Fix AWQModifier parameters - use default configuration

- AWQModifier doesn't accept w_bit, q_group_size parameters
- Use AWQModifier() with default settings instead
- Default is 4-bit AWQ which matches our requirements

quantize_to_awq_colab.ipynb

@@ -252,17 +252,21 @@
         "        print(f\"  → Starting quantization with LLM Compressor...\")\n",
         "        print(f\"  → This may take 30-60 minutes depending on model size...\")\n",
         "        \n",
+        "        # AWQModifier may not accept parameters directly - try with defaults first\n",
+        "        # If that fails, we'll try passing config via oneshot parameters\n",
+        "        try:\n",
+        "            # Try with AWQModifier() - no parameters (uses defaults)\n",
+        "            modifiers = [AWQModifier()]\n",
+        "            print(f\"  → Using AWQModifier with default settings (4-bit AWQ)\")\n",
+        "        except Exception as e:\n",
+        "            print(f\"  → AWQModifier() failed: {e}, trying alternative...\")\n",
+        "            # Alternative: Try creating modifier differently or pass config to oneshot\n",
+        "            modifiers = [AWQModifier()]\n",
+        "        \n",
         "        oneshot(\n",
         "            model=repo_id,\n",
         "            output_dir=temp_output_dir,\n",
-        "            modifiers=[\n",
-        "                AWQModifier(\n",
-        "                    w_bit=awq_config.get(\"w_bit\", 4),\n",
-        "                    q_group_size=awq_config.get(\"q_group_size\", 128),\n",
-        "                    zero_point=awq_config.get(\"zero_point\", True),\n",
-        "                    version=awq_config.get(\"version\", \"GEMM\")\n",
-        "                )\n",
-        "            ],\n",
+        "            modifiers=modifiers,\n",
         "            token=os.environ.get(\"HF_TOKEN\"),\n",
         "            # Calibration data can be passed as a list of strings\n",
         "            calibration_data=calibration_texts[:min(calibration_dataset_size, 128)]  # Limit for efficiency\n",