Model building method and device for estimating hardware configuration based on scene file parameters

Abstract

The invention discloses a model building method for estimating hardware configuration based on scene file parameters. The method comprises the steps of: building a neural network model which comprises at least two linear layers and activation function layers, one activation function layer being arranged between every two adjacent linear layers; determining features corresponding to a scene file; obtaining feature data based on scene file parameters, training the neural network model by using the feature data, inputting the feature data into the neural network model to obtain a feature array; enabling the feature array to correspond to a predicted category; calculating the loss values of predicted category and a real category, and optimizing and updating the parameters of the neural network model according to the loss values; and repeating the steps until the training of the neural network model is completed. The method and the device have the effect of improving the accuracy of hardware configuration estimation.

Description

Technical field [0001] The present application relates to the field of 3D scene rendering, in particular, involving scene file parameters to estimate hardware configuration methods, and devices. Background technique [0002] When rendering the three-dimensional model, the rendering is often necessary to rendering the 3D rendering software (such as: 3ds max), and rendering mainly depends mainly on the CPU's power, GPU's power and machine memory, and during the rendering process, the hardware is not replaced. Configuration, only interrupt rendering resumption. When performing rendering, when the above resources are insufficient, the scene rendering may be stuck inly, or directly crash, which is undebbly very inconvenient for users who have already put a lot of time and resources. [0003] For the expedition of hardware configuration, there is no comprehensive process or tool on the market. It is generally determined by the submitter based on experience. For example, high-level reso...

AI Technical Summary

Problems solved by technology

When performing rendering, when the above resources are insufficient, scene rendering may be stuck indefinitely, or crash directly, which is undoubtedly very inconvenient for users who have invested a lot of time and resources

[0003] For the estimation of hardware configuration, there is currently no complete process or tool on the market. It is generally determined by the submitter based on experience. For example, if the rendering result requires higher resolution and other high-precision requirements, choose a better one as much as possible. Hardware Configuration

Images