Merge pull request #221 from jonnor/install-instructions

Explain install as Python packages

README.md

@@ -65,8 +65,34 @@ However, the available NN libs for MCU are too low-level which make it sooooo di
 Therefore, we build NNoM to help embedded developers for faster and simpler deploying NN model directly to MCU. 
 > NNoM will manage the strucutre, memory and everything else for the developer. All you need to do is feeding your new measurements and getting the results. 
 
+## Installing
 
+NNoM can be installed as a Python package
 
+```
+pip install git+https://github.com/majianjia/nnom@master
+```
+
+NNoM requires [Tensorflow](https://www.tensorflow.org/) version `<= 2.14`.
+There are multiple options for how to install this, see the TensorFlow documentation.
+
+For example:
+```
+pip install 'tensorflow-cpu<=2.14.1'
+```
+
+NOTE: Tensorflow 2.14 supports up until Python 3.11.
+However, *Python 3.12 is not supported*.
+
+## Accessing C files
+
+The C headers and source code in NNoM are distributed in the `nnom_core` Python package.
+You can find its location by running the following command.
+```
+python -c "import nnom_core; print(nnom_core.__path__[0])"
+```
+In your build system, add the `inc/` and `port/` directories as include directories,
+and compile the the `src/*.c` files.
 
 ## Documentations
 

examples/auto_test/SConstruct

@@ -3,6 +3,10 @@ import os
 #if(not os.path.exists('CMSIS_5')):
 #    os.system('git clone https://github.com/ARM-software/CMSIS_5.git')
 
+# Specify the path to NNoM C files
+# When NNoM is installed as Python package, you can find this using
+# python -c "import nnom_core; print(nnom_core.__path__[0])"
+# Otherwise, use the path to the NNoM git repository root.
 ROOT=os.path.abspath('../..')
 
 env = Environment()

pyproject.toml

@@ -21,6 +21,7 @@ version = "0.5.0a1"
 
 dependencies = [
   "keras<3.0.0",
+  "scikit-learn<2.0.0",
 ]
 
 classifiers = [

scripts/nnom.py

@@ -11,7 +11,7 @@
 '''
 
 import sklearn.metrics as skmetrics
-import matplotlib.pyplot as plt
+
 import tensorflow as tf
 import tensorflow.keras.backend as K
 from tensorflow.keras import *
@@ -1137,6 +1137,9 @@ def q2f(d, Q):
     return d*2**-Q
 
 def show_weights(w, name):
+
+    import matplotlib.pyplot as plt
+
     sz = 1
     for s in w.shape:
         sz = sz*s
@@ -1167,6 +1170,9 @@ def show_weights(w, name):
     plt.show()
 
 def compare(a,b,name):
+
+    import matplotlib.pyplot as plt
+
     sz = 1
     for s in a.shape:
         sz = sz*s