time-series-model's Introduction

Time series model using tensorflow

General info
Technologies
Install

General info

This repository contains models that have been converted from image models and layers to time series.

Technologies

python 3.7 ~ 3.10
tensorflow 2.7.0
numpy 1.21.4

Install

install package

cd time-series-model-main
python setup.py

example

import tftime

model = tftime.build_model("sam_efficientnet_b2", (30, 1), 2, None)

import tftime
import tensorflow as tf

inputs = tf.keras.layers.Input((30, 1))

x = tftime.block.ConvBlock(128, "conv1d", "resnet")(inputs)
x = tf.keras.layers.GlobalAvgPool1D()(x)
x = tftime.layers.Output(5, "softmax")

model = tftime.model.SAMModel(inputs, x)
model.compile("adam", categorical_crossentropy, ["accuracy"])

model.fit(x, y)

available model

print(tftime.available_network)

['efficientnet b0 ~ b8' 'sam_efficientnet b0 ~ b8'
 'dense_efficientnet b0 ~ b8' 'sam_dense_efficientnet b0 ~ b8'
 'lambda_efficientnet b0 ~ b8' 'sam_lambda_efficientnet b0 ~ b8'
 'efficientnetv2 b0 ~ b8' 'sam_efficientnetv2 b0 ~ b8' 'resnet b0 ~ b8'
 'sam_resnet b0 ~ b8' 'se_resnet b0 ~ b8' 'sam_se_resnet b0 ~ b8'
 'densenet b0 ~ b8' 'sam_densenet b0 ~ b8' 'se_densenet b0 ~ b8'
 'sam_se_densenet b0 ~ b8' 'lambda_resnet b0 ~ b8'
 'sam_lambda_resnet b0 ~ b8' 'se_lambda_resnet b0 ~ b8'
 'sam_se_lambda_resnet b0 ~ b8' 'convnext b0 ~ b8' 'sam_convnext b0 ~ b8'
 'se_convnext b0 ~ b8' 'sam_se_convnext b0 ~ b8' 'lambda_convnext b0 ~ b8'
 'sam_lambda_convnext b0 ~ b8' 'se_lambda_convnext b0 ~ b8'
 'sam_se_lambda_convnext b0 ~ b8']

Recommend Projects