Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Change some display. Break up some paragraphs. #2

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Change some display. Break up some paragraphs. #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions docs/conf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,3 +41,6 @@
html_logo = "_static/lincc-fw.png"
html_title = "LINCC Frameworks"
html_theme = "sphinx_book_theme"
html_theme_options = {
"show_toc_level": 2
}
137 changes: 48 additions & 89 deletions docs/notebooks/fibad/fibad_demo.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,18 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"# Fibad Demonstration"
"# Demo: Fibad\n",
"\n",
"**Author**: Drew\n",
"\n",
"**Last updated**: Feb 4, 2025\n",
"\n",
"In this notebook, we'll do an end-to-end demonstration of working with FIBAD utilities to perform some image-based inference. This includes:\n",
"\n",
"* fetching the images to train/evaluate/perform inference on\n",
"* train up a model\n",
"* run inference on the model\n",
"* look around at the results"
]
},
{
Expand All @@ -14,11 +25,14 @@
"## Download a sample HSC dataset\n",
"\n",
"This dataset is comprised of approximately 993 cutouts from the Hyper Suprime Cam survey.\n",
"The cutouts were requested to be 8 arcsecs on a side.\n",
"For consistency, we crop each image to [96, 96] pixels at runtime.\n",
"For each object 3 bands have been acquired, I, R, G.\n",
"\n",
"Once unzipped there will be a .fits file for each (object id, band) in the `./data/hsc_8asec_1000` directory.\n"
"Some characteristics of the cut-outs:\n",
"\n",
"* 8 arcsecs on a side\n",
"* crop each image to [96, 96] pixels at runtime.\n",
"* 3 bands have been acquired: `I`, `R`, `G`\n",
"\n",
"We'll create a local directory, `./data/hsc_8asec_1000`, and unzip to create a .fits file for each (object id, band) combo.\n"
]
},
{
Expand Down Expand Up @@ -85,7 +99,31 @@
"source": [
"## Configuration and Training\n",
"\n",
"First we import fibad and create a new fibad object, instantiated (implicitly), with the default configuration file."
"The flow to get a model trained is straightforward:\n",
"\n",
"1. create a new instance\n",
"2. configure that instance\n",
"\n",
" * where's the data?\n",
" * what shape is the data?\n",
" * what kind of model do you want?\n",
" * what are your model hyperparameters?\n",
"\n",
"3. train that model\n",
"\n",
"Reusing the configuration is made easy, too. The fibad instance will store its configuration in `./results/<timestamp>/runtime.config.toml`.\n",
"\n",
"1. If running in another notebook, instantiate a fibad object like so:\n",
"\n",
" ```\n",
" new_fibad_instance = fibad.Fibad(config_file='./results/<timestamped_directory>/runtime_config.toml')\n",
" ```\n",
"\n",
"2. Or from the command line on an HPC system:\n",
"\n",
" ```\n",
" >> fibad train --runtime-config ./results/<timestamped_directory>/runtime_config.toml\n",
" ```"
]
},
{
Expand All @@ -104,24 +142,8 @@
"source": [
"import fibad\n",
"\n",
"f = fibad.Fibad()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For this demo, we'll make a few adjustments to the default configuration settings that the `fibad` object was instantiated with. By accessing the `.config` attribute of the fibad instance, we can modify any configuration value. \n",
"f = fibad.Fibad()\n",
"\n",
"Here we specify the location of our sample data, the data set class, the model to train, number of epochs for training and the batch size."
]
},
{
"cell_type": "code",
"execution_count": 153,
"metadata": {},
"outputs": [],
"source": [
"# Specify the location of the data to use for training\n",
"f.config[\"general\"][\"data_dir\"] = \"./data/hsc_8asec_1000_bu\"\n",
"\n",
Expand All @@ -133,73 +155,10 @@
"\n",
"# Set the number of epochs and batch size for training.\n",
"f.config[\"train\"][\"epochs\"] = 1\n",
"f.config[\"data_loader\"][\"batch_size\"] = 32"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We call the `.train()` method to train the model"
]
},
{
"cell_type": "code",
"execution_count": 154,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"[2025-01-30 21:48:31,624 fibad.data_sets.hsc_data_set:INFO] Processed 993 objects for pruning\n",
"[2025-01-30 21:48:31,625 fibad.data_sets.hsc_data_set:INFO] Checking file dimensions to determine standard cutout size...\n",
"[2025-01-30 21:48:31,628 fibad.data_sets.hsc_data_set:INFO] HSC Data set loader has 993 objects\n",
"[2025-01-30 21:48:31,632 fibad.data_sets.hsc_data_set:INFO] test split contains 199 items\n",
"[2025-01-30 21:48:31,633 fibad.data_sets.hsc_data_set:INFO] train split contains 596 items\n",
"[2025-01-30 21:48:31,633 fibad.data_sets.hsc_data_set:INFO] validate split contains 198 items\n",
"[2025-01-30 21:48:31,644 fibad.models.model_registry:INFO] Using criterion: torch.nn.CrossEntropyLoss with default arguments.\n",
"2025-01-30 21:48:31,646 ignite.distributed.auto.auto_dataloader INFO: Use data loader kwargs for dataset '<fibad.data_sets.hsc': \n",
"\t{'sampler': <torch.utils.data.sampler.SubsetRandomSampler object at 0x7f98004e1ca0>, 'batch_size': 32, 'num_workers': 2, 'pin_memory': True}\n",
"2025-01-30 21:48:31,647 ignite.distributed.auto.auto_dataloader INFO: Use data loader kwargs for dataset '<fibad.data_sets.hsc': \n",
"\t{'sampler': <torch.utils.data.sampler.SubsetRandomSampler object at 0x7f9800949cd0>, 'batch_size': 32, 'num_workers': 2, 'pin_memory': True}\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"2025/01/30 21:48:31 INFO mlflow.system_metrics.system_metrics_monitor: Started monitoring system metrics.\n",
"[2025-01-30 21:48:31,690 fibad.pytorch_ignite:INFO] Training model on device: cuda\n",
"[2025-01-30 21:48:31,691 fibad.pytorch_ignite:INFO] Total epochs: 1\n",
"[2025-01-30 21:48:35,026 fibad.pytorch_ignite:INFO] Total training time: 3.34[s]\n",
"[2025-01-30 21:48:35,027 fibad.pytorch_ignite:INFO] Latest checkpoint saved as: /home/drew/code/fibad/results/20250130-214831-train-cn8s/checkpoint_epoch_1.pt\n",
"[2025-01-30 21:48:35,028 fibad.pytorch_ignite:INFO] Best metric checkpoint saved as: /home/drew/code/fibad/results/20250130-214831-train-cn8s/checkpoint_1_loss=-1077.9679.pt\n",
"2025/01/30 21:48:35 INFO mlflow.system_metrics.system_metrics_monitor: Stopping system metrics monitoring...\n",
"2025/01/30 21:48:35 INFO mlflow.system_metrics.system_metrics_monitor: Successfully terminated system metrics monitoring!\n",
"[2025-01-30 21:48:35,692 fibad.train:INFO] Finished Training\n"
]
}
],
"source": [
"f.train()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The output of the training will be stored in a time-stamped directory under the `./results/`. By default, a copy of the final configuration used in training is persisted as `runtime_config.toml`. To run fibad again with the same configuration, you can reference the runtime_config.toml file.\n",
"\n",
"If running in another notebook, instantiate a fibad object like so:\n",
"```\n",
"new_fibad_instance = fibad.Fibad(config_file='./results/<timestamped_directory>/runtime_config.toml')\n",
"```\n",
"f.config[\"data_loader\"][\"batch_size\"] = 32\n",
"\n",
"Or from the command line on an HPC system:\n",
"```\n",
">> fibad train --runtime-config ./results/<timestamped_directory>/runtime_config.toml\n",
"```"
"# We call the this method to train the model\n",
"f.train()"
]
},
{
Expand Down
6 changes: 5 additions & 1 deletion docs/notebooks/nested/review_demo.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,11 @@
"id": "9d8ebc3c-ac7e-4ce4-ae37-15d54c7afb7d",
"metadata": {},
"source": [
"# LINCC Frameworks `nested-pandas`/`nested-dask` Review Demo\n"
"# Demo: `nested-pandas` and `nested-dask`\n",
"\n",
"**Author**: Doug\n",
"\n",
"**Last updated**: Feb 3, 2025\n"
]
},
{
Expand Down
3 changes: 0 additions & 3 deletions docs/requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,10 +1,7 @@

ipykernel
ipython
jupytext
nbconvert
nbsphinx
sphinx
sphinx-autoapi
sphinx-copybutton
sphinx-book-theme