Added docker stuff to run on Sinzlab server

.gitignore

@@ -86,7 +86,7 @@ target/
 celerybeat-schedule
 
 # dotenv
-.env
+#.env
 
 # virtualenv
 venv/
@@ -97,3 +97,6 @@ ENV/
 
 # Rope project settings
 .ropeproject
+
+#misc
+.history.*

docker-compose.yml

@@ -0,0 +1,48 @@
+version: "2.3"
+services:
+  local_debug:
+    build: ..
+    volumes:
+      - ./notebooks/models:/notebooks/models
+      - ./notebooks:/notebooks
+      - ../ml-utils:/transfer_learning/ml-utils
+      - ../transfer_learning:/projects/transfer_learning
+      - /tmp/djcache:/cache
+    env_file: .env
+    entrypoint: ""
+  local_develop:
+    build: ..
+    volumes:
+      - ./notebooks/models:/notebooks/models
+      - ./notebooks:/notebooks
+      - ../ml-utils:/transfer_learning/ml-utils
+      - ../transfer_learning:/projects/transfer_learning
+      - /tmp/djcache:/cache
+    ports:
+      - "8888:8888"
+    env_file: .env
+  server_develop:
+    build: ..
+    volumes:
+      - ./notebooks/models:/notebooks/models
+      - ./notebooks:/notebooks
+      - ../ml-utils:/transfer_learning/ml-utils
+      - ../transfer_learning:/projects/transfer_learning
+      - /tmp/djcache:/cache
+    ports:
+      - "6001:8888"
+    env_file: .env
+    runtime: nvidia
+  server_job:
+    build: ..
+    volumes:
+      - ./notebooks/models:/notebooks/models
+      - ./notebooks:/notebooks
+      - ../ml-utils:/transfer_learning/ml-utils
+      - ../transfer_learning:/projects/transfer_learning
+      - /tmp/djcache:/cache
+    env_file: .env
+    runtime: nvidia
+    entrypoint: /usr/bin/python3
+    command: run.py
+

setup.py

@@ -0,0 +1,11 @@
+from setuptools import setup
+
+setup(
+   name='transfer_learning',
+   version='0.1dev',
+   description='Experiments about inductive bias transfer',
+   author='Arne Nix',
+   author_email='[email protected]',
+   packages=['transfer_learning'],  #same as name
+   install_requires=[], #external packages as dependencies
+)