Morph MORPH
SIGN IN SIGN UP
pytorch / examples UNCLAIMED

A set of examples around pytorch in Vision, Text, Reinforcement Learning, etc.

0 0 45 Python
Normal View History Raw
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
#!/bin/bash
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
#
# This script runs through the code in each of the python examples.
Fix typos & supplement descriptions
2022-01-20 23:19:54 +08:00
# The purpose is just as an integration test, not to actually train models in any meaningful way.
# For that reason, most of these set epochs = 1 and --dry-run.
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
#
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
# Optionally specify a comma separated list of examples to run. Can be run as:
# * To run all examples:
# ./run_python_examples.sh
# * To run few specific examples:
# ./run_python_examples.sh "dcgan,fast_neural_style"
#
# To test examples on CUDA accelerator, run as:
# USE_CUDA=True ./run_python_examples.sh
#
Use torch.accelerator API in Fast Neural Style example (#1327)
2025-04-28 12:46:10 -07:00
# To test examples on hardware accelerator (CUDA, MPS, XPU, etc.), run as:
# USE_ACCEL=True ./run_python_examples.sh
# NOTE: USE_ACCEL relies on torch.accelerator API and not all examples are converted
# to use it at the moment. Thus, expect failures using this flag on non-CUDA accelerators
# and consider to run examples one by one.
#
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
# Script requires uv to be installed. When executed, script will install prerequisites from
# `requirements.txt` for each example. If ran within activated virtual environment (uv venv,
# python -m venv, conda) this might reinstall some of the packages. To change pip installation
# index or to pass additional pip install options, run as:
# PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \
# ./run_python_examples.sh
#
# To force script to create virtual environment for each example, run as:
# VIRTUAL_ENV=".venv" ./run_python_examples.sh
# Script will remove environments it creates in a teardown step after execution of each example.
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
Update distributed example tests in `run_python_examples.sh` (#1250) * Fix distributed test * fix parallel scripts * install dill * remove dill * run 2 gpu * remove gpucount, use default * Add examples to distributed examples * refactor distributed test * fx ERRORS overwriting * run with base dir * remove distributed from run_python_examples.sh * move basedir to source * separate init --------- Co-authored-by: Sirut Buasai <sirutbuasai27@hotmail.com>
2024-05-03 13:25:49 -07:00
BASE_DIR="$(pwd)/$(dirname $0)"
source $BASE_DIR/utils.sh
Language translation example added (#1131) (#1240) * Intial model setup * Training works * Added inference * Code clean up and commenting * Update to README.md * Add requirements.txt * Updated top level README, added example to CI * Potentially fixed testing (maybe not enough memory?) * Update requirements.txt --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2024-04-02 12:50:04 -04:00
Use torch.accelerator API in Fast Neural Style example (#1327)
2025-04-28 12:46:10 -07:00
# TODO: Leave only USE_ACCEL and drop USE_CUDA once all examples will be converted
# to torch.accelerator API. For now, just add USE_ACCEL as an alias for USE_CUDA.
if [ -n "$USE_ACCEL" ]; then
USE_CUDA=$USE_ACCEL
fi
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
USE_CUDA=${USE_CUDA:-False}
tweak running examples without cuda (#794) * tweak running examples without cuda * rework dry_run handling in mnist, mnist_hogwild
2020-07-02 00:43:39 +03:00
case $USE_CUDA in
"True")
echo "using cuda"
CUDA=1
CUDA_FLAG="--cuda"
Use torch.accelerator API in Fast Neural Style example (#1327)
2025-04-28 12:46:10 -07:00
ACCEL_FLAG="--accel"
tweak running examples without cuda (#794) * tweak running examples without cuda * rework dry_run handling in mnist, mnist_hogwild
2020-07-02 00:43:39 +03:00
;;
"False")
echo "not using cuda"
CUDA=0
CUDA_FLAG=""
Use torch.accelerator API in Fast Neural Style example (#1327)
2025-04-28 12:46:10 -07:00
ACCEL_FLAG=""
tweak running examples without cuda (#794) * tweak running examples without cuda * rework dry_run handling in mnist, mnist_hogwild
2020-07-02 00:43:39 +03:00
;;
"")
exit 1;
;;
esac
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function dcgan() {
Use torch.accelerator in DCGAN example (#1344) Signed-off-by: dggaytan <diana.gaytan.munoz@intel.com>
2025-05-13 13:14:47 -06:00
uv run main.py --dataset fake $ACCEL_FLAG --dry-run || error "dcgan failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
function fast_neural_style() {
if [ ! -d "saved_models" ]; then
echo "downloading saved models for fast neural style"
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run download_saved_models.py
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
fi
test -d "saved_models" || { error "saved models not found"; return; }
echo "running fast neural style model"
Use torch.accelerator API in Fast Neural Style example (#1327)
2025-04-28 12:46:10 -07:00
uv run neural_style/neural_style.py eval --content-image images/content-images/amber.jpg --model saved_models/candy.pth --output-image images/output-images/amber-candy.jpg $ACCEL_FLAG || error "neural_style.py failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
function imagenet() {
if [[ ! -d "sample/val" || ! -d "sample/train" ]]; then
mkdir -p sample/val/n
mkdir -p sample/train/n
Add mps device (#1064) * Add mps device * Add --mps to run_python_examples.sh * Update imagenet with mps device * Use curl in run_python_examples.sh to accommodate macOS devices * Fix for https://github.com/pytorchq/examples/issues/1060
2022-09-20 16:18:33 -07:00
curl -O "https://upload.wikimedia.org/wikipedia/commons/5/5a/Socks-clinton.jpg" || { error "couldn't download sample image for imagenet"; return; }
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
mv Socks-clinton.jpg sample/train/n
cp sample/train/n/* sample/val/n/
fi
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 sample/ || error "imagenet example failed"
imagenet: fix typo addressing args.gpu (#1361)
2025-06-30 11:52:25 -07:00
uv run main.py --epochs 1 --gpu 0 sample/ || error "imagenet example failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Language translation example added (#1131) (#1240) * Intial model setup * Training works * Added inference * Code clean up and commenting * Update to README.md * Add requirements.txt * Updated top level README, added example to CI * Potentially fixed testing (maybe not enough memory?) * Update requirements.txt --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2024-04-02 12:50:04 -04:00
function language_translation() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run -m spacy download en || error "couldn't download en package from spacy"
uv run -m spacy download de || error "couldn't download de package from spacy"
uv run main.py -e 1 --enc_layers 1 --dec_layers 1 --backend cpu --logging_dir output/ --dry_run || error "language translation example failed"
Language translation example added (#1131) (#1240) * Intial model setup * Training works * Added inference * Code clean up and commenting * Update to README.md * Add requirements.txt * Updated top level README, added example to CI * Potentially fixed testing (maybe not enough memory?) * Update requirements.txt --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2024-04-02 12:50:04 -04:00
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function mnist() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "mnist example failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Fixes #1111 : Added Forward-Forward Algorithm (#1114) * add code for ff * completed the Pull reqs. * Final Commit * sorted * Update CODEOWNERS * Update main.py
2023-02-27 05:17:16 +05:30
function mnist_forward_forward() {
Use torch.accelerator API in mnist examples (#1334)
2025-04-30 08:02:42 -06:00
uv run main.py --epochs 1 --no_accel || error "mnist forward forward failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
Fixes #1111 : Added Forward-Forward Algorithm (#1114) * add code for ff * completed the Pull reqs. * Final Commit * sorted * Update CODEOWNERS * Update main.py
2023-02-27 05:17:16 +05:30
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function mnist_hogwild() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run $CUDA_FLAG || error "mnist hogwild failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
add mnist_rnn to test script for CI (#1086) * fix device mismatch issue #1071 * fix device mismatch issue #1071 * add mnist_rnn to test script for CI * support dry_run in test()
2022-10-27 14:31:20 -07:00
function mnist_rnn() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "mnist rnn example failed"
add mnist_rnn to test script for CI (#1086) * fix device mismatch issue #1071 * fix device mismatch issue #1071 * add mnist_rnn to test script for CI * support dry_run in test()
2022-10-27 14:31:20 -07:00
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function regression() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 $CUDA_FLAG || error "regression failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Updated based on PR's feedback - Included `siamese_network` setup to `run_python_examples.sh` - In `siamese_network/main.py`, included more explanation and detailed comments per @msaroufim's feedback.
2022-05-07 19:10:31 -07:00
function siamese_network() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "siamese network example failed"
Updated based on PR's feedback - Included `siamese_network` setup to `run_python_examples.sh` - In `siamese_network/main.py`, included more explanation and detailed comments per @msaroufim's feedback.
2022-05-07 19:10:31 -07:00
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function reinforcement_learning() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run reinforce.py || error "reinforcement learning reinforce failed"
uv run actor_critic.py || error "reinforcement learning actor_critic failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
function snli() {
echo "installing 'en' model if not installed"
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run -m spacy download en || { error "couldn't download 'en' model needed for snli"; return; }
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
echo "training..."
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run train.py --epochs 1 --dev_every 1 --no-bidirectional --dry-run || error "couldn't train snli"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
added fx tests
2022-04-15 10:55:35 -07:00
function fx() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
# uv run custom_tracer.py || error "fx custom tracer has failed" UnboundLocalError: local variable 'tabulate' referenced before assignment
uv run invert.py || error "fx invert has failed"
uv run module_tracer.py || error "fx module tracer has failed"
uv run primitive_library.py || error "fx primitive library has failed"
uv run profiling_tracer.py || error "fx profiling tracer has failed"
uv run replace_op.py || error "fx replace op has failed"
uv run subgraph_rewriter_basic_use.py || error "fx subgraph has failed"
uv run wrap_output_dynamically.py || error "vmap output dynamically has failed"
added fx tests
2022-04-15 10:55:35 -07:00
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function super_resolution() {
Add accelerate API support for Super Resolution example (#1358) Update super_resolution example to support accelerate API
2025-06-23 19:17:56 -06:00
uv run main.py --upscale_factor 3 --batchSize 4 --testBatchSize 100 --nEpochs 1 --lr 0.001 $ACCEL_FLAG || error "super resolution failed"
uv run super_resolve.py --input_image dataset/BSDS300/images/test/16077.jpg --model model_epoch_1.pth --output_filename out.png $ACCEL_FLAG || error "super resolution upscaling failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Fix errors executing run_python_examples.sh "install_deps,run_all" (#787) Co-authored-by: root <root@1fb96699-953f-4d40-97ce-0aa77e57281b.apple.com>
2020-06-09 22:08:19 -07:00
function time_sequence_prediction() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run generate_sine_wave.py || { error "generate sine wave failed"; return; }
uv run train.py --steps 2 || error "time sequence prediction training failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
function vae() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 || error "vae failed"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Implemented Vision Transformer in PyTorch (#1141)
2023-04-26 21:24:13 +05:30
function vision_transformer() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "vision transformer example failed"
Implemented Vision Transformer in PyTorch (#1141)
2023-04-26 21:24:13 +05:30
}
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
function word_language_model() {
Add accelerate API support for Word Language Model example (#1345) * Add accelerate API support for Word Language Model example * Update README for Word Language Model example * Update word_language_model/generate.py for consistency Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update word_language_model/README.md Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update README to change wording on acceleration devices * Remove cuda conditional * Fix flags for word_language_model in ci script --------- Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-08-13 12:31:07 -06:00
uv run main.py --epochs 1 --dry-run $ACCEL_FLAG || error "word_language_model failed"
uv run generate.py $ACCEL_FLAG || error "word_language_model generate failed"
Support torch>=2.6 in word_language_model example (#1347) Update the example usage of `torch.load()` with required safe globals. Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-05-17 20:20:56 -07:00
for model in "RNN_TANH" "RNN_RELU" "LSTM" "GRU" "Transformer"; do
Add accelerate API support for Word Language Model example (#1345) * Add accelerate API support for Word Language Model example * Update README for Word Language Model example * Update word_language_model/generate.py for consistency Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update word_language_model/README.md Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update README to change wording on acceleration devices * Remove cuda conditional * Fix flags for word_language_model in ci script --------- Co-authored-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-08-13 12:31:07 -06:00
uv run main.py --model $model --epochs 1 --dry-run $ACCEL_FLAG || error "word_language_model failed"
uv run generate.py $ACCEL_FLAG || error "word_language_model generate failed"
Support torch>=2.6 in word_language_model example (#1347) Update the example usage of `torch.load()` with required safe globals. Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-05-17 20:20:56 -07:00
done
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
Graph Convolutional Network (#1163) * Implement GCN model architecture * Comments added to GCN example * test added * match random seed with PR screenshot * Update index.rst
2023-06-12 12:22:18 -06:00
function gcn() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "graph convolutional network failed"
Graph Convolutional Network (#1163) * Implement GCN model architecture * Comments added to GCN example * test added * match random seed with PR screenshot * Update index.rst
2023-06-12 12:22:18 -06:00
}
Added Graph Attention Network example (#1174) * add GAT test, add GAT card, update GAT README * add GAT test, add GAT card, update GAT README * update doc build * update tests * revert doc index * update doc index * update doc index * update doc index * revert doc index * revert doc index * revert doc index * revert doc index * Trigger Build * update requirements.txt, remove versions --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2023-08-08 11:49:10 -04:00
function gat() {
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
uv run main.py --epochs 1 --dry-run || error "graph attention network failed"
Added Graph Attention Network example (#1174) * add GAT test, add GAT card, update GAT README * add GAT test, add GAT card, update GAT README * update doc build * update tests * revert doc index * update doc index * update doc index * update doc index * revert doc index * revert doc index * revert doc index * revert doc index * Trigger Build * update requirements.txt, remove versions --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2023-08-08 11:49:10 -04:00
}
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
eval "base_$(declare -f stop)"
function stop() {
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
cd $BASE_DIR
Fix flaky dcgan test by using fake data (#1044)
2022-08-23 17:06:15 -07:00
rm -rf dcgan/fake_samples_epoch_000.png \
dcgan/netD_epoch_0.pth \
dcgan/netG_epoch_0.pth \
tweak running examples without cuda (#794) * tweak running examples without cuda * rework dry_run handling in mnist, mnist_hogwild
2020-07-02 00:43:39 +03:00
dcgan/real_samples.png \
fast_neural_style/saved_models.zip \
fast_neural_style/saved_models/ \
imagenet/checkpoint.pth.tar \
imagenet/lsun/ \
imagenet/model_best.pth.tar \
imagenet/sample/ \
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
language_translation/output/ \
tweak running examples without cuda (#794) * tweak running examples without cuda * rework dry_run handling in mnist, mnist_hogwild
2020-07-02 00:43:39 +03:00
snli/.data/ \
snli/.vector_cache/ \
snli/results/ \
time_sequence_prediction/predict*.pdf \
time_sequence_prediction/traindata.pt \
Graph Convolutional Network (#1163) * Implement GCN model architecture * Comments added to GCN example * test added * match random seed with PR screenshot * Update index.rst
2023-06-12 12:22:18 -06:00
word_language_model/model.pt \
Added Graph Attention Network example (#1174) * add GAT test, add GAT card, update GAT README * add GAT test, add GAT card, update GAT README * update doc build * update tests * revert doc index * update doc index * update doc index * update doc index * revert doc index * revert doc index * revert doc index * revert doc index * Trigger Build * update requirements.txt, remove versions --------- Co-authored-by: Mark Saroufim <marksaroufim@meta.com>
2023-08-08 11:49:10 -04:00
gcn/cora/ \
Revert "Add Swin Transformer Example " (#1383) Revert "Add Swin Transformer Example (#1346)" This reverts commit 892eca0a9469f657bf4c635d1b9b648d267580fc.
2025-08-22 23:21:04 -07:00
gat/cora/ || error "couldn't clean up some files"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
git checkout fast_neural_style/images/output-images/amber-candy.jpg || error "couldn't clean up fast neural style image"
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
base_stop "$1"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
function run_all() {
Update distributed example tests in `run_python_examples.sh` (#1250) * Fix distributed test * fix parallel scripts * install dill * remove dill * run 2 gpu * remove gpucount, use default * Add examples to distributed examples * refactor distributed test * fx ERRORS overwriting * run with base dir * remove distributed from run_python_examples.sh * move basedir to source * separate init --------- Co-authored-by: Sirut Buasai <sirutbuasai27@hotmail.com>
2024-05-03 13:25:49 -07:00
# cpp moved to `run_cpp_examples.sh```
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run dcgan
Update distributed example tests in `run_python_examples.sh` (#1250) * Fix distributed test * fix parallel scripts * install dill * remove dill * run 2 gpu * remove gpucount, use default * Add examples to distributed examples * refactor distributed test * fx ERRORS overwriting * run with base dir * remove distributed from run_python_examples.sh * move basedir to source * separate init --------- Co-authored-by: Sirut Buasai <sirutbuasai27@hotmail.com>
2024-05-03 13:25:49 -07:00
# distributed moved to `run_distributed_examples.sh`
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run fast_neural_style
run imagenet
Update run_python_examples.sh (#1269)
2024-06-08 13:13:14 -07:00
# language_translation
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run mnist
run mnist_forward_forward
run mnist_hogwild
run mnist_rnn
run regression
run reinforcement_learning
run siamese_network
Comment out super resolution paths and run command
2025-08-23 00:03:28 -07:00
# run super_resolution - flaky
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run time_sequence_prediction
run vae
Remove Vision Transformer Example (#1258) * Remove Vision Transformer Example * yolo * Trigger CI
2024-05-24 22:14:30 -07:00
# vision_transformer - example broken see https://github.com/pytorch/examples/issues/1184 and https://github.com/pytorch/examples/pull/1258 for more details
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run word_language_model
run fx
run gcn
run gat
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
}
# by default, run all examples
if [ "" == "$EXAMPLES" ]; then
run_all
else
for i in $(echo $EXAMPLES | sed "s/,/ /g")
do
fix warnings and failures
2020-07-05 18:20:41 +03:00
echo "Starting $i"
Respect each example requirements and use uv (#1330) * Add requirements.txt to examples which miss them Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update numpy requirement for reinforcement_learning to be <2 Current version of the example requires `numpy<2` otherwise the following error can be seen: ``` AttributeError: module 'numpy' has no attribute 'bool8'. Did you mean: 'bool'? ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Update torch requirement for time and word examples to be <2.6 Current version of examples require `torch<2.6` otherwise the following error can be seen: ``` File "/pytorch/examples/time_sequence_prediction/train.py", line 47, in <module> data = torch.load('traindata.pt') ^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/pytorch/examples/time_sequence_prediction/.venv/lib/python3.12/site-packages/torch/serialization.py", line 1524, in load raise pickle.UnpicklingError(_get_wo_message(str(e))) from None ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> * Respect each example requirements and use uv This commit introduces few changes to CI by modifying `run_*_examples.sh` and respective github workflows: * Switched to uv * Added tearup and teardown stages for tests (`start()` and `stop()` methods wrapping up test bodies - these are called automatically) * Tearup (`start()`) installs example dependencies and, optionally (if `VIRTUAL_ENV=.venv` is passed), creates uv virtual environment * Teardown (`stop()`) removes uv virtual environment if it was created (to save space) * If no `VIRTUAL_ENV` set, then scripts expect to be executed in the existing virtual environment. These can be `python -m venv`, `uv env` or `conda env`. In this case example dependencies will be installed in this environment potentially reinstalling existing packages (including `torch`!). * Dropped automated detection of CUDA platform. Now scripts require `USE_CUDA=True` to be passed explicitly * Added `PIP_INSTALL_ARGS` environment variable to be passed to `uv pip install` calls for each example dependencies. This allows to adjust torch indices and other options. Execute all tests in current virtual environment (might rewrite packages): ``` ./run_distributed_examples.sh ``` Execute all tests creating separate environment for each example: ``` VIRTUAL_ENV=.venv ./run_distributed_examples.sh ``` Run with CUDA: ``` USE_CUDA=True ./run_distributed_examples.sh ``` Adjust index: ``` PIP_INSTALL_ARGS="--pre -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html" \ ./run_distributed_examples.sh ``` Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> --------- Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
2025-04-25 21:28:00 -07:00
run $i
fix warnings and failures
2020-07-05 18:20:41 +03:00
echo "Finished $i, status $?"
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
done
fi
if [ "" == "$ERRORS" ]; then
fix warnings and failures
2020-07-05 18:20:41 +03:00
echo "Completed successfully with status $?"
Fix shebang (#686) This script uses some bash-specific features such as - function keyword - == operator - FUNCNAME
2020-03-04 20:05:36 -05:00
else
Update distributed example tests in `run_python_examples.sh` (#1250) * Fix distributed test * fix parallel scripts * install dill * remove dill * run 2 gpu * remove gpucount, use default * Add examples to distributed examples * refactor distributed test * fx ERRORS overwriting * run with base dir * remove distributed from run_python_examples.sh * move basedir to source * separate init --------- Co-authored-by: Sirut Buasai <sirutbuasai27@hotmail.com>
2024-05-03 13:25:49 -07:00
echo "Some python examples failed:"
printf "$ERRORS\n"
test CI flag
2022-01-16 02:15:13 +02:00
#Exit with error (0-255) in case of failure in one of the tests.
Fix CI Job Report when a test fail #943
2022-01-14 05:19:58 +02:00
exit 1
test CI flag
2022-01-16 02:15:13 +02:00
add a script to run all pytorch examples (#591)
2019-07-12 21:19:42 -04:00
fi
Reference in New Issue Copy Permalink