From 37c2e1e0446e8e2f92700eb38614b6107ac9bf8d Mon Sep 17 00:00:00 2001
From: reuvenp <reuvenp@altair-semi.com>
Date: Tue, 14 Jan 2025 19:05:10 +0200
Subject: [PATCH] fix tpc creation after refactor
---
.../keras/example_keras_pruning_mnist.ipynb | 32 ++++++++++---------
1 file changed, 17 insertions(+), 15 deletions(-)
diff --git a/tutorials/notebooks/mct_features_notebooks/keras/example_keras_pruning_mnist.ipynb b/tutorials/notebooks/mct_features_notebooks/keras/example_keras_pruning_mnist.ipynb
index b1b594c81..8bcde15da 100644
--- a/tutorials/notebooks/mct_features_notebooks/keras/example_keras_pruning_mnist.ipynb
+++ b/tutorials/notebooks/mct_features_notebooks/keras/example_keras_pruning_mnist.ipynb
@@ -217,22 +217,23 @@
{
"cell_type": "code",
"source": [
- "from model_compression_toolkit.target_platform_capabilities.target_platform import Signedness\n",
- "tp = mct.target_platform\n",
+ "from mct_quantizers import QuantizationMethod\n",
+ "from model_compression_toolkit.target_platform_capabilities.schema.mct_current_schema import schema, TargetPlatformCapabilities, Signedness, \\\n",
+ " AttributeQuantizationConfig, OpQuantizationConfig\n",
"\n",
"simd_size = 1\n",
"\n",
"def get_tpc():\n",
" # Define the default weight attribute configuration\n",
- " default_weight_attr_config = tp.AttributeQuantizationConfig(\n",
- " weights_quantization_method=tp.QuantizationMethod.UNIFORM,\n",
+ " default_weight_attr_config = AttributeQuantizationConfig(\n",
+ " weights_quantization_method=QuantizationMethod.UNIFORM,\n",
" )\n",
"\n",
" # Define the OpQuantizationConfig\n",
- " default_config = tp.OpQuantizationConfig(\n",
+ " default_config = OpQuantizationConfig(\n",
" default_weight_attr_config=default_weight_attr_config,\n",
" attr_weights_configs_mapping={},\n",
- " activation_quantization_method=tp.QuantizationMethod.UNIFORM,\n",
+ " activation_quantization_method=QuantizationMethod.UNIFORM,\n",
" activation_n_bits=8,\n",
" supported_input_activation_n_bits=8,\n",
" enable_activation_quantization=False,\n",
@@ -240,18 +241,19 @@
" fixed_scale=None,\n",
" fixed_zero_point=None,\n",
" simd_size=simd_size,\n",
- " signedness=Signedness.AUTO\n",
+ " signedness=schema.Signedness.AUTO\n",
" )\n",
+ " \n",
+ " # In this tutorial, we will use the default OpQuantizationConfig for all operator sets.\n",
+ " operator_set=[]\n",
"\n",
" # Create the quantization configuration options and model\n",
- " default_configuration_options = tp.QuantizationConfigOptions([default_config])\n",
- " tp_model = tp.TargetPlatformCapabilities(default_configuration_options,\n",
- " tpc_minor_version=1,\n",
- " tpc_patch_version=0,\n",
- " tpc_platform_type=\"custom_pruning_notebook_tpc\")\n",
- "\n",
- " # Return the target platform capabilities\n",
- " tpc = tp.FrameworkQuantizationCapabilities(tp_model)\n",
+ " default_configuration_options = schema.QuantizationConfigOptions(quantization_configurations=tuple([default_config]))\n",
+ " tpc = TargetPlatformCapabilities(default_qco=default_configuration_options,\n",
+ " tpc_minor_version=1,\n",
+ " tpc_patch_version=0,\n",
+ " tpc_platform_type=\"custom_pruning_notebook_tpc\",\n",
+ " operator_set=tuple(operator_set))\n",
" return tpc\n"
],
"metadata": {