From 76b6f40f5a3875dbc6bf7609f109dc44fcc6112d Mon Sep 17 00:00:00 2001
From: Yong Yue <yongyue@microsoft.com>
Date: Fri, 24 Jul 2026 10:39:29 +0800
Subject: [PATCH] recipe(google/owlvit-base-patch32): add verified CPU
 fp32/fp16 recipes

Add zero-shot-object-detection recipes for google/owlvit-base-patch32
on CPU (fp32 and fp16), verified through L2 numeric parity against
PyTorch (cosine > 0.9999 on all outputs).
---
 ...ero-shot-object-detection_fp16_config.json | 68 +++++++++++++++++++
 ...ero-shot-object-detection_fp32_config.json | 55 +++++++++++++++
 2 files changed, 123 insertions(+)
 create mode 100644 examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp16_config.json
 create mode 100644 examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp32_config.json

diff --git a/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp16_config.json b/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp16_config.json
new file mode 100644
index 000000000..01a68d450
--- /dev/null
+++ b/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp16_config.json
@@ -0,0 +1,68 @@
+{
+  "export": {
+    "opset_version": 17,
+    "batch_size": 1,
+    "export_params": true,
+    "do_constant_folding": true,
+    "verbose": false,
+    "dynamo": false,
+    "enable_hierarchy_tags": true,
+    "clean_onnx": false,
+    "hierarchy_tag_format": "full",
+    "input_tensors": [
+      {
+        "name": "pixel_values",
+        "dtype": "float32",
+        "shape": [1, 3, 768, 768],
+        "value_range": [0, 1]
+      },
+      {
+        "name": "input_ids",
+        "dtype": "int32",
+        "shape": [1, 16],
+        "value_range": [0, 49408]
+      },
+      {
+        "name": "attention_mask",
+        "dtype": "int32",
+        "shape": [1, 16],
+        "value_range": [0, 2]
+      }
+    ],
+    "output_tensors": [
+      {
+        "name": "logits"
+      },
+      {
+        "name": "pred_boxes"
+      },
+      {
+        "name": "text_embeds"
+      },
+      {
+        "name": "image_embeds"
+      }
+    ]
+  },
+  "optim": {},
+  "quant": {
+    "mode": "fp16",
+    "samples": 10,
+    "calibration_method": "minmax",
+    "weight_type": "uint8",
+    "activation_type": "uint16",
+    "per_channel": false,
+    "symmetric": false,
+    "task": "zero-shot-object-detection",
+    "model_id": "google/owlvit-base-patch32",
+    "model_type": "owlvit",
+    "fp16_keep_io_types": true,
+    "fp16_op_block_list": null
+  },
+  "compile": null,
+  "loader": {
+    "task": "zero-shot-object-detection",
+    "model_class": "AutoModelForZeroShotObjectDetection",
+    "model_type": "owlvit"
+  }
+}
diff --git a/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp32_config.json b/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp32_config.json
new file mode 100644
index 000000000..8b59a039d
--- /dev/null
+++ b/examples/recipes/google_owlvit-base-patch32/cpu/cpu/zero-shot-object-detection_fp32_config.json
@@ -0,0 +1,55 @@
+{
+  "export": {
+    "opset_version": 17,
+    "batch_size": 1,
+    "export_params": true,
+    "do_constant_folding": true,
+    "verbose": false,
+    "dynamo": false,
+    "enable_hierarchy_tags": true,
+    "clean_onnx": false,
+    "hierarchy_tag_format": "full",
+    "input_tensors": [
+      {
+        "name": "pixel_values",
+        "dtype": "float32",
+        "shape": [1, 3, 768, 768],
+        "value_range": [0, 1]
+      },
+      {
+        "name": "input_ids",
+        "dtype": "int32",
+        "shape": [1, 16],
+        "value_range": [0, 49408]
+      },
+      {
+        "name": "attention_mask",
+        "dtype": "int32",
+        "shape": [1, 16],
+        "value_range": [0, 2]
+      }
+    ],
+    "output_tensors": [
+      {
+        "name": "logits"
+      },
+      {
+        "name": "pred_boxes"
+      },
+      {
+        "name": "text_embeds"
+      },
+      {
+        "name": "image_embeds"
+      }
+    ]
+  },
+  "optim": {},
+  "quant": null,
+  "compile": null,
+  "loader": {
+    "task": "zero-shot-object-detection",
+    "model_class": "AutoModelForZeroShotObjectDetection",
+    "model_type": "owlvit"
+  }
+}