Build Low Code Automated Tensorflow, What-IF explainable models in just 3 lines of code.

Overview

Downloads Generic badge Generic badge example workflow Open issues

Auto Tensorflow - Mission:

Build Low Code Automated Tensorflow, What-IF explainable models in just 3 lines of code.

To make Deep Learning on Tensorflow absolutely easy for the masses with its low code framework and also increase trust on ML models through What-IF model explainability.

Under the hood:

Built on top of the powerful Tensorflow ecosystem tools like TFX , TF APIs and What-IF Tool , the library automatically does all the heavy lifting internally like EDA, schema discovery, feature engineering, HPT, model search etc. This empowers developers to focus only on building end user applications quickly without any knowledge of Tensorflow, ML or debugging. Built for handling large volume of data / BigData - using only TF scalable components. Moreover the models trained with auto-tensorflow can directly be deployed on any cloud like GCP / AWS / Azure.

Official Launch: https://youtu.be/sil-RbuckG0

Features:

  1. Build Classification / Regression models on CSV data
  2. Automated Schema Inference
  3. Automated Feature Engineering
    • Discretization
    • Scaling
    • Normalization
    • Text Embedding
    • Category encoding
  4. Automated Model build for mixed data types( Continuous, Categorical and Free Text )
  5. Automated Hyper-parameter tuning
  6. Automated GPU Distributed training
  7. Automated UI based What-IF analysis( Fairness, Feature Partial dependencies, What-IF )
  8. Control over complexity of model
  9. No dependency over Pandas / SKLearn
  10. Can handle dataset of any size - including multiple CSV files

Tutorials:

  1. Open In Colab - Auto Classification on CSV data
  2. Open In Colab - Auto Regression on CSV data

Setup:

  1. Install library
    • PIP(Recommended): pip install auto-tensorflow
    • Nightly: pip install git+https://github.com/rafiqhasan/auto-tensorflow.git
  2. Works best on UNIX/Linux/Debian/Google Colab/MacOS

Usage:

  1. Initialize TFAuto Engine
from auto_tensorflow.tfa import TFAuto
tfa = TFAuto(train_data_path='/content/train_data/', test_data_path='/content/test_data/', path_root='/content/tfauto')
  1. Step 1 - Automated EDA and Schema discovery
tfa.step_data_explore(viz=True) ##Viz=False for no visualization
  1. Step 2 - Automated ML model build and train
tfa.step_model_build(label_column = 'price', model_type='REGRESSION', model_complexity=1)
  1. Step 3 - Automated What-IF Tool launch
tfa.step_model_whatif()

API Arguments:

  • Method TFAuto

    • train_data_path: Path where training data is stored
    • test_data_path: Path where Test / Eval data is stored
    • path_root: Directory for running TFAuto( Directory should NOT exist )
  • Method step_data_explore

    • viz: Is data visualization required ? - True or False( Default )
  • Method step_model_build

    • label_column: The feature to be used as Label
    • model_type: Either of 'REGRESSION'( Default ), 'CLASSIFICATION'
    • model_complexity:
      • 0 : Model with default hyper-parameters
      • 1 (Default): Model with automated hyper-parameter tuning
      • 2 : Complexity 1 + Advanced fine-tuning of Text layers

Current limitations:

There are a few limitations in the initial release but we are working day and night to resolve these and add them as future features.

  1. Doesn't support Image / Audio data

Future roadmap:

  1. Add support for Timeseries / Audio / Image data
  2. Add feature to download full pipeline model Python code for advanced tweaking

Release History:

1.3.2 - 27/11/2021 - Release Notes

1.3.1 - 18/11/2021 - Release Notes

1.2.0 - 24/07/2021 - Release Notes

1.1.1 - 14/07/2021 - Release Notes

1.0.1 - 07/07/2021 - Release Notes

Comments
  • Failed to install 1.2.0

    Failed to install 1.2.0

    Describe the bug Does not resolve dependency 👍 Show error when I run; pip install auto-tensorflow I got this message: Could not find a version that matches keras-nightly~=2.5.0.dev

    To Reproduce Steps to reproduce the behavior: pip install auto-tensorflow Expected behavior Install auto-tensorflow

    Versions:

    • Auto-Tensorflow:1.2.0
    • Tensorflow:
    • Tensorflow-Extended:

    Additional context Add any other context about the problem here.

    wontfix 
    opened by HenrryVargas 8
  • Colab Regression Example No Longer Working?

    Colab Regression Example No Longer Working?

    Trying to run the Colab Regression notebook. All dependencies get installed, I Restart and Run All to start the code. It errors out here:

    ##Step 1
    ##Run Data setup -> Infer Schema, find anomalies, create profile and show viz
    tfa.step_data_explore(viz=False)
    
    Data: Pipeline execution started...
    WARNING:apache_beam.runners.interactive.interactive_environment:Dependencies required for Interactive Beam PCollection visualization are not available, please use: `pip install apache-beam[interactive]` to install necessary dependencies to enable all data visualization features.
    WARNING:apache_beam.io.tfrecordio:Couldn't find python-snappy so the implementation of _TFRecordUtil._masked_crc32c is not as fast as it could be.
    WARNING:apache_beam.io.tfrecordio:Couldn't find python-snappy so the implementation of _TFRecordUtil._masked_crc32c is not as fast as it could be.
    ERROR:absl:Execution 2 failed.
    ---------------------------------------------------------------------------
    TypeCheckError                            Traceback (most recent call last)
    [<ipython-input-6-7e17a616f197>](https://localhost:8080/#) in <module>
          1 ##Step 1
          2 ##Run Data setup -> Infer Schema, find anomalies, create profile and show viz
    ----> 3 tfa.step_data_explore(viz=False)
    
    14 frames
    [/usr/local/lib/python3.7/dist-packages/auto_tensorflow/tfa.py](https://localhost:8080/#) in step_data_explore(self, viz)
       1216     Viz: (False) Is data visualization required ?
       1217     '''
    -> 1218     self.pipeline = self.tfadata.run_initial(self._train_data_path, self._test_data_path, self._tfx_root, self._metadata_db_root, self.tfautils, viz)
       1219     self.generate_config_json()
       1220 
    
    [/usr/local/lib/python3.7/dist-packages/auto_tensorflow/tfa.py](https://localhost:8080/#) in run_initial(self, _train_data_path, _test_data_path, _tfx_root, _metadata_db_root, tfautils, viz)
        211     #Run data pipeline
        212     print("Data: Pipeline execution started...")
    --> 213     LocalDagRunner().run(self.pipeline)
        214     self._run = True
        215 
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/portable/tfx_runner.py](https://localhost:8080/#) in run(self, pipeline)
         76     c = compiler.Compiler()
         77     pipeline_pb = c.compile(pipeline)
    ---> 78     return self.run_with_ir(pipeline_pb)
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/local/local_dag_runner.py](https://localhost:8080/#) in run_with_ir(self, pipeline)
         85           with metadata.Metadata(connection_config) as mlmd_handle:
         86             partial_run_utils.snapshot(mlmd_handle, pipeline)
    ---> 87         component_launcher.launch()
         88         logging.info('Component %s is finished.', node_id)
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/portable/launcher.py](https://localhost:8080/#) in launch(self)
        543               executor_watcher.address)
        544           executor_watcher.start()
    --> 545         executor_output = self._run_executor(execution_info)
        546       except Exception as e:  # pylint: disable=broad-except
        547         execution_output = (
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/portable/launcher.py](https://localhost:8080/#) in _run_executor(self, execution_info)
        418     outputs_utils.make_output_dirs(execution_info.output_dict)
        419     try:
    --> 420       executor_output = self._executor_operator.run_executor(execution_info)
        421       code = executor_output.execution_result.code
        422       if code != 0:
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/portable/beam_executor_operator.py](https://localhost:8080/#) in run_executor(self, execution_info, make_beam_pipeline_fn)
         96         make_beam_pipeline_fn=make_beam_pipeline_fn)
         97     executor = self._executor_cls(context=context)
    ---> 98     return python_executor_operator.run_with_executor(execution_info, executor)
    
    [/usr/local/lib/python3.7/dist-packages/tfx/orchestration/portable/python_executor_operator.py](https://localhost:8080/#) in run_with_executor(execution_info, executor)
         57   output_dict = copy.deepcopy(execution_info.output_dict)
         58   result = executor.Do(execution_info.input_dict, output_dict,
    ---> 59                        execution_info.exec_properties)
         60   if not result:
         61     # If result is not returned from the Do function, then try to
    
    [/usr/local/lib/python3.7/dist-packages/tfx/components/statistics_gen/executor.py](https://localhost:8080/#) in Do(self, input_dict, output_dict, exec_properties)
        138             stats_api.GenerateStatistics(stats_options)
        139             | 'WriteStatsOutput[%s]' % split >>
    --> 140             stats_api.WriteStatisticsToBinaryFile(output_path))
        141         logging.info('Statistics for split %s written to %s.', split,
        142                      output_uri)
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/pvalue.py](https://localhost:8080/#) in __or__(self, ptransform)
        135 
        136   def __or__(self, ptransform):
    --> 137     return self.pipeline.apply(ptransform, self)
        138 
        139 
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/pipeline.py](https://localhost:8080/#) in apply(self, transform, pvalueish, label)
        651     if isinstance(transform, ptransform._NamedPTransform):
        652       return self.apply(
    --> 653           transform.transform, pvalueish, label or transform.label)
        654 
        655     if not isinstance(transform, ptransform.PTransform):
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/pipeline.py](https://localhost:8080/#) in apply(self, transform, pvalueish, label)
        661       old_label, transform.label = transform.label, label
        662       try:
    --> 663         return self.apply(transform, pvalueish)
        664       finally:
        665         transform.label = old_label
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/pipeline.py](https://localhost:8080/#) in apply(self, transform, pvalueish, label)
        710 
        711       if type_options is not None and type_options.pipeline_type_check:
    --> 712         transform.type_check_outputs(pvalueish_result)
        713 
        714       for tag, result in ptransform.get_named_nested_pvalues(pvalueish_result):
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/transforms/ptransform.py](https://localhost:8080/#) in type_check_outputs(self, pvalueish)
        464 
        465   def type_check_outputs(self, pvalueish):
    --> 466     self.type_check_inputs_or_outputs(pvalueish, 'output')
        467 
        468   def type_check_inputs_or_outputs(self, pvalueish, input_or_output):
    
    [/usr/local/lib/python3.7/dist-packages/apache_beam/transforms/ptransform.py](https://localhost:8080/#) in type_check_inputs_or_outputs(self, pvalueish, input_or_output)
        495                 hint=hint,
        496                 actual_type=pvalue_.element_type,
    --> 497                 debug_str=type_hints.debug_str()))
        498 
        499   def _infer_output_coder(self, input_type=None, input_coder=None):
    
    TypeCheckError: Output type hint violation at WriteStatsOutput[train]: expected <class 'apache_beam.pvalue.PDone'>, got <class 'str'>
    Full type hint:
    IOTypeHints[inputs=((<class 'tensorflow_metadata.proto.v0.statistics_pb2.DatasetFeatureStatisticsList'>,), {}), outputs=((<class 'apache_beam.pvalue.PDone'>,), {})]
    File "<frozen importlib._bootstrap>", line 677, in _load_unlocked
    File "<frozen importlib._bootstrap_external>", line 728, in exec_module
    File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
    File "/usr/local/lib/python3.7/dist-packages/tensorflow_data_validation/api/stats_api.py", line 113, in <module>
        class WriteStatisticsToBinaryFile(beam.PTransform):
    File "/usr/local/lib/python3.7/dist-packages/apache_beam/typehints/decorators.py", line 776, in annotate_input_types
        *converted_positional_hints, **converted_keyword_hints)
    
    based on:
      IOTypeHints[inputs=None, outputs=((<class 'apache_beam.pvalue.PDone'>,), {})]
      File "<frozen importlib._bootstrap>", line 677, in _load_unlocked
      File "<frozen importlib._bootstrap_external>", line 728, in exec_module
      File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
      File "/usr/local/lib/python3.7/dist-packages/tensorflow_data_validation/api/stats_api.py", line 113, in <module>
          class WriteStatisticsToBinaryFile(beam.PTransform):
      File "/usr/local/lib/python3.7/dist-packages/apache_beam/typehints/decorators.py", line 863, in annotate_output_types
          f._type_hints = th.with_output_types(return_type_hint)  # pylint: disable=protected-access
    
    opened by windowshopr 2
  • Dump when training Text column model on GPUs

    Dump when training Text column model on GPUs

    Describe the bug The model dumps with error when training a model on GPU runtime

    To Reproduce Train a model with Free text column on GPU device

    Expected behavior Should not give any error

    Versions:

    • Auto-Tensorflow: 1.0.1
    • Tensorflow: 2.5.0
    • Tensorflow-Extended: 0.29.0

    Additional context Add any other context about the problem here.

    bug 
    opened by rafiqhasan 2
  • Add automated - advanced feature engineering

    Add automated - advanced feature engineering

    Is your feature request related to a problem? Please describe. Yes

    Describe the solution you'd like Add more feature engineering options for automated consideration:

    1. Squared
    2. Square root
    3. Min-Max scaling( Normalization is already there )
    4. etc

    Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

    Additional context Add any other context or screenshots about the feature request here.

    enhancement 
    opened by rafiqhasan 1
  • Known limitations

    Known limitations

    There are a few limitations in the initial release but we are working day and night to resolve these and add them as future features.

    1. Doesn't support Image / Audio data
    2. Doesn't support - quote delimited CSVs( TFX doesn't support qCSV yet )
    3. Classification only supports integer labels from 0 to N
    enhancement 
    opened by rafiqhasan 1
  • When AutoTF will be released for Time Series ?

    When AutoTF will be released for Time Series ?

    Is your feature request related to a problem? Please describe. A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

    Describe the solution you'd like A clear and concise description of what you want to happen.

    Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

    Additional context Add any other context or screenshots about the feature request here.

    enhancement 
    opened by gulabpatel 1
Releases(1.3.4)
  • 1.3.4(Dec 9, 2022)

    • Fixed bugs
    • Cleaned up PIP dependencies for faster installation

    Full Changelog: https://github.com/rafiqhasan/auto-tensorflow/compare/1.3.3...1.3.4

    Source code(tar.gz)
    Source code(zip)
  • 1.3.3(Dec 9, 2022)

  • 1.3.2(Nov 26, 2021)

    • Added bucketization feature engineering
    • Added more diverse HPT options
    • Replaced RELU with SELU
    • Better accuracy on regression models
    • Changed HPT objective for classification models
    • Multiple improvisations for higher accuracy models

    Full Changelog: https://github.com/rafiqhasan/auto-tensorflow/compare/1.3.1...1.3.2

    Source code(tar.gz)
    Source code(zip)
  • 1.3.1(Nov 18, 2021)

    Features:

    1. Upgraded to TF 2.6.0
    2. Upgraded to TFX 1.4.0
    3. Added new feature engineering functions
    4. Added capability to handle multiple line CSVs
    5. Keras Tuner functionality now more optimised and HPT runs faster
    Source code(tar.gz)
    Source code(zip)
  • 1.2.0(Jul 24, 2021)

    1.2.0 - 07/24/2021

    • Upgraded to TFX 1.0.0
    • Major performance fixes
    • Fixed bugs
    • Added more features:
      • TFX CSVExampleGen speedup
      • Added more feature engineering options
    Source code(tar.gz)
    Source code(zip)
  • 1.1.1(Jul 20, 2021)

    1.1.1 - 07/14/2021

    • Fixed bugs
    • Added more features:
      • Added complexity = 2 for automated tunable textual layers
      • Textual label for Classification
      • Imbalanced label handling
      • GPU fixes
    Source code(tar.gz)
    Source code(zip)
  • 1.0.1(Jul 20, 2021)

Owner
Hasan Rafiq
Technology enthusiast working @ Google: Google Cloud, Machine Learning, Tensorflow, Python
Hasan Rafiq
Turning SymPy expressions into PyTorch modules.

sympytorch A micro-library as a convenience for turning SymPy expressions into PyTorch Modules. All SymPy floats become trainable parameters. All SymP

Patrick Kidger 89 Dec 13, 2022
Neural Logic Inductive Learning

Neural Logic Inductive Learning This is the implementation of the Neural Logic Inductive Learning model (NLIL) proposed in the ICLR 2020 paper: Learn

36 Nov 28, 2022
duralava is a neural network which can simulate a lava lamp in an infinite loop.

duralava duralava is a neural network which can simulate a lava lamp in an infinite loop. Example This is not a real lava lamp but a "fake" one genera

Maximilian Bachl 87 Dec 20, 2022
Applying CLIP to Point Cloud Recognition.

PointCLIP: Point Cloud Understanding by CLIP This repository is an official implementation of the paper 'PointCLIP: Point Cloud Understanding by CLIP'

Renrui Zhang 175 Dec 24, 2022
FEMDA: Robust classification with Flexible Discriminant Analysis in heterogeneous data

FEMDA: Robust classification with Flexible Discriminant Analysis in heterogeneous data. Flexible EM-Inspired Discriminant Analysis is a robust supervised classification algorithm that performs well i

0 Sep 06, 2022
Pytorch implementation of the paper Improving Text-to-Image Synthesis Using Contrastive Learning

T2I_CL This is the official Pytorch implementation of the paper Improving Text-to-Image Synthesis Using Contrastive Learning Requirements Linux Python

42 Dec 31, 2022
Real-Time Seizure Detection using EEG: A Comprehensive Comparison of Recent Approaches under a Realistic Setting

Real-Time Seizure Detection using Electroencephalogram (EEG) This is the repository for "Real-Time Seizure Detection using EEG: A Comprehensive Compar

AITRICS 30 Dec 17, 2022
Chinese license plate recognition

AgentCLPR 简介 一个基于 ONNXRuntime、AgentOCR 和 License-Plate-Detector 项目开发的中国车牌检测识别系统。 车牌识别效果 支持多种车牌的检测和识别(其中单层车牌识别效果较好): 单层车牌: [[[[373, 282], [69, 284],

AgentMaker 26 Dec 25, 2022
[SIGGRAPH 2022 Journal Track] AvatarCLIP: Zero-Shot Text-Driven Generation and Animation of 3D Avatars

AvatarCLIP: Zero-Shot Text-Driven Generation and Animation of 3D Avatars Fangzhou Hong1*  Mingyuan Zhang1*  Liang Pan1  Zhongang Cai1,2,3  Lei Yang2 

Fangzhou Hong 749 Jan 04, 2023
Code for "The Intrinsic Dimension of Images and Its Impact on Learning" - ICLR 2021 Spotlight

dimensions Estimating the instrinsic dimensionality of image datasets Code for: The Intrinsic Dimensionaity of Images and Its Impact On Learning - Phi

Phil Pope 41 Dec 10, 2022
This repo provides the base code for pytorch-lightning and weight and biases simultaneous integration.

Write your model faster with pytorch-lightning-wadb-code-backbone This repository provides the base code for pytorch-lightning and weight and biases s

9 Mar 29, 2022
Exemplo de implementação do padrão circuit breaker em python

fast-circuit-breaker Circuit breakers existem para permitir que uma parte do seu sistema falhe sem destruir todo seu ecossistema de serviços. Michael

James G Silva 17 Nov 10, 2022
End-to-End Object Detection with Fully Convolutional Network

This project provides an implementation for "End-to-End Object Detection with Fully Convolutional Network" on PyTorch.

472 Dec 22, 2022
Related resources for our EMNLP 2021 paper

Plan-then-Generate: Controlled Data-to-Text Generation via Planning Authors: Yixuan Su, David Vandyke, Sihui Wang, Yimai Fang, and Nigel Collier Code

Yixuan Su 61 Jan 03, 2023
Single Image Random Dot Stereogram for Tensorflow

TensorFlow-SIRDS Single Image Random Dot Stereogram for Tensorflow SIRDS is a means to present 3D data in a 2D image. It allows for scientific data di

Greg Peatfield 5 Aug 10, 2022
A new codebase for Group Activity Recognition. It contains codes for ICCV 2021 paper: Spatio-Temporal Dynamic Inference Network for Group Activity Recognition and some other methods.

Spatio-Temporal Dynamic Inference Network for Group Activity Recognition The source codes for ICCV2021 Paper: Spatio-Temporal Dynamic Inference Networ

40 Dec 12, 2022
Source code for our Paper "Learning in High-Dimensional Feature Spaces Using ANOVA-Based Matrix-Vector Multiplication"

NFFT4ANOVA Source code for our Paper "Learning in High-Dimensional Feature Spaces Using ANOVA-Based Matrix-Vector Multiplication" This package uses th

Theresa Wagner 1 Aug 10, 2022
An Unpaired Sketch-to-Photo Translation Model

Unpaired-Sketch-to-Photo-Translation We have released our code at https://github.com/rt219/Unsupervised-Sketch-to-Photo-Synthesis This project is the

38 Oct 28, 2022
Spontaneous Facial Micro Expression Recognition using 3D Spatio-Temporal Convolutional Neural Networks

Spontaneous Facial Micro Expression Recognition using 3D Spatio-Temporal Convolutional Neural Networks Abstract Facial expression recognition in video

Bogireddy Sai Prasanna Teja Reddy 103 Dec 29, 2022
PyTorch implementation of the ExORL: Exploratory Data for Offline Reinforcement Learning

ExORL: Exploratory Data for Offline Reinforcement Learning This is an original PyTorch implementation of the ExORL framework from Don't Change the Alg

Denis Yarats 52 Jan 01, 2023