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

Jump to bottom

Fix validate lightning models inference tutorial #287

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Fix validate lightning models inference tutorial #287

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4de1e0b
fix: Fix validate lightning models inference tutorial
ktagowski Feb 15, 2023
8393b16
Merge branch 'main' into fix/fix-inference-tutorial
ktagowski Feb 15, 2023
b2f1180
Merge branch 'main' into fix/fix-inference-tutorial
ktagowski Apr 17, 2023
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
2 changes: 1 addition & 1 deletion .github/workflows/ghp_deploy.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@ name: CD
on:
push:
branches:
["main", "master", "271-create-documentation-and-library-presentation"]
["main", "master", "271-create-documentation-and-library-presentation", "fix/fix-inference-tutorial"]
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please remove all beside main, I forgot to remove it. If you want to run is manually you can use GH actions menu and run specific job

workflow_dispatch:
jobs:
deploy_ghp:
Expand Down
247 changes: 36 additions & 211 deletions nbs/01_Tutorials/validate_lightning_models_inference.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,30 +30,24 @@
"outputs": [],
"source": [
"#| hide\n",
"%load_ext autoreload\n",
"%autoreload 2"
"import os\n",
"import warnings\n",
"\n",
"# disable warnings\n",
"warnings.simplefilter(\"ignore\")\n",
"# set os environ variable for multiprocesses\n",
"os.environ[\"PYTHONWARNINGS\"] = \"ignore\"\n",
"\n",
"os.chdir(\"..\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1019b750-cebe-438b-b1ab-434d6f756864",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
" from .autonotebook import tqdm as notebook_tqdm\n"
]
}
],
"outputs": [],
"source": [
"#| eval: false\n",
"import os\n",
"\n",
"os.chdir(\"..\")\n",
"from typing import Any, Dict\n",
"\n",
"import pytorch_lightning as pl\n",
Expand All @@ -75,25 +69,15 @@
"execution_count": null,
"id": "2d0e06e2-3c5a-420b-b065-31d5ccd6b255",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"2023-02-13 23:05:26,246 - embeddings.utils.utils - WARNING - String 'hf-internal-testing/tiny-albert' contains '/'. Replacing it with '__'. Cleaned_text: hf-internal-testing__tiny-albert.\n",
"2023-02-13 23:05:26,247 - embeddings.utils.utils - WARNING - String 'clarin-pl/polemo2-official' contains '/'. Replacing it with '__'. Cleaned_text: clarin-pl__polemo2-official.\n",
"2023-02-13 23:05:26,254 - embeddings.utils.utils - WARNING - String 'hf-internal-testing/tiny-albert' contains '/'. Replacing it with '__'. Cleaned_text: hf-internal-testing__tiny-albert.\n",
"2023-02-13 23:05:26,256 - embeddings.utils.utils - WARNING - String 'clarin-pl/polemo2-official' contains '/'. Replacing it with '__'. Cleaned_text: clarin-pl__polemo2-official.\n"
]
}
],
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"embedding_name_or_path = \"hf-internal-testing/tiny-albert\"\n",
"dataset_name = \"clarin-pl/polemo2-official\"\n",
"\n",
"dataset_path = build_output_path(DATASET_PATH, embedding_name_or_path, dataset_name)\n",
"output_path = build_output_path(RESULTS_PATH, embedding_name_or_path, dataset_name)"
"output_path = build_output_path(\".\", embedding_name_or_path, dataset_name, mkdirs=True)"
]
},
{
Expand All @@ -109,26 +93,10 @@
"execution_count": null,
"id": "095d1c88-900f-4275-a879-f9efdb73265a",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using custom data configuration default-e0c1ce6ddfd81769\n",
"Found cached dataset polemo2-official (/root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70)\n",
"100%|██████████| 3/3 [00:00<00:00, 817.23it/s]\n",
"Loading cached split indices for dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-a54edce9681df8b7.arrow and /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-09cf731207f31628.arrow\n",
"Loading cached split indices for dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-c48721732fabb729.arrow and /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-f9d782422a65c7e6.arrow\n",
"Loading cached split indices for dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-0db6321193feb3ec.arrow and /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-4e6c26839c3e4adf.arrow\n",
"Loading cached processed dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-e32b75da1d28bfd0.arrow\n",
"Loading cached processed dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-98cbedcc70a23855.arrow\n",
"Loading cached processed dataset at /root/.cache/huggingface/datasets/clarin-pl___polemo2-official/default-e0c1ce6ddfd81769/0.0.0/2b75fdbe5def97538e81fb120f8752744b50729a4ce09bd75132bfc863a2fd70/cache-b2cbb8ab856bac0f.arrow\n",
" \r"
]
}
],
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"def preprocess_data(path: str) -> Dict[str, Any]:\n",
" pipeline = HuggingFacePreprocessingPipeline(\n",
" dataset_name=dataset_name,\n",
Expand All @@ -141,7 +109,7 @@
" persist_path=path,\n",
" sample_missing_splits=None,\n",
" ignore_test_subset=False,\n",
" downsample_splits=(0.01, 0.01, 0.05),\n",
" downsample_splits=(0.01, 0.01, 0.01),\n",
" seed=441,\n",
" )\n",
" pipeline.run()\n",
Expand Down Expand Up @@ -171,7 +139,8 @@
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"config = LightningAdvancedConfig(\n",
" finetune_last_n_layers=0,\n",
" task_train_kwargs={\"max_epochs\": 1, \"deterministic\": True,},\n",
Expand Down Expand Up @@ -199,140 +168,10 @@
"execution_count": null,
"id": "148a0089-f461-4948-93fa-04f2e34ac9e0",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|██████████| 1/1 [00:00<00:00, 6.73ba/s]\n",
"100%|██████████| 1/1 [00:00<00:00, 25.69ba/s]\n",
"100%|██████████| 1/1 [00:00<00:00, 26.51ba/s]\n",
"Casting the dataset: 100%|██████████| 1/1 [00:00<00:00, 113.84ba/s]\n",
"Casting the dataset: 100%|██████████| 1/1 [00:00<00:00, 68.70ba/s]\n",
"Casting the dataset: 100%|██████████| 1/1 [00:00<00:00, 103.15ba/s]\n",
"Some weights of the model checkpoint at hf-internal-testing/tiny-albert were not used when initializing AlbertForSequenceClassification: ['predictions.LayerNorm.bias', 'predictions.LayerNorm.weight', 'predictions.bias', 'predictions.decoder.weight', 'predictions.decoder.bias', 'predictions.dense.bias', 'predictions.dense.weight']\n",
"- This IS expected if you are initializing AlbertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
"- This IS NOT expected if you are initializing AlbertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
"Some weights of AlbertForSequenceClassification were not initialized from the model checkpoint at hf-internal-testing/tiny-albert and are newly initialized: ['classifier.bias', 'albert.pooler.weight', 'albert.pooler.bias', 'classifier.weight']\n",
"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n",
"GPU available: True, used: False\n",
"TPU available: False, using: 0 TPU cores\n",
"IPU available: False, using: 0 IPUs\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/trainer.py:1579: UserWarning: GPU available but not used. Set the gpus flag in your trainer `Trainer(gpus=1)` or script `--gpus=1`.\n",
" rank_zero_warn(\n",
"\n",
" | Name | Type | Params\n",
"------------------------------------------------------------------\n",
"0 | model | AlbertForSequenceClassification | 352 K \n",
"1 | metrics | MetricCollection | 0 \n",
"2 | train_metrics | MetricCollection | 0 \n",
"3 | val_metrics | MetricCollection | 0 \n",
"4 | test_metrics | MetricCollection | 0 \n",
"------------------------------------------------------------------\n",
"132 Trainable params\n",
"352 K Non-trainable params\n",
"352 K Total params\n",
"1.410 Total estimated model params size (MB)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Validation sanity check: 0%| | 0/1 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/data_loading.py:111: UserWarning: The dataloader, val_dataloader 0, does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` (try 48 which is the number of cpus on this machine) in the `DataLoader` init to improve performance.\n",
" rank_zero_warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" "
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/data_loading.py:111: UserWarning: The dataloader, train_dataloader, does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` (try 48 which is the number of cpus on this machine) in the `DataLoader` init to improve performance.\n",
" rank_zero_warn(\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/data_loading.py:407: UserWarning: The number of training samples (2) is smaller than the logging interval Trainer(log_every_n_steps=50). Set a lower value for log_every_n_steps if you want to see logs for the training epoch.\n",
" rank_zero_warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 0: 100%|██████████| 3/3 [00:01<00:00, 2.07it/s, loss=1.39, v_num=, train/BaseLR=5e-6, train/LambdaLR=5e-6, val/MulticlassAccuracy=0.375, val/MulticlassPrecision=0.0938, val/MulticlassRecall=0.250, val/MulticlassF1Score=0.136]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/data_loading.py:111: UserWarning: The dataloader, test_dataloader 0, does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` (try 48 which is the number of cpus on this machine) in the `DataLoader` init to improve performance.\n",
" rank_zero_warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Testing: 100%|██████████| 2/2 [00:02<00:00, 1.02s/it]--------------------------------------------------------------------------------\n",
"DATALOADER:0 TEST RESULTS\n",
"{'test/Loss': 1.3823001384735107,\n",
" 'test/MulticlassAccuracy': 0.4054054021835327,\n",
" 'test/MulticlassF1Score': 0.14423076808452606,\n",
" 'test/MulticlassPrecision': 0.10135135054588318,\n",
" 'test/MulticlassRecall': 0.25}\n",
"--------------------------------------------------------------------------------\n",
"Testing: 100%|██████████| 2/2 [00:02<00:00, 1.04s/it]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Restoring states from the checkpoint path at /app/resources/results/hf-internal-testing__tiny-albert/clarin-pl__polemo2-official/20230213_230526/checkpoints/epoch=0-step=1.ckpt\n",
"Loaded model weights from checkpoint at /app/resources/results/hf-internal-testing__tiny-albert/clarin-pl__polemo2-official/20230213_230526/checkpoints/epoch=0-step=1.ckpt\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/pytorch_lightning/trainer/data_loading.py:111: UserWarning: The dataloader, predict_dataloader 0, does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` (try 48 which is the number of cpus on this machine) in the `DataLoader` init to improve performance.\n",
" rank_zero_warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Predicting: 100%|██████████| 2/2 [00:01<?, ?it/s]\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/app/embeddings/metric/hugging_face_metric.py:27: FutureWarning: load_metric is deprecated and will be removed in the next major version of datasets. Use 'evaluate.load' instead, from the new library 🤗 Evaluate: https://huggingface.co/docs/evaluate\n",
" datasets.load_metric(metric, **init_kwargs) if isinstance(metric, str) else metric\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
" _warn_prf(average, modifier, msg_start, len(result))\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
" _warn_prf(average, modifier, msg_start, len(result))\n",
"/opt/conda/envs/embeddings/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
" _warn_prf(average, modifier, msg_start, len(result))\n"
]
}
],
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"pipeline = LightningClassificationPipeline(\n",
" embedding_name_or_path=embedding_name_or_path,\n",
" output_path=output_path,\n",
Expand All @@ -352,26 +191,17 @@
"### Load model from chechpoint automatically generated with Trainer"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "ee9e824c-00f1-45b0-9e32-1bd33f364f3a",
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"ckpt_path = output_path / \"checkpoints\" / \"last.ckpt\"\n",
"ckpt_path"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "2785fcbc-1c95-4d23-807f-a14569992354",
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"ckpt_path = output_path / \"checkpoints\" / \"epoch=0-step=1.ckpt\"\n",
"\n",
"task_from_ckpt = TextClassificationTask.from_checkpoint(\n",
" checkpoint_path=ckpt_path, output_path=output_path,\n",
")"
Expand All @@ -392,7 +222,8 @@
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"model_from_ckpt = TextClassificationModule.load_from_checkpoint(str(ckpt_path))"
]
},
Expand Down Expand Up @@ -429,7 +260,8 @@
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"test_dataloader = pipeline.datamodule.test_dataloader()\n",
"preds = task_from_ckpt.predict(test_dataloader)\n",
"preds"
Expand All @@ -450,7 +282,8 @@
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"task_from_ckpt.trainer.datamodule = pipeline.datamodule\n",
"preds_with_names = task_from_ckpt.predict(test_dataloader, return_names=True)\n",
"preds_with_names"
Expand All @@ -471,7 +304,8 @@
"metadata": {},
"outputs": [],
"source": [
"#| eval: false\n",
"#|exec_doc\n",
"\n",
"trainer = pl.Trainer(default_root_dir=str(output_path))\n",
"preds_from_model = trainer.predict(model_from_ckpt, dataloaders=test_dataloader)\n",
"preds_from_model"
Expand All @@ -480,18 +314,9 @@
],
"metadata": {
"kernelspec": {
"display_name": "embeddings",
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python",
"version": "3.9.16"
},
"vscode": {
"interpreter": {
"hash": "86b40992624b6ecf125385760a49d2b554d653d5c84d942a6f4a5512888cc722"
}
}
},
"nbformat": 4,
Expand Down