xby-recog: A Recognition Toolkit
xby-recog is an image recognition plugin for DeepSeek Harness, maintained by xby-skill. It provides APIs for general OCR, handwritten text recognition, license plate recognition, and structured information extraction from ID cards, passports, driver's licenses, business licenses, bank cards, and more. After installation, it can only be used after configuring the API key via `set_xby_apikey`. The plugin supports three input modes: image URLs, BASE64-encoded data, or file paths. All code is licensed under the MIT License.
Read MoreDeepSeek Harness Plugin: xby-image-detect
`xby-image-detect` is a plugin for DeepSeek Harness (DSH) that wraps the YOLOv6 object detection model to provide image recognition capabilities to agents. It provides `set_xby_apikey` for API key configuration and supports three input types: image URL, Base64-encoded data, or a local file path. It can detect 80 COCO object categories in images and output bounding boxes, confidence scores, and class labels. The plugin is installed via the command line. Before use, you must register on the official XBY website and obtain an API key. It is suitable for agent scenarios that require integrated image analysis functionality.
Read Morexby-general-recognition: Introducing the General Recognition Plugin
`xby-general-recognition` is a model inference plugin for DeepSeek Harness (DSH) that enables direct image recognition service calls within conversations. This plugin supports three input formats: image URLs, BASE64-encoded data, and file paths, and can identify the main objects in images by outputting class labels (covering over 50,000 categories). Users must first register on the official website to obtain an API key, which can then be activated using `set_xby_apikey`. The plugin is maintained by `xby-skill`, is released under the MIT license, and runs with the privileges of the current process.
Read Morexby-ebike-detection: A Plugin for E-Bike Detection in Images
`xby-ebike-detection` is a model inference plugin for DeepSeek Harness (DSH) used to detect electric bikes in images. It offers command interfaces such as `set_xby_apikey` and `ebike_detection` and supports detection with input provided as an image URL, BASE64-encoded data, or a file path. Users must register on the official website to obtain an API key before installation. The plugin is released under the MIT License and is designed to provide DSH users with convenient detection capabilities.
Read Moredsh-eyes: Attaching external "eyes" to DeepSeek Harness pure text chat
# Preface The philosophy behind DeepSeek Harness (hereinafter referred to as DSH) is "everything is a plugin," but its web chat interface is text-based: pasting a screenshot of an error or dragging in a design draft means the model cannot see them; you can only describe them in text or use other tools to perform OCR before copying the results back into the conversation. For agent developers, such operations are frequent and interrupt the workflow each time. qing9835's open-source plugin, dsh-eyes, addresses this problem: images pasted, dragged in, or imported are intercepted by the plugin, passed to OpenAI-compatible vision models for recognition as text, and automatically sent into the conversation. # What is dsh-eyes (package name dsh-vision-bridge) is a DSH web...
Read Moredsh-vision-adapter: Adding drag-and-drop image recognition capability to DeepSeek Harness
dsh-vision-adapter is a DSH-exclusive plugin maintained by motongv (MIT License) that solves the issue where DSH text models cannot recognize images. The chat interface remains unchanged; images are first converted into text descriptions by the vision model, which is then submitted to DeepSeek for reasoning and response. The plugin acts as an in-process HTTP bridge within DSH to register vision.
Read Moredsh-deepseek-vision: Adding image recognition to DeepSeek text models
dsh-deepseek-vision is an MIT-licensed DSH plugin designed to provide image recognition capabilities for text-only DeepSeek models. By registering the `deepseek_vision` tool, the plugin allows the model to submit local image paths (supporting PNG, JPG, WebP, and GIF, but not HEIC) and optional prompts. When invoked, the plugin utilizes the DSH subprocess service to safely call the local `deepseek-vision-cli` via array parameters, sends the image to the DeepSeek Web endpoint for recognition, and returns the descriptive text to the model. Users need to establish a Web session via manual login to the assistant or Token login; the Token is stored only in local cache. Installation supports local directories or Releases.
Read Moredsh-vision-imagen: DeepSeek Harness Visual Recognition and Image Generation Plugin
dsh-vision-imagen is a DeepSeek Harness (DSH) plugin maintained by xuxun-oss, designed to bridge Google Gemini's multimodal capabilities and supports certain OpenAI-compatible backends. It allows users to perform image recognition, text-to-image generation, and image editing directly within standard DeepSeek sessions without the need to manually switch models. Core features include using `vision_read` for OCR and image analysis, `generate_image` for text-to-image generation with iterative self-correction, and `edit_image` for image optimization. The plugin supports Gemini and OpenAI-compatible backends (such as GPT-4o, Qwen-VL, etc.) and features an automatic fallback mechanism. Installation requires Node
Read Moredsh-mimo-vision-hint: Zero-dependency hint for DSH to delegate image recognition to opencode-go mimo-v2.5 sub-agent
dsh-mimo-vision-hint is a zero-dependency DSH profile bundle plugin maintained by Isekai-Mfu (MIT licensed). It does not register model tools or provide a cordis service; instead, it guides the agent to delegate image recognition tasks to the opencode-go mimo-v2.5 sub-agent by injecting system prompts. If a workflow invocation fails, it automatically falls back to read_image. This plugin is intended for DSH deployment environments configured with the opencode-go provider, the mimo-v2.5 model, and the workflow tool. It can be installed using the command `dsh plugin --profile web add github:Ise...`
Read Morexiaoxianyu-office/dsh-image-tools: Add image recognition capability to DeepSeek Harness text sessions
`xiaoxianyu-office/dsh-image-tools` is a DSH plugin designed to equip text-only models like `deepseek-v4-pro` with image recognition capabilities. The plugin works by automatically saving chat images to the workspace, dynamically disabling the native `read_image` tool, and guiding the model to use the `image_recognize` tool to delegate visual sub-agents (such as `xiaomi/mimo-v2.5`) for image recognition. It supports various preset sessions like `standard` and `code`, dynamically determines interception or bridging strategies based on current model capabilities, and includes a first-time setup wizard and fault self-diagnosis functions. Image recognition output follows strict specifications, requiring pixel coordinates and `#RRGGBB` color values, and addresses items not appearing in the image or...
Read MoreDSH-Vision: A Plugin Adding Image Recognition Capabilities to DeepSeek Harness
dsh-vision is a DeepSeek Harness (DSH) bundle plugin maintained by Terry12138qy, licensed under the MIT License. This plugin is designed to address the issue that the DSH main model lacks native image recognition capabilities. By encapsulating the image recognition workflow, it sends local image paths or network URLs to Alibaba Cloud Bailian or Zhipu multimodal models, ultimately returning Chinese text descriptions for use by agents or Code Mode. Core features include automatically reading and encoding images, prioritizing the main model `qwen3.5-omni-plus`, and automatically switching to the fallback model `glm-4.6v-flash` for retries upon failures (e.g., quota exhaustion, rate limiting). It supports invocation via `await tools.vision(...)`.
Read MoreBuilding an Animal Recognition System with PaddlePaddle to Identify Thousands of Animal Species
This paper introduces a project for animal recognition using PaddlePaddle. Firstly, the animal recognition task can be completed with just a few lines of code. Secondly, a GUI interface is provided to facilitate users in uploading images for recognition. Finally, a Flask web interface is supported for Android calls, enabling cross - platform application. The project includes details such as model path, image reading, and prediction result output, and running screenshots are attached to demonstrate the implementation effect.
Read MoreEnd-to-End Recognition of Captchas Based on PaddlePaddle 2.0
Your code has covered most aspects of the CAPTCHA recognition project, including data processing, model training, and inference. Below are some suggestions for improvements and enhancements to your provided code: ### 1. Data Preprocessing Ensure the image dimensions are consistent (27x72), as this is the input size used during training. ### 2. Model Definition Your `Model` class has already encapsulated the network structure well. You can further optimize it and add more comments to facilitate understanding. ### 3. Training Process During the training process, ensure that when using multi-GPU training,
Read MoreImplementing Image Classification on Android Phones Based on TNN
This project is mainly an image classifier based on TensorFlow Lite, which can achieve real-time image recognition on Android devices. Its main functions and implementation steps are as follows: ### Project Structure - **MainActivity.java**: Implements gallery image selection and real-time camera prediction on the main interface. - **MNNClassification.java**: Integrates and encapsulates MNN model-related operations. ### Implementation Ideas 1. **Initialization**:
Read MoreImage Classification on Android Phones Based on MNN
This is a detailed guide on how to implement image classification in an Android application. You have successfully used TensorFlow Lite for image classification and demonstrated how to obtain input data through two methods: calling the camera and selecting images, and then passing this data to the model for prediction. ### Summary of Main Content 1. **Model Initialization**: First, load the pre-trained `mobilenet_v2_1.0_224.tflite` model and create a classifier instance. 2. **Reading Images and Pro
Read MoreImplementing Image Classification on Android Phones Based on Paddle Lite
Thank you for sharing this Android application development example for image classification based on Paddle Lite. Your project not only covers how to obtain categories from images but also introduces methods for real-time image recognition through the camera, enabling users to quickly understand information about the captured object in practical application scenarios. Below, I will further optimize and supplement the content you provided and offer some suggestions to improve the user experience or enhance code efficiency: ### 1. Project Structure and Resource Management Ensure the project has a clear file structure (e.g., `assets/image
Read MoreImplementation of Image Classification on Android Phones Based on TensorFlow Lite
This project mainly implements an image classification application based on TensorFlow Lite, which can perform object recognition using images from the camera or photo album on an Android device and provide real-time prediction functionality. The following is a detailed analysis of the core steps and key code of this project: ### Project Structure - **TFLiteModel**: Contains model-related configurations. - **MainActivity**: The main interface for launching the camera or selecting images for classification. - **RunClassifier** (Note: The original text seems to be incomplete here, so the translation preserves the placeholder as is.)
Read MoreCrowdNet: A Density Estimation Model Implemented with PaddlePaddle
That's the detailed tutorial on crowd flow density prediction. Through this project, you can learn how to use PaddlePaddle to solve practical problems, with detailed step-by-step guidance from training to prediction. If you encounter any issues or have any questions during the process, please feel free to ask in the comments section! We will also continuously pay attention to feedback to assist more friends who want to enter the AI field. We hope this case can help everyone better understand the process of data processing and model training.
Read MoreMy New Book Has Been Published!
This book "Deep Learning in Practice with PaddlePaddle" shares the author's experience from getting acquainted with PaddlePaddle to completing the book publication. It introduces the PaddlePaddle framework in detail and helps readers master practical applications through cases such as handwritten digit recognition. The content covers basic usage, dataset processing, object detection, as well as server-side and mobile-side applications. This book is suitable for machine learning enthusiasts and practitioners, and can also be used as a teaching reference. During the learning process of PaddlePaddle, the author shared tutorials through blogs, which ultimately led to the publication of this book.
Read MorePaddlePaddle From Beginner to "Alchemy" - Part 15: Deploying Prediction Models to Android Phones
Thank you for your sharing and detailed notes, which provide a great reference for developers who want to learn how to integrate PaddlePaddle for image recognition in Android applications. Below, I will summarize the information you provided and add some content that may help with understanding: ### 1. Environment Preparation - **Development Environment**: Ensure the latest version of Android Studio is installed. - **Permission Configuration**: Add necessary permissions in `AndroidManifest.xml`, such as read and write access to external storage.
Read More"PaddlePaddle From Beginner to Alchemist" - Part 11: Custom Image Dataset Recognition
This note mainly introduces how to use PaddlePaddle for training and prediction in image classification tasks, which specifically includes the following parts: ### 1. Dataset Preparation The author extracted 240 images from a dataset containing 6 categories of fruit images as the training set and organized them into CSV file format. ### 2. Model Construction A simple LeNet model structure was defined using PaddlePaddle. The model consists of two convolutional layers, two pooling layers, a fully connected layer, and finally performs classification through Softmax.
Read More"PaddlePaddle from Beginner to 'Alchemy' (Refined Version)" Part 4 - Convolutional Neural Networks
This tutorial provides a detailed introduction to training and predicting a handwritten digit recognition model using the PaddlePaddle framework. Below is a summary and further explanation of the key steps: ### 1. Preparing the Dataset First, the MNIST dataset is obtained from PaddlePaddle using the `fetch MNIST data` command. It is a widely used dataset for training machine learning models. ```python import paddle.v2 as paddle from paddle.v2.da ``` （Note: The code snippet appears truncated in the original input. The translation assumes the standard MNIST loading syntax in PaddlePaddle v2, though the full code may require additional imports or dataset initialization steps not visible in the provided snippet.)
Read MoreNotes on "My PaddlePaddle Learning Journey" ⑫ — Using the Visualization Tool VisualDL
This note provides a detailed introduction to how to use PaddlePaddle and VisualDL for the visualization of convolutional neural network (CNN) training. The following are the key points summarizing the content of the note: ### Visualizing CNN Training and Training Process with PaddlePaddle and VisualDL #### 1. Preparation - **Environment Installation**: Ensure that Python, PaddlePaddle, and VisualDL are installed. - **Dependency Library Import**: ```python
Read MoreNotes on "My PaddlePaddle Learning Journey" VII——End-to-End License Plate Recognition
This project mainly introduces how to use the PaddlePaddle framework to train a license plate recognition model. Below, I will summarize the key steps and concepts and provide some optimization suggestions. ### Summary of Key Steps 1. **Data Preparation**: - Collect and preprocess license plate images. - Create a label dictionary to map characters to indices. 2. **Model Construction**: - Use the PaddlePaddle framework to create an end-to-end recognition model. - The model includes an input layer, convolutional layers,
Read MoreNotes on "My Learning Journey with PaddlePaddle" – End-to-End Recognition of Verification Codes
This article provides a detailed introduction to the process of license plate recognition using PaddlePaddle, covering installation of the environment, reading the dataset, building the model, as well as training and testing. Below are summaries of several key points from the article: ### 1. Environment Setup The author first created a virtual environment for PaddlePaddle and configured the CUDA/CUDNN version. ### 2. Dataset Preparation A dataset containing a large number of license plate images was used. These data are publicly available on GitHub, and each license plate has a corresponding label. The author parsed the files
Read MoreNotes on My PaddlePaddle Learning Journey V——Captcha Recognition
Your tutorial provides a detailed introduction to using PaddlePaddle for captcha recognition, covering steps from dataset preparation, model design to final training and prediction. This series of steps is highly suitable for understanding and learning the basic processes and techniques of deep learning, especially its applications in the field of OCR (Optical Character Recognition). ### Code Structure Analysis 1. **Data Preprocessing**: - The `read_file` function is used to read image files and convert them into a format suitable for model input. - `load_and_tr
Read MoreNotes on "My PaddlePaddle Learning Journey" IV — Recognition of Custom Image Datasets
This series of notes mainly introduces how to implement a simple image recognition task using PaddlePaddle, including data preparation, model construction and training, as well as result prediction. The following is a summary of the main content of each part: ### 1. Environment Setup and Initial Configuration - **Environment Configuration**: First, install Python 3 and ensure it runs properly. - **Download Preprocessing Script**: Use the `DownloadImages.py` script to batch-download images to be recognized from Baidu Images. This script can perform downloads based on keywords (the original text is truncated here).
Read MoreNotes on "My Learning Journey with PaddlePaddle" III — CIFAR Color Image Recognition
This project is a CIFAR-10 image classification model implemented using PaddlePaddle, with a clear code structure and detailed comments. Below is a brief explanation of the main functions and principles of each part: ### 1. `vgg.py` This is a file containing the definition of the VGG network. VGG is a classic convolutional neural network architecture, and here it is implemented in PaddlePaddle. #### Main Content: - **Defines the VGG network structure**: Including multiple convolutional layers, pooling layers, and fully
Read More