CNN algorithms like DenseNet DenseNet stress parameter efficiency, which usually results in less FLOPs. However, what I am struggling to understand is why this is important. For DenseNet, in particular, it has low inference speed. Isn't the purpose of decreased parameter size/FLOPs to decrease the time for inference? Is there another real world reason, such as perhaps less energy used, for these optimizations?
What is the purpose of decreased FLOPs and parameter size if they are not for increased speed?
921 views Asked by ddd At
1
There are 1 answers
Related Questions in MACHINE-LEARNING
- Trained ML model with the camera module is not giving predictions
- Keras similarity calculation. Enumerating distance between two tensors, which indicates as lists
- How to get content of BLOCK types LAYOUT_TITLE, LAYOUT_SECTION_HEADER and LAYOUT_xx in Textract
- How to predict input parameters from target parameter in a machine learning model?
- The training accuracy and the validation accuracy curves are almost parallel to each other. Is the model overfitting?
- ImportError: cannot import name 'HuggingFaceInferenceAPI' from 'llama_index.llms' (unknown location)
- Which library can replace causal_conv1d in machine learning programming?
- Fine-Tuning Large Language Model on PDFs containing Text and Images
- Sketch Guided Text to Image Generation
- My ICNN doesn't seem to work for any n_hidden
- Optuna Hyperband Algorithm Not Following Expected Model Training Scheme
- How can I resolve this error and work smoothly in deep learning?
- ModuleNotFoundError: No module named 'llama_index.node_parser'
- Difference between model.evaluate and metrics.accuracy_score
- Give Bert an input and ask him to predict. In this input, can Bert apply the first word prediction result to all subsequent predictions?
Related Questions in CONV-NEURAL-NETWORK
- The meaning of an out_channel in nn.Conv2d pytorch
- Difference between model.evaluate and metrics.accuracy_score
- PyTorch training on M2 GPU slower than Colab CPU
- LeNet5 & Self-Organizing Maps - RuntimeError: Trying to backward through the graph a second time - PyTorch
- How I resolve : InvalidArgumentError: Graph execution error?
- Ensemble Learning
- Implementing tflite quantized inference in python
- how to set a custom prior for Convolution2DReparameterization?
- this code is not working on tensorflow 2.16.0+ version
- How to cut out a fragment of an image in a contour
- Can I treat CNN channels separately to make placement predictions?
- Image channel error while training CNN model
- What is the size difference between Convolutional Neural Networks that take as input an RGB vs. monochrome image?
- Struggling to put my input images in the right format for CNN (tensorflow)
- Siamese Network Backpropagation
Related Questions in FLOPS
- Achieving More FMA3 Performance Than The Theoretical Maximum
- The Intel MKL LINPACK test indicates too big performance
- Which instructions are included in the hardware event `INST_SIMD_ALU`?
- I want to measure FLOPs! I want to measure the amount of computation of the call function set to @tf.funcion, what should I do
- find number of MAC operations in python
- How to compute the model complexity of FasterRCNNFPN pretrained from torchvision?
- How to calculate the FLOPS of a Python program?
- Counting FLOPS in tensorflow
- calculate flops in a custom pytorch model
- calculate program execution time in embedded device from python code
- why is Rpeak different from Rmax when measuring performance?
- Why do I get higher Whetstone FLOPS from SiSoft Sandra when I disable extensions (SSE, AVX, FMA)?
- Is it possible that the inference time is large while number of parameters and flops are low in pytorch?
- Difficulty understanding FLOPS in this scenario
- TensorFlow Object Detection API - determining FLOPS and number of Parameters
Related Questions in DENSENET
- How do I use DenseNet121 in MONAI?
- Streamlit importing HDF5 FILES
- How to add metadata to a tensorflow lite image recongition model Densenet121 with pretrained weights from imagenet?
- How to compute GradCAM for keras regression model with multiple outputs
- Comparison between DenseNet and GCN models shows abnormally high performance for GCN - Request for data leakage validation
- RuntimeError: mat1 and mat2 shapes cannot be multiplied (229376x7 and 1024x512)
- Input 0 of layer "dense" is incompatible with the layer: expected min_ndim=2, found ndim=1. Full shape received: (None,)
- OperatorNotAllowedInGraphError: Iterating over a symbolic `tf.Tensor` is not allowed: AutoGraph did convert this function
- ImportError: cannot import name 'Input' from 'keras.models'
- How to solve `NameError: name 'compression' is not defined`?
- How to unfreeze layers from a densenet? (PyTorch)
- how to change the out_features of densenet121 model?
- Is it possible to perform quantization on densenet169 and how?
- Multihead model based on DenseNet201 using Keras
- DenseNet 121 implementation using tensorflow giving AlreadyExistsError: Another metric with the same name already exists
Popular Questions
- How do I undo the most recent local commits in Git?
- How can I remove a specific item from an array in JavaScript?
- How do I delete a Git branch locally and remotely?
- Find all files containing a specific text (string) on Linux?
- How do I revert a Git repository to a previous commit?
- How do I create an HTML button that acts like a link?
- How do I check out a remote Git branch?
- How do I force "git pull" to overwrite local files?
- How do I list all files of a directory?
- How to check whether a string contains a substring in JavaScript?
- How do I redirect to another webpage?
- How can I iterate over rows in a Pandas DataFrame?
- How do I convert a String to an int in Java?
- Does Python have a string 'contains' substring method?
- How do I check if a string contains a specific word?
Trending Questions
- UIImageView Frame Doesn't Reflect Constraints
- Is it possible to use adb commands to click on a view by finding its ID?
- How to create a new web character symbol recognizable by html/javascript?
- Why isn't my CSS3 animation smooth in Google Chrome (but very smooth on other browsers)?
- Heap Gives Page Fault
- Connect ffmpeg to Visual Studio 2008
- Both Object- and ValueAnimator jumps when Duration is set above API LvL 24
- How to avoid default initialization of objects in std::vector?
- second argument of the command line arguments in a format other than char** argv or char* argv[]
- How to improve efficiency of algorithm which generates next lexicographic permutation?
- Navigating to the another actvity app getting crash in android
- How to read the particular message format in android and store in sqlite database?
- Resetting inventory status after order is cancelled
- Efficiently compute powers of X in SSE/AVX
- Insert into an external database using ajax and php : POST 500 (Internal Server Error)
There is a difference between overall inference time vs. per parameter/FLOPs training efficiency. Having lower parameter/FLOPs in training does not guarantee higher speed in inference. Because overall inference depends on the architecture and how predictions are computed.