Add Stable Diffusion int8 example

examples/.config/model_params_pytorch_3x.json

@@ -228,6 +228,22 @@
       "main_script": "main.py",
       "batch_size": 1
     },
+    "sd21_static_int8":{
+      "model_src_dir": "diffusion_model/diffusers/stable_diffusion/static_quant",
+      "dataset_location": "/tf_dataset2/datasets/coco2017/coco/",
+      "input_model": "",
+      "main_script": "main.py",
+      "batch_size": 1,
+      "iters": 10
+    },
+    "lcm_static_int8":{
+      "model_src_dir": "diffusion_model/diffusers/stable_diffusion/static_quant",
+      "dataset_location": "/tf_dataset2/datasets/coco2017/coco/",
+      "input_model": "",
+      "main_script": "main.py",
+      "batch_size": 1,
+      "iters": 10
+    },
     "resnet18_mixed_precision": {
       "model_src_dir": "cv/mixed_precision",
       "dataset_location": "/tf_dataset/pytorch/ImageNet/raw",

examples/3.x_api/pytorch/diffusion_model/diffusers/stable_diffusion/static_quant/README.md

@@ -0,0 +1,63 @@
+# Stable Diffusion
+
+Stable Diffusion quantization and inference best known configurations with static quant.
+
+## Model Information
+
+| **Framework** | **Model Repo** |
+|:-------------:|:-------------------------------------------------------------------:|
+|   PyTorch     |       https://huggingface.co/stabilityai/stable-diffusion-2-1       | 
+|   PyTorch     |       https://huggingface.co/SimianLuo/LCM_Dreamshaper_v7           |
+
+# Pre-Requisite
+
+### Datasets
+
+Download the 2017 [COCO dataset](https://cocodataset.org) using the `download_dataset.sh` script.
+Export the `DATASET_DIR` environment variable to specify the directory where the dataset will be downloaded. This environment variable will be used again when running training scripts.
+```
+export DATASET_DIR=<directory where the dataset will be saved>
+bash download_dataset.sh
+```
+
+# Quantization and Inference
+quantization
+```shell
+python main.py \
+    --model_name_or_path stabilityai/stable-diffusion-2-1 \
+    --dataset_path=${DATASET_DIR} \
+    --quantized_model_path=${INT8_MODEL} \
+    --compile_inductor \
+    --precision=int8-bf16 \
+    --calibration
+```
+inference
+```shell
+python main.py \
+    --model_name_or_path stabilityai/stable-diffusion-2-1 \
+    --dataset_path=${DATASET_DIR} \
+    --precision=int8-bf16 \
+    --benchmark \
+    -w 1 \
+    -i 10 \
+    --quantized_model_path=${INT8_MODEL} \ 
+    --compile_inductor
+```
+## FID evaluation
+We have also evaluated FID scores on COCO2017 validation dataset for BF16 model, mixture of BF16 and INT8 model. FID results are listed below.
+
+| Model                | BF16  | INT8+BF16 |
+|----------------------|-------|-----------|
+| stable-diffusion-2-1 | 27.8630 |  27.8618    |
+| SimianLuo/LCM_Dreamshaper_v7|42.1710| 42.3138|
+
+To evaluated FID score on COCO2017 validation dataset for mixture of BF16 and INT8 model, you can use below command.
+
+```bash
+python main.py \
+    --dataset_path=${DATASET_DIR} \
+    --precision=int8-bf16 \
+    --accuracy \
+    --quantized_model_path=${INT8_MODEL} \ 
+    --compile_inductor
+```

examples/3.x_api/pytorch/diffusion_model/diffusers/stable_diffusion/static_quant/download_dataset.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+#
+# Copyright (c) 2025 Intel Corporation
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+DATASET_DIR=${DATASET_DIR-$PWD}
+
+dir=$(pwd)
+mkdir ${DATASET_DIR}; cd ${DATASET_DIR}
+curl -O http://images.cocodataset.org/zips/val2017.zip; unzip val2017.zip
+curl -O http://images.cocodataset.org/annotations/annotations_trainval2017.zip; unzip annotations_trainval2017.zip
+cd $dir