Documentation Updates
VisionPro Deep Learning 3.1
-
Added "{}" to the example JSON file in the Cognex Label Import/Export Plug-in page.
VisionPro Deep Learning 3.0
-
The document structure and contents organization is reformed for the clear understanding, quick and easy access to the necessary contents, and future enhancement and rearrangement of the contents.
-
Fixed a typo "8192x819" to "8192x8192" in Known Issues - Tools - General section
-
Fixed contents into that the Expert Mode does not support Edit Mask and Edit Region in Red Analyze . For example, see Edit Mask (Optional)
-
Updates about Adaptation Modes
-
Green Classify High Detail Mode Parameter
-
Red Analyze High Detail Mode Parameter
-
Updates About Label Check
-
Updates on Parameter Search and its Auto mode.
-
Updates on VisionPro Deep Learning GUI refinement
-
Interpret Results (all tools)
-
Updates on Red Analyze High Detail Validation Loss Formula
-
Updates about Indicators (Training Set Indicator, Validation Set Indicator) on Image Display Area, Updates about Filters for Validation Set
-
Updates about View Browser List Mode for all tools
-
Define Region of Interest (ROI) and Mask to Construct View (all tools)
-
Collect Images (all tools)
-
Create an Image Sample Set (all tools)
-
Modify Training Set (all tools)
-
Review Markings (Red Analyze, Green Classify)
-
Create Label (Labeling) (Red Analyze, Green Classify)
-
-
Updates about TensorRT for Runtime API and Batch Processing for Runtime Green Classify
-
Updates about Cognex Label Import/Export Plugin
-
Updates about Image Processing Logic Change for Red Analyze High Detail
-
Updates about Minimum Image Size for Green Classify High Detail and Red Analyze High Detail.
-
Collect Images (Red Analyze High Detail)
-
Updates about Validation Loss Scale Change
-
Updates about Image Resize Improvement for Green Classify High Detail
-
Removing Red Analyze High Detail Quick Contents
-
Removing Stride parameter from Red Analyze High Detail Training parameters
-
Updates about Migrating SuaKIT Segmentation and Classification Projects into VisionPro Deep Learning
-
Updates on FAQ
-
Updates on VisionPro Deep Learning 3.0 Known Issues
-
Tools - Red Analyze
-
Tools - Green Classify
-
Parameter Search
-
GPU
-
-
Updates on VisionPro Deep Learning 2.1 Known Issues
-
Parameter Search
-
-
Updates on High Detail training logic regarding Patience Epochs and Minimum Epochs in Training parameters.
-
Updates on Best Epoch update logic in Loss Inspector
-
Updates on the compatible VisionPro version
-
Updates on the GPU driver versions for NVIDIA RTX / Quadro® / Data Center driver
-
Updates on Designer Plug-in support
VisionPro Deep Learning 2.1
-
Updates on Red Analyze Tool and Green Classify Tool for newly added High Detail Quick modes
-
Updates on Edit Deployment Target page that introduces the CIE1 deployment is no longer available
-
Updates on Parameter Search and Use the Parameter Search Utility
-
Updates on Image File and Format Requirements.
-
Updates on the contents explaining statistical results
-
Statistics
-
False Positive and False Negative Results
-
Blue Locate Tool – Statistics
-
Blue Read Tool – Statistics
-
Green Classify Tool – Confusion Matrix
-
Green Classify Tool – F Score
-
Red Analyze Tool – Statistics – Confusion Matrix
-
Red Analyze Tool – Statistics – Region Area Metrics
-
-
Updates on the model scores for Blue Locate/Read
-
Blue Locate Tool – Node Model
-
Blue Locate Tool – Layout Model
-
Blue Read Tool – Node Model
-
Blue Read Tool – Regex Model
-
Blue Read Tool – String Model
-
-
Updates on C#, C++ example codes (...\ProgramData\Cognex\VisionPro Deep Learning\2.1\Examples\). There were many improvements in example codes and some of the notable changes are:
-
The unit of angle for input is changed from "degree" to "radian" (example: Perturbation of Red Analyze)
-
For Green Classify, the codes that load label information per image is added.
-
The parameter path used for vidi_runtime_get_parameter() and vidi_runtime_set_parameter() is changed to be a "full path" which should include "tool/" within it.
Example: "workspaces/(workspace_name)/streams/(stream_name)/tools/(tool_name)/tool/processing/sampling_density"
Note: When either "roi/", "runtime_parameters/", or "post_process_parameters/" are used in the parameter path, this path should not include "tool/".
Example 1: "workspaces/(workspace_name)/streams/(stream_name)/tools/(tool_name)/roi/..."
Example 2: "workspaces/(workspace_name)/streams/(stream_name)/tools/(tool_name)/runtime_parameters/..."
-
-
Updates on page contents.
VisionPro Deep Learning 2.0
-
Updates for VisionPro Deep Learning 2.0 release
-
The documentation for the integration with Cognex Designer is no longer provided
-
VisionPro Deep Learning 1.1
-
Updates for VisionPro Deep Learning 1.1 release
-
Designer Plugin script examples for each tool have been updated for VisionPro Deep Learning 1.1.
-
Updates of the comments in C#, C++ example codes (...\ProgramData\Cognex\VisionPro Deep Learning\1.1\Examples\)
-
Updates about how to use multiple GPUs for VisionPro Deep Learning in Multiple GPUs Utilization.
-
Updates about the introduction to the various training parameters for VisionPro Deep Learning in Training Parameters
-
Updates about how to import SuaKIT projects in Import SuaKIT Project.
VisionPro Deep Learning 1.0.1
-
Updates for VisionPro Deep Learning 1.0.1 release
-
Updates for Migrating To This Release
The migration records from ViDi Suite 3.4 to ViDi Suite 4.1.0 have been reorganized for the correct migration history.