All deep learning models were trained and analyzed by using the 64-bit Ubuntu 16.04.5 LTS operating system on a workstation with 8 GB memory and an NVIDIA GeForce GTX 1080 8 GB graphics processing unit. The optimizer, weight decay, and momentum were common among all the CNNs. In this study, the optimizer used stochastic gradient descent, with a weight decay of 0 and momentum of 0.9. Learning rates of 0.001 and 0.01 were used for both ResNet and EfficientNet. All the models analyzed a maximum of 100 epochs. We used the early stopping method to terminate the data training to prevent overfitting if the validation error did not update 20 times in a row. This process was performed 30 times on all CNN models for statistical evaluation.
Geforce gtx 1080 8 gb
Manufactured by NVIDIA
The GeForce GTX 1080 is a high-performance graphics processing unit (GPU) with 8 GB of video memory. It is designed for use in computer systems for various applications that require advanced graphics processing capabilities.
Automatically generated - may contain errors
Lab products found in correlation
3 protocols using geforce gtx 1080 8 gb
1
Comparative Analysis of CNN Models
2
Deep Learning Environment Setup
Check if the same lab product or an alternative is used in the 5 most similar protocols
+ Expand
Python 3.6 was used in the development of the model. The algorithms were trained and validated using Google Colaboratory. In this study, a personal computer (PC) with an Intel I7 8700 K 3.70 GHz processor, 32 GB DDR4 RAM, Nvidia GeForce GTX 1080 8 GB, Anaconda with Python, and TensorFlow was used. The latest Anaconda was downloaded from https://www.anaconda.com/products/distribution accessed on 1 July 2022. This study aimed to create a Conda and Deep Learning Conda environment. A PC usually has a graphic card and a graphics processing unit (GPU) environment that can be used for deep learning.
3
Quantitative Phase Imaging Deep Learning
Check if the same lab product or an alternative is used in the 5 most similar protocols
+ Expand
For this research we used ResNet-50 [16] initiated with weights pre-trained on ImageNet [17] (source: github/fchollet). We added a global average pooling (GAP) layer after the last convolutional layer, followed by 2 neurons activated by sigmoid functions (Fig. 2b). Our resulting model, therefore, outputs 2 separate estimations representing 0-1 scaled mKO2 and mAG sum fluorescence levels for each 3-frame QPI clip passed to the model. To avoid overfitting during training we introduced 50% dropout after the GAP layer. We also performed data augmentation using simple mirroring and rotations to increase the training set size and make our model more robust to changes in orientation. This expanded our training dataset by 16X to 70480 examples.
We validated the model using 5-fold cross validation and used the median lowest loss as a guide to the number of epochs for training the final model (Fig. 3a). Cross validation folds were split on the list of individual cell tracks, not 3-frame clips, to ensure QPI frames from the same cell were never used for both training and validation at the same time. The final model was trained for 45 epochs using stochastic gradient descent with Nesterov momentum of 0.9 and decay of 1e-6. We used a learning rate of 1e-5. Our loss function was mean absolute error (MAE). Training was performed using a single Nvidia GeForce GTX 1080 (8GB).
We validated the model using 5-fold cross validation and used the median lowest loss as a guide to the number of epochs for training the final model (Fig. 3a). Cross validation folds were split on the list of individual cell tracks, not 3-frame clips, to ensure QPI frames from the same cell were never used for both training and validation at the same time. The final model was trained for 45 epochs using stochastic gradient descent with Nesterov momentum of 0.9 and decay of 1e-6. We used a learning rate of 1e-5. Our loss function was mean absolute error (MAE). Training was performed using a single Nvidia GeForce GTX 1080 (8GB).
About PubCompare
Our mission is to provide scientists with the largest repository of trustworthy protocols and intelligent analytical tools, thereby offering them extensive information to design robust protocols aimed at minimizing the risk of failures.
We believe that the most crucial aspect is to grant scientists access to a wide range of reliable sources and new useful tools that surpass human capabilities.
However, we trust in allowing scientists to determine how to construct their own protocols based on this information, as they are the experts in their field.
Ready to get started?
Sign up for free.
Registration takes 20 seconds.
Available from any computer
No download required
Revolutionizing how scientists
search and build protocols!