Tamamlanmış

simple project TensorFlow. From a spreadsheet it learn to predict values with CNN

This project is a simple project TensorFlow. From a spreadsheet it learn to predict values with a single tensorflow CNN.

LoadData

Parameter CNNDatasAll are an object pandas with columns names, all values are numerical and are normalized. It contains all datas for training

Parameter PercentageOfEvaluationData is how many % of data will be for evaluation (shuffle)

Parameter PercentageOfTrainingData is how many % of data will be for training (shuffle)

Parameter NamesOfColumnsOutput is an array with the list of name of column which are the CNN output to predict

Build CNNDataForInput from CNNDatasAll by removing PercentageOfEvaluationData % of lines randomly, and move this removed lines into CNNDataInputForEvaluation

With pandas, Remove columns NamesOfColumnsOutput from CNNDataForInput and move theses column to CNNDataForOutput

With pandas, Remove columns NamesOfColumnsOutput from CNNDataInputForEvaluation and move theses column to CNNDataOutputForEvaluation

Maybe with : Xtrain, Xval, ytrain, yval = train_test_split( [login to view URL]( columns=["quality", 'type'] ), AllData[["quality", 'type']], test_size=%%% )

BuildTrainCNN

Parameter : CNNDataForInput , CNNDataForOutput

Parameter : MaxLinesForTraining or default=0=All lines in CNNDataForInput and CNNDataForOutput

Parameter : batch_size, Start_epochs (To resume a previous training), End_epochs

Parameter : CNNShape (Simple array 2D, one line=one layer, on each line : 4 parameters for the layer CNN, The parameters are : size in neurons, layer name , activation=softmax,elu,selu,softplus,softsign,relu,tanh,sigmoid,hard_sigmoid,exponential,linear. LayerType = Conv3D,Conv2D,Conv1D,Dense,AveragePooling3D)

Build CNN into tensorflow

Compile CNN with 2 parameters : Loss, Optimizer

loss = absolute_difference, cosine_distance, hinge_loss, huber_loss, log_loss, mean_squared_error

optimizer = Adadelta,Adagrad,Adam,Adamax,Nadam,RMSprop,SGD

Run CNN training with parameters

It must be possible to continue training, the parameter Start_epoch is 0 the first run, then it is >0 if the training continue

SaveCNN

Save CNN on disk in json : Parameters CNNShape, NamesOfColumnsOutput (columns names of CNNDataForOutput), Loss, Optimizer, NamesOfColumnsInput (columns names of CNNDataForInput), CNN Weights

LoadBuildCNN

Load CNN from disk json : Parameters CNNShape, NamesOfColumnsOutput (columns names of CNNDataForOutput), Loss, Optimizer, NamesOfColumnsInput (columns names of CNNDataForInput), CNN Weights

Build CNN into tensorflow

Compile CNN

EvalCNN

Parameter : MaxLinesToEvaluate or default=0=All lines in CNNDataInputForEvaluation and CNNDataOutputForEvaluation

Run Evaluation on CNNDataInputForEvaluation,CNNDataOutputForEvaluation

Return : Weight for each inputs (for each input, need to average the weight of all first layer neurons connected to each input)

Return : Accuracy

Return : precision(true_positives, true_negatives, false_positives and false_negatives)

PredictCNN

Parameter : Pandas-Data with same columns names as parameter : NamesOfColumnsInput

Load CNN from disk json : Parameters CNNShape, NamesOfColumnsOutput (columns names of CNNDataForOutput), Loss, Optimizer, NamesOfColumnsInput (columns names of CNNDataForInput), CNN Weights

Build CNN into tensorflow

Compile CNN

RunCNN and add predicted values : NamesOfColumnsOutput into the Pandas-Data

TensorBoard

Tensorboard must be available with all metrics during eval and training

Beceriler: Neural Networks, Tensorflow

Daha fazlasını gör: tensorflow tutorial python, tensorflow tutorial for beginners, tensorflow basics, tensorflow cnn, keras tutorial, tensorflow tutorial for beginners pdf, tensorflow tutorial pdf, tensorflow neural network, php mysql simple project database, simple project php, linked list simple project java, simple project vba using sql, simple project jsp servlet, simple project management script, possible simple project, simple project timekeeping program, simple project management php script, simple project marketing research, simple project spreadsheet, project management from simple to complex pdf

İşveren Hakkında:
( 137 değerlendirme ) --, Ukraine

Proje NO: #19410550

Seçilen:

anveshrddy3

Hey! I have worked on a similar project before and I am sure I can help you out with your problem. I am also certified in both CNNs and Tensorflow. Let's talk about the problem in detail over chat. Cheers!

%selectedBids___i_period_sub_7% gün içinde 83%project_currencyDetails_sign_sub_9% %project_currencyDetails_code_sub_10%
(0 Değerlendirme)
0.0

Bu iş için 4 freelancer ortalamada $183 teklif veriyor

WorldOfScience

I can do this project efficiently and effectively within the designated time period and i have all of skills in AI and Machine [login to view URL] i have supervised hundreds of Research projects related to Neural Networks a Daha Fazla

$100 USD in 6 gün içinde
(2 Değerlendirme)
3.8
usamamw141

Hey there, I am Usama and I am expert in CNN, Python, Tensorflow and few other Deep learning frameworks. Feel free to inbox me now so that we can discuss about project, deadlines and budget. Also tell me will csv have Daha Fazla

$150 USD in 3 gün içinde
(4 Değerlendirme)
3.6
muller1998926

Hello, dear! I am interested in your project. I have extensive experience in time series analysis, data forecasting, stock market prediction, data cleaning, data transformation, classification and clustering, feature Daha Fazla

$400 USD in 3 gün içinde
(2 Değerlendirme)
2.6