✨ Enhancements and Bug fixes

What's Changed

• 🌟 format docs by @aniketmaurya in #170
• [Snyk] Fix for 3 vulnerabilities by @snyk-bot in #171
• Tensorboard callbacks by @arvindmuralie77 in #173
• update docs 📝 by @aniketmaurya in #175
• anti pattern fixes from deepsource by @skp-github in #176
• [pre-commit.ci] pre-commit suggestions by @pre-commit-ci in #177
• Demo & fixes by @aniketmaurya in #180
• Upgrade deps by @aniketmaurya in #181
• 🐛 minor bug fixes and reformat by @aniketmaurya in #182
• fix examples and Flash trainer by @aniketmaurya in #183
• refactor apis by @aniketmaurya in #184
• Fix examples by @aniketmaurya in #185

New Contributors

• @snyk-bot made their first contribution in #171
• @arvindmuralie77 made their first contribution in #173
• @skp-github made their first contribution in #176

Full Changelog: v0.0.8...v0.0.8.post1

v0.0.8

What's Changed

• 🤩 refactor core by @aniketmaurya in #136
• cleanup APIs by @aniketmaurya in #137
• added conda installation instruction by @sugatoray in #144
• recursively exclude tests folder and its contents by @sugatoray in #141
• add model.save test by @aniketmaurya in #147
• remove redundant to_item by @aniketmaurya in #152
• refactor Tracker by @aniketmaurya in #153
• Change methods not using its bound instance to staticmethods by @deepsource-autofix in #156
• refactor metrics by @aniketmaurya in #159
• add dataoader length by @aniketmaurya in #160
• fix model checkpoint folder not found by @aniketmaurya in #162
• Fix metrics update by @aniketmaurya in #163
• Replace multiple == checks with in by @deepsource-autofix in #167
• increment current_epoch after each epoch by @aniketmaurya in #169
• Wandb Implementation by @aniketmaurya in #168

Full Changelog: v0.0.7...v0.0.8

v0.0.8.dev1

What's Changed

• optional pl dependency by @aniketmaurya in #133
• minor fixes by @aniketmaurya in #134
• 📚 update example by @aniketmaurya in #135
• 🤩 refactor core by @aniketmaurya in #136
• cleanup APIs by @aniketmaurya in #137
• remove dependencies by @aniketmaurya in #139
• added conda installation instruction by @sugatoray in #144
• recursively exclude tests folder and its contents by @sugatoray in #141
• pin flash version to 0.5.1 by @aniketmaurya in #143
• refactor backend by @aniketmaurya in #146
• Remove flit by @aniketmaurya in #148
• add model.save test by @aniketmaurya in #147
• Update main.yml by @aniketmaurya in #149
• remove redundant to_item by @aniketmaurya in #152
• [pre-commit.ci] pre-commit suggestions by @pre-commit-ci in #155
• refactor Tracker by @aniketmaurya in #153
• Change methods not using its bound instance to staticmethods by @deepsource-autofix in #156
• 📝 fix documentation & examples by @aniketmaurya in #158
• refactor metrics by @aniketmaurya in #159
• add dataoader length by @aniketmaurya in #160
• fix model checkpoint folder not found by @aniketmaurya in #162
• Fix metrics update by @aniketmaurya in #163

Full Changelog: v0.0.7...v0.0.8.dev1

v0.0.7.post2

What's Changed

• optional pl dependency by @aniketmaurya in #133
• minor fixes by @aniketmaurya in #134
• 📚 update example by @aniketmaurya in #135
• 🤩 refactor core by @aniketmaurya in #136
• cleanup APIs by @aniketmaurya in #137

Full Changelog: v0.0.7...v0.0.7.post2

Improved Callback and Loggers

Highlights

• ☄️ comet integration #129
• add model checkpoint callback #121
• 📝 add csv logger #116
• 🚀 add train_eval_callback #111
• 🪄 add Average Meter #109
• fix device issue in metric calculation PR #106

What's Changed

• [pre-commit.ci] pre-commit suggestions by @pre-commit-ci in #98
• clean docs by @aniketmaurya in #100
• fix optimizer by @aniketmaurya in #102
• fix device issue in metric calculation by @aniketmaurya in #106
• ✨ refactor tuner by @aniketmaurya in #107
• 📚 add example & better intro by @aniketmaurya in #108
• 🪄 add Average Meter by @aniketmaurya in #109
• Fixes metrics device handling by @aniketmaurya in #110
• 🚀 add train_eval_callback by @aniketmaurya in #111
• refactor training callback by @aniketmaurya in #112
• 🔥Data refactor by @aniketmaurya in #113
• 🚀 Bump test coverage by @aniketmaurya in #114
• fix live display error on multiple runs by @aniketmaurya in #115
• 📝 add csv logger by @aniketmaurya in #116
• ✨refactor base model by @aniketmaurya in #118
• ☘️ add codecarbon callback by @aniketmaurya in #119
• add model checkpoint callback by @aniketmaurya in #121
• fix loss compile by @aniketmaurya in #126
• add requirement decorator by @aniketmaurya in #128
• Refactor callback runner by @aniketmaurya in #130
• ☄️ comet integration by @aniketmaurya in #129

New Contributors

• @pre-commit-ci made their first contribution in #98

Full Changelog: v0.0.6...v0.0.7

Model Training & Auto Tuner

0.0.6

• 🎉 Revamp Callbacks and Training #94
• ✨ refactor data handling 📝 docs update. PR #91
• integrate torchmetrics. PR #80
• callbacks & 🤑 ProgressCallback. PR #76
• 🔥 Add AutoModel Tuner. PR #74
• refactor APIs - Simplify API & add model.compile(...). PR #73
• 🤗 integrate HF Accelerator. PR #71

0.0.5

What is New?

• Keras style Model Training API 🎉
• Remote Dataset Loader - create dataloader from any cloud Bucket ☁️
• Datagenerator for Image Classification (more to come soon...)

📚 Documentation & Examples

To install $ pip install -U gradsflow

Release Notes

• 🔥 Add custom training loop with model.fit. PR #63 Done by @aniketmaurya
• ☁️ Add ray.data - remote dataset loader. PR #61 Done by @aniketmaurya
• 🎉 Add AutoDataset - Encapsulate datamodule and dataloaders. PR #59 Done by @aniketmaurya
• 🌟 Add Autotask feature. PR #54 Done by @gagan3012
• ✨ Add AutoTrainer to support plain torch training loop and other torch frameworks. PR #53

0.0.5a0

Release Notes

🔥 Add custom training loop with model.fit. PR #63 Done by @aniketmaurya
☁️ Add ray.data - remote dataset loader. PR #61 Done by @aniketmaurya
🎉 Add AutoDataset - Encapsulate datamodule and dataloaders. PR #59 Done by @aniketmaurya
🌟 Add Autotask feature. PR #54 Done by @gagan3012
✨ Add AutoTrainer to support plain torch training loop and other torch frameworks. PR #53

$ pip install gradsflow==0.0.5a0

📚 Get started with Documentation

0.0.4

Latest Changes

• fix best checkpoints model loading #52
• 🚀 feature/fix train arguments docs #44
• Publish Python 🐍 distributions 📦 to PyPI #42

Distributed HPO with Ray ♾

0.0.3

Latest Changes

• add optuna visualizations 🎨 . PR #27 by @aniketmaurya.
• add max_steps for HPO. PR #25 by @aniketmaurya.
• 📝 update docs & license. PR #23 by @aniketmaurya.
• fetch best trial model. PR #21 by @aniketmaurya.
• migrate to ray_tune 🌟. Read more here. PR #36 by @aniketmaurya.
• render jupyter notebooks in documentation. PR #38 by @aniketmaurya.
• remove optuna #39
• Publish Python 🐍 distributions 📦 to PyPI #42

Get Started Now

📚 Documentation: docs.gradsflow.com

$ pip install -U gradsflow

Example

from gradsflow import AutoImageClassifier

from flash.core.data.utils import download_data
from flash.image import ImageClassificationData


data_dir = "/Users/aniket/personal/gradsflow/gradsflow/data/"
download_data("https://pl-flash-data.s3.amazonaws.com/hymenoptera_data.zip", data_dir)
datamodule = ImageClassificationData.from_folders(
    train_folder=f"{data_dir}/hymenoptera_data/train/",
    val_folder=f"{data_dir}/hymenoptera_data/val/",
)

model = AutoImageClassifier(
    datamodule,
    max_epochs=2,
    n_trials=4,
    optimization_metric="val_accuracy",
    timeout=50,
)

print("AutoImageClassifier initialised!")
model.hp_tune(gpu=1)