Add IPython extension (#8)

* First cut of a IPython extension

* Add ipython to requirements

* Specify IPython dep like in line_profiler

* Try to fix travis due to ipython

* Add call to y() in mlrun cell magic demo

* Make IPython an optional dependency

* Add IPython magic instructions to README

* Fix syntax error in setup.py

* Catch NameError instead of Exception when evaling user provided function names

* Make string quotes consistent

* Add --dump-profile and --return args to mlrun magic

* Add --gpu specifier to mlrun magic

* Make target_gpu kwarg compatible with python 2.7

* Improve demo.ipynb and add --quiet flag

* Rerun notebook

* Tidy up IPython extension source

* Add links to relevant sections from overview in README, add details of IPython support

* Add link to demo.ipynb from README.md

.gitignore

@@ -1,3 +1,116 @@
-__pycache__
-*.pyc
-*.egg-info
+#### joe made this: http://goel.io/joe
+
+#####=== IPythonNotebook ===#####
+# Temporary data
+.ipynb_checkpoints/
+
+#####=== Python ===#####
+
+# 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/
+*.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
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+#####=== JetBrains ===#####
+# Covers JetBrains IDEs: IntelliJ, RubyMine, PhpStorm, AppCode, PyCharm, CLion, Android Studio
+
+*.iml
+
+## Directory-based project format:
+.idea/
+# if you remove the above rule, at least ignore the following:
+
+# User-specific stuff:
+# .idea/workspace.xml
+# .idea/tasks.xml
+# .idea/dictionaries
+
+# Sensitive or high-churn files:
+# .idea/dataSources.ids
+# .idea/dataSources.xml
+# .idea/sqlDataSources.xml
+# .idea/dynamic.xml
+# .idea/uiDesigner.xml
+
+# Gradle:
+# .idea/gradle.xml
+# .idea/libraries
+
+# Mongo Explorer plugin:
+# .idea/mongoSettings.xml
+
+## File-based project format:
+*.ipr
+*.iws
+
+## Plugin-specific files:
+
+# IntelliJ
+/out/
+
+# mpeltonen/sbt-idea plugin
+.idea_modules/
+
+# JIRA plugin
+atlassian-ide-plugin.xml
+
+# Crashlytics plugin (for Android Studio and IntelliJ)
+com_crashlytics_export_strings.xml
+crashlytics.properties
+crashlytics-build.properties
+
+.ropeproject

.travis.yml

@@ -3,10 +3,19 @@ language: python
 matrix:
   include:
   - python: '2.7'
+    env:
+      IPYTHON_VERSION='5.8'
   - python: '3.5'
+    env:
+      IPYTHON_VERSION='7.9'
   - python: '3.6'
+    env:
+      IPYTHON_VERSION='7'
   - python: '3.7'
+    env:
+      IPYTHON_VERSION='7'
 install:
+- pip install IPython==$IPYTHON_VERSION
 - python setup.py install
 - pip install -r requirements.txt
 script:

README.md

@@ -6,10 +6,13 @@ pytorch_memlab
 
 A simple and accurate **CUDA** memory management laboratory for pytorch,
 it consists of different parts about the memory:
-  - A `line_profiler` style CUDA memory profiler with simple API.
-  - A reporter to inspect tensors occupying the CUDA memory.
-  - An interesting feature to temporarily move all the CUDA tensors into
-  CPU memory for courtesy, and of course the backward transferring.
+
+  - [A `line_profiler` style CUDA memory profiler with simple API.](#memory-profiler)
+  - [A reporter to inspect tensors occupying the CUDA memory.](#memory-reporter)
+  - [An interesting feature to temporarily move all the CUDA tensors into
+    CPU memory for courtesy, and of course the backward transferring.](#courtesy)
+  - [IPython support through `%mlrun`/`%%mlrun` line/cell magic
+    commands.](#ipython-support)
 
 Installation
 -----
@@ -130,6 +133,49 @@ func()
 
 More samples can be found in `test/test_line_profiler.py`
 
+### IPython support
+
+Make sure you have `IPython` installed, or have installed `pytorch-memlab` with
+`pip install pytorch-memlab[ipython]`.
+
+First, load the extension:
+
+```python
+%%load_ext pytorch_memlab
+```
+
+This makes the `%mlrun` and `%%mlrun` line/cell magics available for use. For
+example, in a new cell run the following to profile an entire cell
+
+```python
+%%mlrun -f func
+import torch
+from pytorch_memlab import profile, set_target_gpu
+def func():
+    net1 = torch.nn.Linear(1024, 1024).cuda(0)
+    set_target_gpu(1)
+    net2 = torch.nn.Linear(1024, 1024).cuda(1)
+    set_target_gpu(0)
+    net3 = torch.nn.Linear(1024, 1024).cuda(0)
+```
+
+Or you can invoke the profiler for a single statement on via the `%mlrun` cell
+magic.
+
+```python
+import torch
+from pytorch_memlab import profile, set_target_gpu
+def func(input_size):
+    net1 = torch.nn.Linear(input_size, 1024).cuda(0)
+%mlrun -f func func(2048)
+```
+
+See `%mlrun?` for help on what arguments are supported. You can set the GPU
+device to profile, dump profiling results to a file, and return the
+`LineProfiler` object for post-profile inspection.
+
+Find out more by checking out the [demo Jupyter notebook](./demo.ipynb)
+
 
 ### Memory Reporter