From 9e6ce3b0d6e018f78d1bf785ca2c888aaf2f9fbc Mon Sep 17 00:00:00 2001
From: William Falcon <falc@fb.com>
Date: Sat, 31 Aug 2019 03:19:08 -0400
Subject: [PATCH] testing loop docs

---
 docs/Trainer/Testing loop.md | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
 create mode 100644 docs/Trainer/Testing loop.md

diff --git a/docs/Trainer/Testing loop.md b/docs/Trainer/Testing loop.md
new file mode 100644
index 0000000000..8424be20cb
--- /dev/null
+++ b/docs/Trainer/Testing loop.md	
@@ -0,0 +1,31 @@
+To ensure you don't accidentally use test data to guide training decisions Lightning makes running the test set deliberate.   
+
+---
+#### test
+You have two options to run the test set. 
+First case is where you test right after a full training routine.
+``` {.python}
+# run full training
+trainer.fit(model)
+
+# run test set
+trainer.test()
+```   
+
+Second case is where you load a model and run the test set   
+```{.python}
+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
+)
+
+# init trainer with whatever options
+trainer = Trainer(...)
+    
+# test (pass in the model)
+trainer.test(model)
+```
+In this second case, the options you pass to trainer will be used when running the test set (ie: 16-bit, dp, ddp, etc...)  
+