightning Documentation</title>
<link href='https://fonts.googleapis.com/css?family=Lato:400,700|Roboto+Slab:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>
<link rel="stylesheet" href="../../css/theme.css" type="text/css" />
<link rel="stylesheet" href="../../css/theme_extra.css" type="text/css" />
<link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
<script>
// Current page data
var mkdocs_page_name = "Methods";
var mkdocs_page_input_path = "LightningModule/methods.md";
var mkdocs_page_url = null;
</script>
<script src="../../js/jquery-2.1.1.min.js" defer></script>
<script src="../../js/modernizr-2.8.3.min.js" defer></script>
<script src="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
</head>
<body class="wy-body-for-nav" role="document">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
<div class="wy-side-nav-search">
<a href="../.." class="icon icon-home"> Pytorch lightning Documentation</a>
<div role="search">
<form id ="rtd-search-form" class="wy-form" action="../../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" title="Type search term here" />
</form>
</div>
</div>
<div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<ul class="current">
<li class="toctree-l1">
<a class="" href="../..">PYTORCH-LIGHTNING DOCUMENTATION</a>
</li>
<li class="toctree-l1">
<span class="caption-text">LightningModule</span>
<ul class="subnav">
<li class="">
<a class="" href="../RequiredTrainerInterface/">Lightning Module interface</a>
</li>
<li class=" current">
<a class="current" href="./">Methods</a>
<ul class="subnav">
<li class="toctree-l3"><a href="#freeze">freeze</a></li>
<li class="toctree-l3"><a href="#load_from_metrics">load_from_metrics</a></li>
<li class="toctree-l3"><a href="#unfreeze">unfreeze</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l1">
<span class="caption-text">Trainer</span>
<ul class="subnav">
<li class="">
<a class="" href="../../Trainer/">Trainer</a>
</li>
<li class="">
<a class="" href="../../Trainer/Checkpointing/">Checkpointing</a>
</li>
<li class="">
<a class="" href="../../Trainer/Distributed training/">Distributed training</a>
</li>
<li class="">
<a class="" href="../../Trainer/Logging/">Logging</a>
</li>
<li class="">
<a class="" href="../../Trainer/SLURM Managed Cluster/">SLURM Managed Cluster</a>
</li>
<li class="">
<a class="" href="../../Trainer/Training Loop/">Training Loop</a>
</li>
<li class="">
<a class="" href="../../Trainer/Validation loop/">Validation loop</a>
</li>
<li class="">
<a class="" href="../../Trainer/debugging/">Debugging</a>
</li>
<li class="">
<a class="" href="../../Trainer/hooks/">Hooks</a>
</li>
</ul>
</li>
</ul>
</div>
&nbsp;
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
<nav class="wy-nav-top" role="navigation" aria-label="top navigation">
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../..">Pytorch lightning Documentation</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="wy-breadcrumbs">
<li><a href="../..">Docs</a> &raquo;</li>
<li>LightningModule &raquo;</li>
<li>Methods</li>
<li class="wy-breadcrumbs-aside">
<a href="https://github.com/williamFalcon/pytorch-lightning/edit/master/docs/LightningModule/methods.md"
class="icon icon-github"> Edit on GitHub</a>
</li>
</ul>
<hr/>
</div>
<div role="main">
<div class="section">
<p>Lightning modules are strict superclasses of torch.nn.Module. A LightningModule offers the following in addition to that API.</p>
<hr />
<h3 id="freeze">freeze</h3>
<p>Freeze all params for inference</p>
<pre><code class="python">model = MyLightningModule(...)
model.freeze()
</code></pre>
<hr />
<h3 id="load_from_metrics">load_from_metrics</h3>
<p>This is the easiest/fastest way which uses the meta_tags.csv file from test-tube to rebuild the model.
The meta_tags.csv file can be found in the test-tube experiment save_dir. </p>
<pre><code class="python">pretrained_model = MyLightningModule.load_from_metrics(
weights_path='/path/to/pytorch_checkpoint.ckpt',
tags_csv='/path/to/test_tube/experiment/version/meta_tags.csv',
on_gpu=True,
map_location=None
)
# predict
pretrained_model.freeze()
y_hat = pretrained_model(x)
</code></pre>
<p><strong>Params</strong> </p>
<table>
<thead>
<tr>
<th>Param</th>
<th>description</th>
</tr>
</thead>
<tbody>
<tr>
<td>weights_path</td>
<td>Path to a pytorch checkpoint</td>
</tr>
<tr>
<td>tags_csv</td>
<td>Path to meta_tags.csv file generated by the test-tube Experiment</td>
</tr>
<tr>
<td>on_gpu</td>
<td>if True, puts model on GPU. Make sure to use transforms option if model devices have changed</td>
</tr>
<tr>
<td>map_location</td>
<td>A dictionary mapping saved weight GPU devices to new GPU devices</td>
</tr>
</tbody>
</table>
<p><strong>Returns</strong> </p>
<p>LightningModule - The pretrained LightningModule</p>
<hr />
<h3 id="unfreeze">unfreeze</h3>
<p>Unfreeze all params for inference</p>
<pre><code class="python">model = MyLightningModule(...)
model.unfreeze()
</code></pre>
</div>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="../../Trainer/" class="btn btn-neutral float-right" title="Trainer">Next <span class="icon icon-circle-arrow-right"></span></a>
<a href="../RequiredTrainerInterface/" class="btn btn-neutral" title="Lightning Module interface"><span class="icon icon-circle-arrow-left"></span> Previous</a>
</div>
<hr/>
<div role="contentinfo">
<!-- Copyright etc -->
</div>
Built with <a href="http://www.mkdocs.org">MkDocs</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<div class="rst-versions" role="note" style="cursor: pointer">
<span class="rst-current-version" data-toggle="rst-current-version">
<a href="https://github.com/williamFalcon/pytorch-lightning/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
<span><a href="../RequiredTrainerInterface/" style="color: #fcfcfc;">&laquo; Previous</a></span>
<span style="margin-left: 15px"><a href="../../Trainer/" style="color: #fcfcfc">Next &raquo;</a></span>
</span>
</div>
<script>var base_url = '../..';</script>
<script src="../../js/theme.js" defer></script>
<script src="../../search/main.js" defer></script>
</body>
</html>