Skip to content

add stable diffusion from huggingface #877

add stable diffusion from huggingface

add stable diffusion from huggingface #877

Workflow file for this run

name: CI
on:
push:
branches: [ main ]
pull_request:
branches: [ main ]
concurrency:
group: ${{ github.ref }}
cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
jobs:
linter:
runs-on: ubuntu-latest
container: ubuntu:24.04
name: Flake8, intellectual property compliance
steps:
- name: Install deps
run:
apt-get update && apt-get install -y git python3-pip && pip3 install --break-system-packages flake8 urlextract
- name: Git checkout w/o submodules
uses: actions/checkout@v4
with:
submodules: false
- name: Lint with flake8
run:
python3 -m flake8
- name: Ensure runner files don't do imports in global scope and check if env checking codeblock prepended
run:
python3 -m unittest tests.test_imports
- name: Git checkout w/ submodules
uses: actions/checkout@v4
with:
fetch-depth: 0
submodules: True
- name: Check for intellectual property compliance
run: |
git config --global --add safe.directory $(pwd)
python3 -m unittest tests.test_ip
test_x86:
runs-on: ubuntu-latest
container: ubuntu:24.04
name: x86-64 - Ubuntu 24.04 - all frameworks (native)
env:
PYTHONPATH: ./
COCO_IMG_PATH: aio_objdet_dataset
COCO_ANNO_PATH: aio_objdet_dataset/annotations.json
OMP_NUM_THREADS: 4
S3_URL_CRITEO_DATASET: ${{ secrets.S3_URL_CRITEO_DATASET }}
S3_URL_RESNET_50_V15_TF_FP32: ${{ secrets.S3_URL_RESNET_50_V15_TF_FP32 }}
S3_URL_SSD_INCEPTION_V2_TF_FP32: ${{ secrets.S3_URL_SSD_INCEPTION_V2_TF_FP32 }}
S3_URL_ALPACA_PYTORCH_FP32: ${{ secrets.S3_URL_ALPACA_PYTORCH_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Install git
run:
apt-get update && apt-get install -y git
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run:
FORCE_INSTALL=1 bash setup_deb.sh
- name: Unittest
run: |
python3 -m unittest tests.test_pytorch_models
- name: End-user smoke test
run: |
wget https://ampereaimodelzoo.s3.eu-central-1.amazonaws.com/aio_objdet_dataset.tar.gz > /dev/null 2>&1
tar -xf aio_objdet_dataset.tar.gz > /dev/null
wget $S3_URL_RESNET_50_V15_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/resnet_50_v15/run.py -m resnet_50_v15_tf_fp32.pb -p fp32 -f tf --timeout=60
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/mobilenet_v2/run.py -p fp32 -f pytorch --timeout=60
wget https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/object_detection/yolo_v8/run.py -m yolov8n.pt -f pytorch -p fp32 --timeout=60
python3 speech_recognition/whisper/run.py -m small.en
wget $S3_URL_SSD_INCEPTION_V2_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/object_detection/ssd_inception_v2/run.py -m ssd_inception_v2_tf_fp32.pb -p fp32 --timeout=60
wget https://zenodo.org/records/4735647/files/resnet50_v1.onnx > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/resnet_50_v1/run.py -m resnet50_v1.onnx -p fp32 -f ort
wget https://s3.amazonaws.com/onnx-model-zoo/vgg/vgg16/vgg16.tar.gz > /dev/null 2>&1
tar -xf vgg16.tar.gz > /dev/null
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/vgg_16/run.py -m vgg16/vgg16.onnx -p fp32 -f ort
test_arm64:
runs-on: self-hosted
container:
image: oraclelinux:9
options: --memory=170g
name: ARM64 - Oracle Linux 9 - all frameworks (native)
env:
PYTHONPATH: ./
COCO_IMG_PATH: aio_objdet_dataset
COCO_ANNO_PATH: aio_objdet_dataset/annotations.json
OMP_NUM_THREADS: 32
S3_URL_CRITEO_DATASET: ${{ secrets.S3_URL_CRITEO_DATASET }}
S3_URL_RESNET_50_V15_TF_FP32: ${{ secrets.S3_URL_RESNET_50_V15_TF_FP32 }}
S3_URL_SSD_INCEPTION_V2_TF_FP32: ${{ secrets.S3_URL_SSD_INCEPTION_V2_TF_FP32 }}
S3_URL_ALPACA_PYTORCH_FP32: ${{ secrets.S3_URL_ALPACA_PYTORCH_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
S3_URL_COVOST2_DATASET: ${{ secrets.S3_URL_COVOST2_DATASET }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Install git
run:
yum install -y git
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run:
bash setup_rhel.sh
- name: Unittest
run: |
python3 -m unittest tests.test_pytorch_models
- name: End-user smoke test
run: |
wget https://ampereaimodelzoo.s3.eu-central-1.amazonaws.com/aio_objdet_dataset.tar.gz > /dev/null 2>&1
tar -xf aio_objdet_dataset.tar.gz > /dev/null
wget $S3_URL_RESNET_50_V15_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/resnet_50_v15/run.py -m resnet_50_v15_tf_fp32.pb -p fp32 -f tf --timeout=60
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/mobilenet_v2/run.py -p fp32 -f pytorch --timeout=60
wget https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/object_detection/yolo_v8/run.py -m yolov8n.pt -f pytorch -p fp32 --timeout=60
python3 speech_recognition/whisper/run.py -m small.en
wget $S3_URL_SSD_INCEPTION_V2_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/object_detection/ssd_inception_v2/run.py -m ssd_inception_v2_tf_fp32.pb -p fp32 --timeout=60
wget https://zenodo.org/records/4735647/files/resnet50_v1.onnx > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/resnet_50_v1/run.py -m resnet50_v1.onnx -p fp32 -f ort
wget https://s3.amazonaws.com/onnx-model-zoo/vgg/vgg16/vgg16.tar.gz > /dev/null 2>&1
tar -xf vgg16.tar.gz > /dev/null
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/vgg_16/run.py -m vgg16/vgg16.onnx -p fp32 -f ort
test_pytorch_arm64_sh:
if: false
runs-on: self-hosted
container:
image: ubuntu:22.04
options: --memory=170g
name: Ampere Altra - Ampere optimized PyTorch (shell installer)
env:
PYTHONPATH: ./
AIO_NUM_THREADS: 32
AIO_DEBUG_MODE: 0
S3_URL_CRITEO_DATASET: ${{ secrets.S3_URL_CRITEO_DATASET }}
S3_URL_ALPACA_PYTORCH_FP32: ${{ secrets.S3_URL_ALPACA_PYTORCH_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
S3_URL_COVOST2_DATASET: ${{ secrets.S3_URL_COVOST2_DATASET }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Install Ampere optimized PyTorch
run: |
apt-get update && apt-get install -y wget
bash -c "$(wget -qO- https://ampereaidevelopus.s3.amazonaws.com/releases/1.10.0/binaries/install_ampere_pytorch_u22_1_10_0.sh)"
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run:
bash setup_deb.sh
- name: Unittest
run: |
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 -m unittest tests.test_pytorch_models
- name: benchmark.py test
run: |
PYTHONPATH=/__w/ampere_model_library/ampere_model_library python3 benchmark.py --no-interactive --model resnet_50_v1.5
test_pytorch_arm64_docker:
runs-on: self-hosted
container:
image: amperecomputingai/pytorch:latest
options: --memory=170g
name: Ampere Altra - Ampere optimized PyTorch (Docker image)
env:
PYTHONPATH: ./
COCO_IMG_PATH: aio_objdet_dataset
COCO_ANNO_PATH: aio_objdet_dataset/annotations.json
AIO_NUM_THREADS: 32
AIO_DEBUG_MODE: 0
S3_URL_CRITEO_DATASET: ${{ secrets.S3_URL_CRITEO_DATASET }}
S3_URL_ALPACA_PYTORCH_FP32: ${{ secrets.S3_URL_ALPACA_PYTORCH_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
S3_URL_COVOST2_DATASET: ${{ secrets.S3_URL_COVOST2_DATASET }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run: |
bash setup_deb.sh
echo $HF_HUB_TOKEN > ~/.cache/huggingface/token
- name: Unittest
run: |
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 -m unittest tests.test_pytorch_models
- name: benchmark.py test
run: |
{ echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; echo "y"; } | PYTHONPATH=/__w/ampere_model_library/ampere_model_library python3 benchmark.py
# testing second time to ensure that left-over files don't interrupt, etc. - this time no-interactive mode
PYTHONPATH=/__w/ampere_model_library/ampere_model_library python3 benchmark.py --no-interactive --memory 30 --max-threads 24
- name: AML end-user smoke test
run: |
wget https://ampereaimodelzoo.s3.eu-central-1.amazonaws.com/aio_objdet_dataset.tar.gz > /dev/null 2>&1
tar -xf aio_objdet_dataset.tar.gz > /dev/null
wget https://github.com/tloen/alpaca-lora/raw/main/alpaca_data.json > /dev/null 2>&1
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 natural_language_processing/text_generation/llama2/run.py -m meta-llama/Llama-2-7b-chat-hf --dataset_path=alpaca_data.json
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 recommendation/dlrm_torchbench/run.py -p fp32
IGNORE_DATASET_LIMITS=1 AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 computer_vision/classification/resnet_50_v15/run.py -m resnet50 -p fp32 -b 16 -f pytorch
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 speech_recognition/whisper/run.py -m tiny.en
IGNORE_DATASET_LIMITS=1 python3 computer_vision/classification/mobilenet_v2/run.py -p fp32 -f pytorch --timeout=60
wget https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8l.pt > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 computer_vision/object_detection/yolo_v8/run.py -m yolov8l.pt -p fp32 -f pytorch
wget -O bert_large_mlperf.pt https://zenodo.org/records/3733896/files/model.pytorch?download=1 > /dev/null 2>&1
AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 natural_language_processing/extractive_question_answering/bert_large/run_mlperf.py -m bert_large_mlperf.pt -p fp32 -f pytorch
test_tensorflow_arm64:
runs-on: self-hosted
container:
image: amperecomputingai/tensorflow:latest
options: --memory=170g
name: Ampere Altra - Ampere optimized TensorFlow (Docker image)
env:
PYTHONPATH: ./
COCO_IMG_PATH: aio_objdet_dataset
COCO_ANNO_PATH: aio_objdet_dataset/annotations.json
AIO_NUM_THREADS: 32
AIO_DEBUG_MODE: 0
S3_URL_RESNET_50_V15_TF_FP32: ${{ secrets.S3_URL_RESNET_50_V15_TF_FP32 }}
S3_URL_SSD_INCEPTION_V2_TF_FP32: ${{ secrets.S3_URL_SSD_INCEPTION_V2_TF_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run: |
bash setup_deb.sh
echo $HF_HUB_TOKEN > ~/.cache/huggingface/token
- name: End-user smoke test
run: |
wget https://ampereaimodelzoo.s3.eu-central-1.amazonaws.com/aio_objdet_dataset.tar.gz > /dev/null 2>&1
tar -xf aio_objdet_dataset.tar.gz > /dev/null
# AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 natural_language_processing/extractive_question_answering/bert_large/run_huggingface.py -m bert-large-cased-whole-word-masking-finetuned-squad
wget $S3_URL_RESNET_50_V15_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 computer_vision/classification/resnet_50_v15/run.py -m resnet_50_v15_tf_fp32.pb -b 32 -p fp32 -f tf --timeout=60
wget $S3_URL_SSD_INCEPTION_V2_TF_FP32 > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 python3 computer_vision/object_detection/ssd_inception_v2/run.py -m ssd_inception_v2_tf_fp32.pb -b 8 -p fp32 --timeout=60
test_onnxrt_arm64:
runs-on: self-hosted
container:
image: amperecomputingai/onnxruntime:latest
options: --memory=170g
name: Ampere Altra - Ampere optimized ONNXRunTime (Docker image)
env:
PYTHONPATH: ./
COCO_IMG_PATH: aio_objdet_dataset
COCO_ANNO_PATH: aio_objdet_dataset/annotations.json
AIO_NUM_THREADS: 32
AIO_DEBUG_MODE: 0
S3_URL_RESNET_50_V15_TF_FP32: ${{ secrets.S3_URL_RESNET_50_V15_TF_FP32 }}
S3_URL_SSD_INCEPTION_V2_TF_FP32: ${{ secrets.S3_URL_SSD_INCEPTION_V2_TF_FP32 }}
S3_URL_IMAGENET_DATASET: ${{ secrets.S3_URL_IMAGENET_DATASET }}
S3_URL_IMAGENET_DATASET_LABELS: ${{ secrets.S3_URL_IMAGENET_DATASET_LABELS }}
S3_URL_COCO_DATASET: ${{ secrets.S3_URL_COCO_DATASET }}
S3_URL_COCO_DATASET_ANNOTATIONS: ${{ secrets.S3_URL_COCO_DATASET_ANNOTATIONS }}
HF_HUB_TOKEN: ${{ secrets.HF_HUB_TOKEN }}
steps:
- name: Git checkout & pull submodules
uses: actions/checkout@v4
with:
submodules: true
- name: Set up AML
run:
bash setup_deb.sh
- name: End-user smoke test
run: |
wget https://ampereaimodelzoo.s3.eu-central-1.amazonaws.com/aio_objdet_dataset.tar.gz > /dev/null 2>&1
tar -xvf aio_objdet_dataset.tar.gz > /dev/null
wget https://zenodo.org/records/4735647/files/resnet50_v1.onnx > /dev/null 2>&1
IGNORE_DATASET_LIMITS=1 AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 computer_vision/classification/resnet_50_v1/run.py -m resnet50_v1.onnx -p fp32 -f ort
wget https://s3.amazonaws.com/onnx-model-zoo/vgg/vgg16/vgg16.tar.gz > /dev/null 2>&1
tar -xf vgg16.tar.gz > /dev/null
IGNORE_DATASET_LIMITS=1 AIO_IMPLICIT_FP16_TRANSFORM_FILTER=".*" python3 computer_vision/classification/vgg_16/run.py -m vgg16/vgg16.onnx -p fp32 -f ort