Fix (notebooks): make requirements clearer for dynamic export

Xilinx · Mar 6, 2024 · 0fb2ea2 · 0fb2ea2
1 parent 1c7c530
commit 0fb2ea2
Showing 1 changed file with 13 additions and 13 deletions.
diff --git a/notebooks/ONNX_export_tutorial.ipynb b/notebooks/ONNX_export_tutorial.ipynb
@@ -213,7 +213,7 @@
        "        "
       ],
       "text/plain": [
-       "<IPython.lib.display.IFrame at 0x7f2a6afb6f50>"
+       "<IPython.lib.display.IFrame at 0x7fb62ae3fe50>"
       ]
      },
      "execution_count": 4,
@@ -331,7 +331,7 @@
        "        "
       ],
       "text/plain": [
-       "<IPython.lib.display.IFrame at 0x7f2a6afa9e90>"
+       "<IPython.lib.display.IFrame at 0x7fb734383710>"
       ]
      },
      "execution_count": 6,
@@ -460,7 +460,7 @@
        "        "
       ],
       "text/plain": [
-       "<IPython.lib.display.IFrame at 0x7f2a69d7ff10>"
+       "<IPython.lib.display.IFrame at 0x7fb629e8a010>"
       ]
      },
      "execution_count": 8,
@@ -605,7 +605,7 @@
        "        "
       ],
       "text/plain": [
-       "<IPython.lib.display.IFrame at 0x7f2a69d2e710>"
+       "<IPython.lib.display.IFrame at 0x7fb62ae37190>"
       ]
      },
      "execution_count": 10,
@@ -704,7 +704,7 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "2024-03-01 03:24:07.215804006 [W:onnxruntime:, graph.cc:1283 Graph] Initializer linear.bias appears in graph inputs and will not be treated as constant value/weight. This may prevent some of the graph optimizations, like const folding. Move it out of graph inputs if there is no need to override it, by either re-generating the model with latest exporter/converter or with the tool onnxruntime/tools/python/remove_initializer_from_input.py.\n"
+      "2024-03-06 02:12:47.492497092 [W:onnxruntime:, graph.cc:1283 Graph] Initializer linear.bias appears in graph inputs and will not be treated as constant value/weight. This may prevent some of the graph optimizations, like const folding. Move it out of graph inputs if there is no need to override it, by either re-generating the model with latest exporter/converter or with the tool onnxruntime/tools/python/remove_initializer_from_input.py.\n"
      ]
     }
    ],
@@ -842,18 +842,18 @@
    },
    "outputs": [
     {
-     "name": "stdout",
+     "name": "stderr",
      "output_type": "stream",
      "text": [
-      "True\n"
+      "/scratch/fabian/brevitas/src/brevitas/export/onnx/standard/manager.py:26: UserWarning: ONNX opset version set to 13, override with opset_version=\n",
+      "  warnings.warn(f\"ONNX opset version set to {DEFAULT_OPSET}, override with {ka}=\")\n"
      ]
     },
     {
-     "name": "stderr",
+     "name": "stdout",
      "output_type": "stream",
      "text": [
-      "/scratch/fabian/brevitas/src/brevitas/export/onnx/standard/manager.py:26: UserWarning: ONNX opset version set to 13, override with opset_version=\n",
-      "  warnings.warn(f\"ONNX opset version set to {DEFAULT_OPSET}, override with {ka}=\")\n"
+      "True\n"
      ]
     }
    ],
@@ -924,14 +924,14 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "You can also export dynamically quantized models to ONNX, but there are some limitations. Supported exports include:\n",
+    "You can also export dynamically quantized models to ONNX, but there are some limitations. The ONNX DynamicQuantizeLinear requires the following settings:\n",
     "- Asymmetric quantization (and therefore *unsigned*)\n",
     "- Min-max scaling\n",
     "- Rounding to nearest\n",
     "- Per tensor scaling\n",
     "- Bit width set to 8\n",
     "\n",
-    "Note that ONNX DynamicQuantizeLinear requires these specific requirements for quantizers. This is shown in the following example:"
+    "This is shown in the following example:"
    ]
   },
   {
@@ -993,7 +993,7 @@
        "        "
       ],
       "text/plain": [
-       "<IPython.lib.display.IFrame at 0x7efe085d8ad0>"
+       "<IPython.lib.display.IFrame at 0x7fb62856ccd0>"
       ]
      },
      "execution_count": 15,