diff --git a/notebooks/FINN-HowToWorkWithONNX.ipynb b/notebooks/FINN-HowToWorkWithONNX.ipynb
index 44a030c11c88f4084b7da2b765a6563efcb8dc8d..fa4e381a64ae00be03e476695a3aab0457dbec3f 100644
--- a/notebooks/FINN-HowToWorkWithONNX.ipynb
+++ b/notebooks/FINN-HowToWorkWithONNX.ipynb
@@ -31,7 +31,7 @@
},
{
"cell_type": "code",
- "execution_count": 1,
+ "execution_count": 39,
"metadata": {},
"outputs": [],
"source": [
@@ -41,6 +41,7 @@
" 'Add',\n",
" inputs=['in1', 'in2'],\n",
" outputs=['sum1'],\n",
+ " name='Add1'\n",
")"
]
},
@@ -55,7 +56,7 @@
},
{
"cell_type": "code",
- "execution_count": 81,
+ "execution_count": 40,
"metadata": {},
"outputs": [],
"source": [
@@ -63,24 +64,28 @@
" 'Add',\n",
" inputs=['sum1', 'in3'],\n",
" outputs=['sum2'],\n",
+ " name='Add2',\n",
")\n",
"\n",
"Add3_node = onnx.helper.make_node(\n",
" 'Add',\n",
" inputs=['abs1', 'abs1'],\n",
" outputs=['sum3'],\n",
+ " name='Add3',\n",
")\n",
"\n",
"Abs_node = onnx.helper.make_node(\n",
" 'Abs',\n",
" inputs=['sum2'],\n",
" outputs=['abs1'],\n",
+ " name='Abs'\n",
")\n",
"\n",
"Round_node = onnx.helper.make_node(\n",
" 'Round',\n",
" inputs=['sum3'],\n",
" outputs=['out1'],\n",
+ " name='Round',\n",
")\n"
]
},
@@ -93,7 +98,7 @@
},
{
"cell_type": "code",
- "execution_count": 82,
+ "execution_count": 41,
"metadata": {},
"outputs": [],
"source": [
@@ -114,7 +119,7 @@
},
{
"cell_type": "code",
- "execution_count": 83,
+ "execution_count": 42,
"metadata": {},
"outputs": [],
"source": [
@@ -149,7 +154,7 @@
},
{
"cell_type": "code",
- "execution_count": 84,
+ "execution_count": 43,
"metadata": {},
"outputs": [],
"source": [
@@ -166,7 +171,7 @@
},
{
"cell_type": "code",
- "execution_count": 85,
+ "execution_count": 44,
"metadata": {},
"outputs": [
{
@@ -186,7 +191,7 @@
},
{
"cell_type": "code",
- "execution_count": 86,
+ "execution_count": 45,
"metadata": {
"scrolled": true
},
@@ -220,7 +225,7 @@
},
{
"cell_type": "code",
- "execution_count": 92,
+ "execution_count": 46,
"metadata": {},
"outputs": [],
"source": [
@@ -243,7 +248,7 @@
},
{
"cell_type": "code",
- "execution_count": 93,
+ "execution_count": 47,
"metadata": {},
"outputs": [],
"source": [
@@ -261,7 +266,7 @@
},
{
"cell_type": "code",
- "execution_count": 94,
+ "execution_count": 48,
"metadata": {},
"outputs": [],
"source": [
@@ -280,7 +285,7 @@
},
{
"cell_type": "code",
- "execution_count": 95,
+ "execution_count": 49,
"metadata": {},
"outputs": [],
"source": [
@@ -299,7 +304,7 @@
},
{
"cell_type": "code",
- "execution_count": 96,
+ "execution_count": 50,
"metadata": {},
"outputs": [
{
@@ -307,16 +312,16 @@
"output_type": "stream",
"text": [
"The output of the ONNX model is: \n",
- "[[24. 2. 6. 2.]\n",
- " [ 5. 1. 9. 5.]\n",
- " [ 6. 1. 8. 8.]\n",
- " [14. 4. 0. 1.]]\n",
+ "[[ 4. 17. 24. 20.]\n",
+ " [ 4. 7. 13. 21.]\n",
+ " [ 4. 3. 3. 14.]\n",
+ " [ 9. 12. 0. 5.]]\n",
"\n",
"The output of the reference function is: \n",
- "[[24. 2. 6. 2.]\n",
- " [ 5. 1. 9. 5.]\n",
- " [ 6. 1. 8. 8.]\n",
- " [14. 4. 0. 1.]]\n",
+ "[[ 4. 17. 24. 20.]\n",
+ " [ 4. 7. 13. 21.]\n",
+ " [ 4. 3. 3. 14.]\n",
+ " [ 9. 12. 0. 5.]]\n",
"\n",
"The results are the same!\n"
]
@@ -360,12 +365,19 @@
},
{
"cell_type": "code",
- "execution_count": 104,
+ "execution_count": 57,
"metadata": {},
"outputs": [],
"source": [
"from finn.core.modelwrapper import ModelWrapper\n",
- "finn_model = ModelWrapper(onnx_model)"
+ "finn_model = ModelWrapper(onnx_model)\n",
+ "\n",
+ "def get_node_id(model):\n",
+ " node_index = {}\n",
+ " node_ind = 0\n",
+ " for node in model.graph.node:\n",
+ " node_index[node_ind] = node.name\n",
+ " node_ind += 1"
]
},
{
@@ -377,7 +389,7 @@
},
{
"cell_type": "code",
- "execution_count": 105,
+ "execution_count": 60,
"metadata": {},
"outputs": [],
"source": [
@@ -401,38 +413,23 @@
},
{
"cell_type": "code",
- "execution_count": 106,
+ "execution_count": 62,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
- "Found adder node: \n",
- "input: \"in1\"\n",
- "input: \"in2\"\n",
- "output: \"sum1\"\n",
- "op_type: \"Add\"\n",
- "\n",
- "Found adder node: \n",
- "input: \"sum1\"\n",
- "input: \"in3\"\n",
- "output: \"sum2\"\n",
- "op_type: \"Add\"\n",
- "\n",
- "Found adder node: \n",
- "input: \"abs1\"\n",
- "input: \"abs1\"\n",
- "output: \"sum3\"\n",
- "op_type: \"Add\"\n",
- "\n"
+ "Found adder node: Add1\n",
+ "Found adder node: Add2\n",
+ "Found adder node: Add3\n"
]
}
],
"source": [
"add_nodes = identify_adder_nodes(finn_model)\n",
"for node in add_nodes:\n",
- " print(\"Found adder node: \\n{}\".format(node))"
+ " print(\"Found adder node: {}\".format(node.name))"
]
},
{
@@ -444,7 +441,7 @@
},
{
"cell_type": "code",
- "execution_count": 107,
+ "execution_count": 63,
"metadata": {},
"outputs": [],
"source": [
@@ -475,7 +472,7 @@
},
{
"cell_type": "code",
- "execution_count": 108,
+ "execution_count": 64,
"metadata": {},
"outputs": [],
"source": [
@@ -492,34 +489,75 @@
},
{
"cell_type": "code",
- "execution_count": 113,
+ "execution_count": 68,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
- "[input: \"in1\"\n",
- "input: \"in2\"\n",
- "output: \"sum1\"\n",
- "op_type: \"Add\"\n",
- ", input: \"sum1\"\n",
- "input: \"in3\"\n",
- "output: \"sum2\"\n",
- "op_type: \"Add\"\n",
- "]\n"
+ "Found following pair that could be replaced by a sum node:\n",
+ "Add1\n",
+ "Add2\n"
]
}
],
"source": [
- "i = 0\n",
"for node in add_nodes:\n",
" add_pair = adder_pair(finn_model, node)\n",
" if len(add_pair) != 0:\n",
- " print(add_pair)\n",
- " i+=1"
+ " substitute_pair = add_pair\n",
+ " print(\"Found following pair that could be replaced by a sum node:\")\n",
+ " for node in add_pair:\n",
+ " print(node.name)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "<font size=\"3\">Now that the pair to be replaced has been identified, a sum node can be instantiated and inserted into the graph at the correct position. \n",
+ "\n",
+ "First of all, the inputs must be determined. For this the adder nodes inputs are used minus the input, which corresponds to the output of the other adder node.</font>"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 74,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "input_list = []\n",
+ "for i in range(len(substitute_pair)):\n",
+ " if i == 0:\n",
+ " for j in range(len(substitute_pair[i].input)):\n",
+ " if substitute_pair[i].input[j] != substitute_pair[i+1].output[0]:\n",
+ " input_list.append(substitute_pair[i].input[j])\n",
+ " else:\n",
+ " for j in range(len(substitute_pair[i].input)):\n",
+ " if substitute_pair[i].input[j] != substitute_pair[i-1].output[0]:\n",
+ " input_list.append(substitute_pair[i].input[j])\n",
+ " "
]
},
+ {
+ "cell_type": "code",
+ "execution_count": 73,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "['in1', 'in2', 'in3']"
+ ]
+ },
+ "execution_count": 73,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": []
+ },
{
"cell_type": "code",
"execution_count": null,