Fixed download error due to missing directories. Added docker docs. F…

…ixed package redownload due to TF package naming conventions.

README.md

@@ -52,6 +52,16 @@ pip install .
 ### (Alternatively) No install:
 If you do not want to install you could just drop the `efficientnet_lite/efficientnet_lite.py` file directly into your project.
 
+### Docker
+You can also install this package as an extension to official Tensorflow docker container:  
+
+Build: `docker build -t efficientnet_lite_keras .`  
+Run: `docker run -it --rm efficientnet_lite_keras`
+
+For GPU support or different TAG you can (for example) pass  
+`--build-arg IMAGE_TAG=2.5.0-gpu`  
+in build command.
+
 ### Verify installation
 If all goes well you should be able to import:  
 `from efficientnet_lite import *` 
@@ -107,7 +117,7 @@ description of Lite variants.
 ### (Optionally) Convert the weights
 I am hosting the converted weights on DropBox. If, for some reason, you wish to download and convert original weights yourself, I prepered the utility scripts: 
 1. `bash scripts/download_all_weights.sh`
-2. `bash scripts/conver_all_weights.sh`
+2. `bash scripts/convert_all_weights.sh`
 
 # Bibliography
 [1] [Original repository](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite)  

setup.py

@@ -1,8 +1,33 @@
+from pkg_resources import DistributionNotFound, get_distribution
 from setuptools import setup
 
 with open("README.md", encoding="utf-8") as f:
     long_description = "\n" + f.read()
 
+
+def _package_exists(name: str) -> bool:
+    """Check whether package is present in the system."""
+    try:
+        get_distribution(name)
+    except DistributionNotFound:
+        return False
+    else:
+        return True
+
+
+def _get_tensorflow_requirement():
+    """Avoid re-download and misdetection of package."""
+    lower = 2.2
+    upper = 2.6
+
+    if _package_exists("tensorflow-cpu"):
+        return [f"tensorflow-cpu>={lower},<{upper}"]
+    elif _package_exists("tensorflow-gpu"):
+        return [f"tensorflow-gpu>={lower},<{upper}"]
+    else:
+        return [f"tensorflow>={lower},<{upper}"]
+
+
 setup(
     name="efficientnet-lite-keras",
     version="1.0",
@@ -20,5 +45,5 @@
         "Programming Language :: Python :: 3",
     ],
     packages=["efficientnet_lite"],
-    install_requires=["tensorflow>=2.2,<2.6"],
+    install_requires=_get_tensorflow_requirement(),
 )