{"schema_version":"onlylabs.public_analysis_evidence.v1","title":"Baidu (ERNIE) analysis evidence pack","description":"Public onlylabs evidence pack for cited agent analysis: captured pages, ranked public signals, and stored web-search provenance used by the background analysis workflow.","url":"https://onlylabs.fyi/analysis/baidu","json_url":"https://onlylabs.fyi/analysis/baidu/evidence.json","generated_at":"2026-06-11T18:06:27.454Z","org":{"slug":"baidu","name":"Baidu (ERNIE)","category":"frontier-lab","category_label":"Frontier lab","dossier_url":"https://onlylabs.fyi/labs/baidu"},"analysis":{"url":"https://onlylabs.fyi/analysis/baidu","json_url":"https://onlylabs.fyi/analysis/baidu/analysis.json","generated_at":"2026-06-08T15:59:08.147+00:00"},"workflow":{"version":"synthesize-analyses","provider":null,"model":null,"agent":null,"public_pack_mode":"local-pages-and-events","live_web_fetches":false,"note":"Public evidence exports do not trigger live Exa calls; stored Exa provenance is included when analysis metadata contains it."},"stats":{"pages":28,"events":140,"web":0,"evidence":88,"signal_desks":{"hiring":19,"forks":1,"releases":27,"talking":12,"repos":1},"data_radar_lanes":{"data":0,"evals":2,"infrastructure":2,"safety":0,"product":3},"data_radar_matches":7,"stored_analysis_evidence":null,"stored_analysis_web":null,"stored_analysis_signal_desks":null,"stored_analysis_data_radar_lanes":null,"stored_analysis_data_radar_matches":null},"stored_web_provenance":null,"evidence":[{"ref":"P1","kind":"page","title":"PaddlePaddle/Paddle-Lite repository metadata","date":"2026-06-11T03:58:02.398344+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/Paddle-Lite","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/Paddle-Lite\n\nDescription: PaddlePaddle High Performance Deep Learning Inference Engine for Mobile and Edge (飞桨高性能深度学习端侧推理引擎）\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 7259\n\nForks: 1627\n\nOpen issues: 58\n\nCreated: 2017-09-20T11:41:42Z\n\nPushed: 2026-04-27T09:16:13Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n# Paddle Lite\n\n[English](README_en.md) | 简体中文\n\n[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](https://www.paddlepaddle.org.cn/lite) [![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle-Lite.svg)](https://github.com/PaddlePaddle/Paddle-Lite/releases) [![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE) \n\nPaddle Lite 是一个高性能、轻量级、灵活性强且易于扩展的深度学习推理框架，定位于支持包括移动端、嵌入式以及边缘端在内的多种硬件平台。\n\n当前 Paddle Lite 不仅在百度内部业务中得到全面应用，也成功支持了众多外部用户和企业的生产任务。\n\n## 快速入门\n\n使用 Paddle Lite，只需几个简单的步骤，就可以把模型部署到多种终端设备中，运行高性能的推理任务，使用流程如下所示：\n\n**一. 准备模型**\n\nPaddle Lite 框架直接支持模型结构为 [PaddlePaddle](https://github.com/PaddlePaddle/Paddle) 深度学习框架产出的模型格式。目前 PaddlePaddle 用于推理的模型是通过 [save_inference_model](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/static/save_inference_model_cn.html#save-inference-model) 这个 API 保存下来的。\n如果您手中的模型是由诸如 Caffe、Tensorflow、PyTorch 等框架产出的，那么您可以使用 [X2Paddle](https://github.com/PaddlePaddle/X2Paddle) 工具将模型转换为 PaddlePaddle 格式。\n\n**二. 模型优化**\n\nPaddle Lite 框架拥有优秀的加速、优化策略及实现，包含量化、子图融合、Kernel 优选等优化手段。优化后的模型更轻量级，耗费资源更少，并且执行速度也更快。\n这些优化通过 Paddle Lite 提供的 opt 工具实现。opt 工具还可以统计并打印出模型中的算子信息，并判断不同硬件平台下 Paddle Lite 的支持情况。您获取 PaddlePaddle 格式的模型之后，一般需要通过该 opt 工具做模型优化。opt 工具的下载和使用，请参考[模型优化方法](https://www.paddlepaddle.org.cn/lite/develop/user_guides/model_optimize_tool.html)。\n\n**三. 下载或编译**\n\nPaddle Lite 提供了 Android/iOS/x86/macOS 平台的官方 Release 预测库下载，我们优先推荐您直接下载 [Paddle Lite 预编译库](https://www.paddlepaddle.org.cn/lite/develop/quick_start/release_lib.html)，或者从 Release notes 处获取最新的[预编译编译库](https://github.com/PaddlePaddle/Paddle-Lite/releases)。\n\nPaddle Lite 已支持多种环境下的源码编译，为了避免复杂、繁琐的环境搭建过程，我们建议您使用 [Docker 统一编译环境搭建](https://www.paddlepaddle.org.cn/lite/develop/source_compile/docker_env.html) 进行编译。当然，您也可以根据宿主机和目标设备的 CPU 架构和操作系统，在[源码编译](https://www.paddlepaddle.org.cn/lite/develop/source"},{"ref":"P2","kind":"page","title":"PaddlePaddle/Anakin repository metadata","date":"2026-06-11T03:58:01.436902+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/Anakin","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/Anakin\n\nDescription: High performance Cross-platform Inference-engine, you could run Anakin on x86-cpu,arm, nv-gpu, amd-gpu,bitmain and cambricon devices.\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 536\n\nForks: 135\n\nOpen issues: 68\n\nCreated: 2018-05-18T05:32:25Z\n\nPushed: 2022-09-23T22:22:29Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Anakin2.0\n\n[![Build Status](https://travis-ci.org/PaddlePaddle/Anakin.svg?branch=developing)](https://travis-ci.org/PaddlePaddle/Anakin)\n[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)\n[![Coverage Status](https://coveralls.io/repos/github/xklnono/Anakin/badge.svg)](https://coveralls.io/github/xklnono/Anakin)\n\nWelcome to the Anakin GitHub.\n\nAnakin is a cross-platform, high-performance inference engine, which is originally\ndeveloped by Baidu engineers and is a large-scale application of industrial products.\n\nPlease refer to our [release announcement](https://github.com/PaddlePaddle/Anakin/releases) to track the latest feature of Anakin.\n\n## Features\n\n- **Flexibility**\n\nAnakin is a cross-platform, high-performance inference engine, supports a wide range of neural network architectures and different hardware platforms. It is easy to run Anakin on GPU / x86 / ARM platform.\n\nAnakin has integrated with NVIDIA TensorRT and open source this part of integrated API to provide services, developers can call the API directly or modify it as needed, which will be more flexible for development requirements.\n\n- **High performance**\n\nIn order to give full play to the performance of hardware, we optimized the\nforward prediction at different levels.\n- Automatic graph fusion. The goal of all performance optimizations under a\ngiven algorithm is to make the ALU as busy as possible. Operator fusion\ncan effectively reduce memory access and keep the ALU busy.\n\n- Memory reuse. Forward prediction is a one-way calculation. We reuse\nthe memory between the input and output of different operators, thus\nreducing the overall memory overhead.\n\n- Assembly level optimization. Saber is a underlying DNN library for Anakin, which\nis deeply optimized at assembly level.\n\n## NV GPU Benchmark\n### Machine And Envi"},{"ref":"P3","kind":"page","title":"PaddlePaddle/VisualDL repository metadata","date":"2026-06-11T03:58:00.952223+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/VisualDL","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/VisualDL\n\nDescription: Deep Learning Visualization Toolkit（『飞桨』深度学习可视化工具 ）\n\nLanguage: HTML\n\nLicense: Apache-2.0\n\nStars: 4883\n\nForks: 634\n\nOpen issues: 157\n\nCreated: 2017-12-20T12:34:31Z\n\nPushed: 2025-01-22T06:02:52Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n[**中文**](./README_CN.md)\n\n<p align=\"center\">\n<img src=\"https://raw.githubusercontent.com/PaddlePaddle/VisualDL/develop/frontend/packages/core/public/images/logo-visualdl.svg?sanitize=true\" width=\"70%\"/>\n</p>\n\n<p align=\"center\">\n<a href=\"https://actions-badge.atrox.dev/PaddlePaddle/VisualDL/goto?ref=develop\"><img alt=\"Build Status\" src=\"https://img.shields.io/endpoint.svg?url=https%3A%2F%2Factions-badge.atrox.dev%2FPaddlePaddle%2FVisualDL%2Fbadge%3Fref%3Ddevelop&style=flat-square\" alt=\"Build Status\" /></a>\n<a href=\"https://pypi.org/project/visualdl/\"><img src=\"https://img.shields.io/pypi/v/visualdl?style=flat-square\" alt=\"PyPI\" /></a>\n<a href=\"https://pypi.org/project/visualdl/#files\"><img src=\"https://img.shields.io/pypi/dm/visualdl?style=flat-square\" alt=\"Downloads\" /></a>\n<a href=\"https://github.com/PaddlePaddle/VisualDL/blob/develop/LICENSE\"><img src=\"https://img.shields.io/github/license/paddlepaddle/visualdl?style=flat-square\" alt=\"License\" /></a>\n</p>\n\n<p align=\"center\">\n<a href=\"javascript:void(0)\"><img src=\"https://img.shields.io/badge/QQ_Group-1045783368-52B6EF?style=social&logo=tencent-qq&logoColor=000&logoWidth=20\" alt=\"QQ Group\" /></a>\n</p>\n\n## Introduction\n\nVisualDL, a visualization analysis tool of PaddlePaddle, provides a variety of charts to show the trends of parameters, and visualizes model structures, data samples, histograms of tensors, PR curves , ROC curves and high-dimensional data distributions. It enables users to understand the training process and the model structure more clearly and intuitively so as to optimize models efficiently.\n\nVisualDL provides various visualization functions, including **tracking metrics in real-time, visualizing the model structure, displaying the data sample, visualizing the relationship between hyperparameters and model metrics, presenting the changes of distributions of tensors, showing the pr curves, projecting high-dimens"},{"ref":"P4","kind":"page","title":"PaddlePaddle/PaddleSpeech repository metadata","date":"2026-06-11T03:58:00.894021+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleSpeech","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleSpeech\n\nDescription: Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation and Keyword Spotting. Won NAACL2022 Best Demo Award.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 12614\n\nForks: 1957\n\nOpen issues: 271\n\nCreated: 2017-11-14T12:36:30Z\n\nPushed: 2026-06-10T06:42:49Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n([简体中文](./README_cn.md)|English)\n<p align=\"center\">\n<img src=\"./docs/images/PaddleSpeech_logo.png\" />\n</p>\n\n<p align=\"center\">\n<a href=\"./LICENSE\"><img src=\"https://img.shields.io/badge/license-Apache%202-red.svg\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleSpeech/releases\"><img src=\"https://img.shields.io/github/v/release/PaddlePaddle/PaddleSpeech?color=ffa\"></a>\n<a href=\"support os\"><img src=\"https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-pink.svg\"></a>\n<a href=\"\"><img src=\"https://img.shields.io/badge/python-3.8+-aff.svg\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleSpeech/graphs/contributors\"><img src=\"https://img.shields.io/github/contributors/PaddlePaddle/PaddleSpeech?color=9ea\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleSpeech/commits\"><img src=\"https://img.shields.io/github/commit-activity/m/PaddlePaddle/PaddleSpeech?color=3af\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleSpeech/issues\"><img src=\"https://img.shields.io/github/issues/PaddlePaddle/PaddleSpeech?color=9cc\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleSpeech/stargazers\"><img src=\"https://img.shields.io/github/stars/PaddlePaddle/PaddleSpeech?color=ccf\"></a>\n<a href=\"=https://pypi.org/project/paddlespeech/\"><img src=\"https://img.shields.io/pypi/dm/PaddleSpeech\"></a>\n<a href=\"=https://pypi.org/project/paddlespeech/\"><img src=\"https://static.pepy.tech/badge/paddlespeech\"></a>\n<a href=\"https://huggingface.co/spaces\"><img src=\"https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue\"></a>\n</p>\n<div align=\"center\"> \n<h4>\n<a href=\"#quick-start\"> Quick Start </a>\n| <a href=\"#documents\"> Documents </a>\n| <a href=\"#model-list\"> Models List </a>\n| <a href=\""},{"ref":"P5","kind":"page","title":"PaddlePaddle/continuous_evaluation repository metadata","date":"2026-06-11T03:58:00.389812+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/continuous_evaluation","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/continuous_evaluation\n\nDescription: Macro Continuous Evaluation Platform for Paddle.\n\nLanguage: Python\n\nStars: 19\n\nForks: 14\n\nOpen issues: 24\n\nCreated: 2018-03-14T12:04:51Z\n\nPushed: 2020-03-11T01:32:24Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Paddle Continous Evaluate Framework"},{"ref":"P6","kind":"page","title":"PaddlePaddle/Paddle2ONNX repository metadata","date":"2026-06-11T03:58:00.370929+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/Paddle2ONNX","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/Paddle2ONNX\n\nDescription: ONNX Model Exporter for PaddlePaddle\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 927\n\nForks: 196\n\nOpen issues: 16\n\nCreated: 2018-03-27T23:46:14Z\n\nPushed: 2026-03-18T09:31:20Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n# Paddle2ONNX\n\n简体中文 | [English](README_en.md)\n\n## 1 Paddle2ONNX 简介\n\nPaddle2ONNX 支持将 **PaddlePaddle** 模型格式转化到 **ONNX** 模型格式。通过 ONNX 可以完成将 Paddle 模型到多种推理引擎的部署，包括 TensorRT/OpenVINO/MNN/TNN/NCNN，以及其它对 ONNX 开源格式进行支持的推理引擎或硬件。\n\n## 2 Paddle2ONNX 环境依赖\n\nPaddle2ONNX 依赖PaddlePaddle3.0，我们建议您在以下环境下使用 Paddle2ONNX ：\n\n- PaddlePaddle == 3.0.0\n- onnxruntime >= 1.10.0\n\n## 3 安装 Paddle2ONNX\n\n如果您只是想要安装 Paddle2ONNX 且没有二次开发的需求，你可以通过执行以下代码来快速安装 Paddle2ONNX\n\n```\npip install paddle2onnx\n```\n\n如果你希望对 Paddle2ONNX 进行二次开发，请按照[Github 源码安装方式](docs/zh/compile_local.md)编译Paddle2ONNX。\n\n## 4 快速使用教程\n\n### 4.1 获取PaddlePaddle部署模型\n\nPaddle2ONNX 在导出模型时，需要传入部署模型格式，包括两个文件\n\n- `model_name.json`: 表示模型结构\n- `model_name.pdiparams`: 表示模型参数\n\n### 4.2 调整Paddle模型\n\n如果对Paddle模型的输入输出需要做调整，可以前往[Paddle 相关工具](./tools/paddle/README.md)查看教程。\n\n### 4.3 使用命令行转换 PaddlePaddle 模型\n\n你可以通过使用命令行并通过以下命令将Paddle模型转换为ONNX模型\n\n```bash\npaddle2onnx --model_dir model_dir \\\n--model_filename model.json \\\n--params_filename model.pdiparams \\\n--save_file model.onnx\n```\n\n可调整的转换参数如下表:\n\n| 参数 | 参数说明 |\n|----------------------------|-----------------------------------------------------------------------------------------------------------------|\n| --model_dir | 配置包含 Paddle 模型的目录路径 |\n| --model_filename | **[可选]** 配置位于 `--model_dir` 下存储网络结构的文件名 |\n| --params_filename | **[可选]** 配置位于 `--model_dir` 下存储模型参数的文件名 |\n| --save_file | 指定转换后的模型保存目录路径 |\n| --opset_version | **[可选]** 配置转换为ONNX的OpSet版本，目前支持7~19等多个版本，默认为 9 |\n| --enable_auto_update_opset | **[可选]** 是否开启opset version自动升级功能，当低版本opset无法转换时，自动选择更高版本的opset进行转换， 默认为 True |\n| --enable_onnx_checker | **[可选]** 配置是否检查导出为 ONNX 模型的正确性, 建议打开此开关， 默认为 True |\n| --enable_dist_prim_all | **[可选]** 是否开启组合算子拆解，默为 False |\n| --optimize_tool | **[可选]** ONNX模型优化工具，可选择onnxoptimizer、polygraphy、None, 默认为 onnxoptimizer |\n| --enable_verbose | **[可选]** 是否打印更更详细的日志信息，默认为 False |\n| --version | **[可选]** 查看 paddle2onnx 版本 |\n\n### 4.4 裁剪ONNX\n\n如果你需要调整 ONNX 模型，请参考 [ONNX 相关工具](./"},{"ref":"P7","kind":"page","title":"PaddlePaddle/PARL repository metadata","date":"2026-06-11T03:58:00.151558+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PARL","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PARL\n\nDescription: A high-performance distributed training framework for Reinforcement Learning \n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 3457\n\nForks: 815\n\nOpen issues: 134\n\nCreated: 2018-04-25T17:54:22Z\n\nPushed: 2025-09-13T06:29:18Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\".github/PARL-logo.png\" alt=\"PARL\" width=\"500\"/>\n</p>\n\nEnglish | [简体中文](./README.cn.md)\n\n[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](https://parl.readthedocs.io/en/latest/index.html) [![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](https://parl.readthedocs.io/zh_CN/latest/) [![Documentation Status](https://img.shields.io/badge/手册-中文-brightgreen.svg)](./docs/zh_CN/Overview.md) [![Release](https://img.shields.io/badge/release-v2.2.1-blue.svg)](https://github.com/PaddlePaddle/PARL/releases)\n\n> PARL is a flexible and high-efficient reinforcement learning framework.\n\n<!-- toc -->\n\n- [About PARL](#about-parl)\n- [Features](#features)\n- [Abstractions](#abstractions)\n- [Model](#model)\n- [Algorithm](#algorithm)\n- [Agent](#agent)\n- [Parallelization](#parallelization)\n- [Install:](#install)\n- [Dependencies](#dependencies)\n- [Getting Started](#getting-started)\n- [Examples](#examples)\n- [Waymax-RL(2025 Update, GPU-RL Autonomous Driving)](#waymax-rl2025-update-gpu-rl-autonomous-driving)\n- [xparl Security](#xparl-security)\n- [Security Considerations](#security-considerations)\n\n# About PARL\n## Features\n**Reproducible**. We provide algorithms that stably reproduce the result of many influential reinforcement learning algorithms.\n\n**Large Scale**. Ability to support high-performance parallelization of training with thousands of CPUs and multi-GPUs.\n\n**Reusable**. Algorithms provided in the repository could be directly adapted to a new task by defining a forward network and training mechanism will be built automatically.\n\n**Extensible**. Build new algorithms quickly by inheriting the abstract class in the framework.\n\n## Abstractions\n<img src=\".github/abstractions.png\" alt=\"abstractions\" width=\"400\"/>\nPARL aims to build an agent for training algorithms to perform comp"},{"ref":"P8","kind":"page","title":"PaddlePaddle/paddle-ce-latest-kpis repository metadata","date":"2026-06-11T03:58:00.147488+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/paddle-ce-latest-kpis","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/paddle-ce-latest-kpis\n\nDescription: Paddle Continuous Evaluation, keep updating.\n\nLanguage: Python\n\nStars: 25\n\nForks: 39\n\nOpen issues: 7\n\nCreated: 2018-04-02T02:40:42Z\n\nPushed: 2021-12-13T11:27:18Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Paddle Continuous Evaluation Baselines\n\n## Howtos\n\n### Add New Evaluation Task\n\nReference [mnist task](https://github.com/PaddlePaddle/paddle-ce-latest-kpis/tree/develop/models/mnist), \nthe following files are required by CE framework:\n\n- `run.xsh` , a script to start this evaluation execution\n- this script can be any bash script, just place `#!/bin/bash` or \n`#/bin/xonsh` to the head if it is written in the `bash` or `xonsh` language\n- `continuous_evaluation.py` to include all the `KPI`s this task tracks\n- `latest_kpis` directory, include all the baseline files\n\n### PR and Add to Service\n- PR to `fast` branch, and run `ce-kpi-fast-test` test on teamcity,\n- if passed, PR from `fast` to `master` branch.\n\n### Add new KPI to track\nReference the interface [kpi.py](https://github.com/PaddlePaddle/continuous_evaluation/blob/develop/continuous_evaluation_py23/kpi.py), there are two basic KPIs:\n\n- LessWorseKpi\n- GreaterWorseKpi"},{"ref":"P9","kind":"page","title":"PaddlePaddle/docs repository metadata","date":"2026-06-11T03:57:59.641968+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/docs","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/docs\n\nDescription: Documentations for PaddlePaddle\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 282\n\nForks: 907\n\nOpen issues: 51\n\nCreated: 2018-06-07T08:58:28Z\n\nPushed: 2026-06-09T09:35:04Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n# PaddlePaddle docs\n\nEnglish | [简体中文](./README_cn.md) | [日本語](./README_ja.md)\n\nSource files for contents presented at [PaddlePaddle documentation](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/index_cn.html).\n\nNote: English version API docs are generally docstrings in [PaddlePaddle/Paddle](https://github.com/PaddlePaddle/Paddle), documents for [other PaddlePaddle projects](https://www.paddlepaddle.org.cn/overview) are being managed in their respective ways.\n\n## Codebase structure\n\n- [docs](docs): PaddlePaddle 2.0 & above docs source file.\n- [docs/api](docs/api): PaddlePaddle API docs.\n- [docs/guides](docs/guides): PaddlePaddle guides docs.\n- [docs/tutorial](docs/tutorial): PaddlePaddle tutorial docs.\n- [ci_scripts](ci_scripts): docs CI scripts.\n\n## How to build\n\n- pre-requirements\n- docker\n- Instructions\n- step1: clone docs\n```\ngit clone https://github.com/PaddlePaddle/docs\n```\n- step2: build docs\n```\ncd docs\nmkdir output\nbash docs-build.sh -f absolute_path_docs\n```\n- step3: preview docs\nThe output of docs will be generated in docs/output.\n\n## How to contribute\n\nPaddlePaddle welcomes documentation contributions, please see [CONTRIBUTING.md](./CONTRIBUTING.md) for details.\n\n## License\n\n[Apache License 2.0](LICENSE)"},{"ref":"P10","kind":"page","title":"PaddlePaddle/tape repository metadata","date":"2026-06-11T03:57:59.579337+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/tape","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/tape\n\nLanguage: C++\n\nStars: 14\n\nForks: 9\n\nOpen issues: 6\n\nCreated: 2018-06-18T18:53:15Z\n\nPushed: 2020-01-14T09:42:14Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Dynamic Graph on Fluid\n\nPaddlePaddle Fluid is targeting the autodiff without tape, which, however, is very\nchallenging and we are still way from there. DyNet and PyTorch provide a good design\nidea, the *tape*, that significantly eases the challenge. Also, DyNet provides\na C++ API that is as convenient as Python but with higher efficiency and could\nconveniently integrate with industrial/production systems. This package, `tape`,\ncombines the good of\n\n1. tape from PyTorch and DyNet\n2. C++ API and core from DyNet\n3. rich set of operators from PaddlePaddle\n\n## Overview\n\nWe can implement Dynet-like Tape(See this [survey](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/survey/dynamic_graph.md))\nby wrapping Paddle Fluid's `Operator` and `Variable`.\n\nThe user API is straight forward since\n\n1. it is imperative. And it uses host language's control flow logic.\n1. it avoids extra concepts such as `Scope` and `Executor`.\n\nAll of these benefits come at the cost of just adding one line `reset_global_tape`\nat every iteration.\n\n## Code Structure\n\nIn short, the `Tape` contains a vector of `OpHandle`s. And an `OpHandle` contains its\n`type`, the pointers to the `Variable`s, and necessary attributes.\n\n```c++\nclass Variable {\npublic:\nVriableHandle Grad(); // returns its gradient variable\nprivate:\nframework::VarDesc desc_; // compile time infershape, necessary for lazy execution\nframework::Variable var_; // run time variable, holds data memory\n};\n\nusing VariableHandle = shared_ptr<Variable>;\n\nstruct OpHandle {\nstring type_;\nmap<string, vector<VariableHandle>> inputs_;\nmap<string, vector<VariableHandle>> outputs_;\nAttributeMap attrs_;\n};\n\nclass Tape {\npublic:\nvoid AddOp(OpHandle); // add op\nvoid Forward(); // execute the tape_\nvoid Backward(); // execute the backward of the tape_\nprivate:\nvector<OpHandle> tape_;\n};\n```\n\nWe uses `Function` to indicate layers. It takes care of parameter\ninitialization and `AddOp` to the Tape when it is called.\n\n```c++\nclass Linear {\npublic:\nLinear(int in_di"},{"ref":"P11","kind":"page","title":"PaddlePaddle/PaddleFormers repository metadata","date":"2026-06-11T03:57:59.358723+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleFormers","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleFormers\n\nDescription: PaddleFormers is an easy-to-use library of pre-trained large language model zoo based on PaddlePaddle.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 12983\n\nForks: 2194\n\nOpen issues: 241\n\nCreated: 2018-12-21T06:00:48Z\n\nPushed: 2026-06-11T03:41:53Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"https://github.com/user-attachments/assets/9d1c1937-7fac-48f8-9d61-f7ac67b61b18\" align=\"middle\" width=\"500\" />\n</p>\n\n------------------------------------------------------------------------------------------\n\n<p align=\"center\">\n<a href=\"\"><img src=\"https://img.shields.io/badge/python-3.10+-aff.svg\"></a>\n<a href=\"\"><img src=\"https://img.shields.io/badge/os-linux%2C%20win-pink.svg\"></a>\n<a href=\"./LICENSE\"><img src=\"https://img.shields.io/badge/license-Apache%202-dfd.svg\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleFormers/stargazers\"><img src=\"https://img.shields.io/github/stars/PaddlePaddle/PaddleFormers?color=ccf\"></a>\n</p>\n\n<h4 align=\"center\">\n<a href=#最新更新> 最新更新 </a> |\n<a href=#特性> 特性 </a> |\n<a href=#安装> 安装 </a> |\n<a href=#快速体验> 快速体验 </a> |\n<a href=#社区交流> 社区交流 </a>\n</h4>\n\n# PaddleFormers\n## 📝简介\nPaddleFormers 是基于百度深度学习框架 PaddlePaddle 搭建的 Transformers 库，旨在为 PaddlePaddle 生态提供与 Hugging Face Transformers 项目对等的模型接口与功能体验，支持大语言模型（LLM）与视觉语言模型（VLM）的训练能力。PaddleFormers 充分发挥 PaddlePaddle 在高性能训练方面的内置优势，全面支持包括张量并行、流水线并行和专家并行在内的主流大模型分布式训练策略，以及自动混合精度等加速技术，在 DeepSeek-V3、GLM-4.5-Air 等重点模型上，训练性能明显超越 Megatron-LM ，实现了高效的预训练与后训练性能。\n\n结合业界主流优化方法与飞桨在业务实践中积累的高效特性，PaddleFormers 致力于打造**高性能、低资源占用**的训练体验，帮助用户高效便捷地完成大模型训练，而无需关注底层复杂的优化细节。\n\n## 🆕最新更新\n* 2026.03.31 - PaddleFormers v1.1 正式发布！在这个版本中我们支持了 GLM-4.5 系列模型的单步与多步 MTP 训练能力。依托 MTP 架构优势，开发者可显著提升推理效率；同时针对 MTP 模块训练场景，我们新增主干网络冻结开关，灵活满足各类模型精细化调优需求。此外，我们对视觉理解类模型进行了深度优化，Qwen3-VL 30B-A3B 模型性能相比上个版本提升48%，领先Megatron-LM 6%。\n* 2026.01.21 - PaddleFomers v1.0版本发布啦！我们提供了针对 LLM 和 VLM 等模型的训练能力，针对 DeepSeek-V3模型和 GLM-4.5-Air 等重点模型，我们实现了极致性能优化（训练性能明显超越 Megatron-LM ）。针对 PaddleOCR-VL，我们在昆仑芯 P800、天数天垓150等国产计算芯片上进行了适配，更好的满足国内用户需求。\n\n## ✨特性\n* **丰富的模型支持：** PaddleFormers 实现了对于 100+ 主流的大语言模型和视觉语言模型的训练能力支持，涵盖了 DeepSeek-V3、GLM-4.5系列、Qwen2和 Qwen3系列、Qwen3-VL 等前沿模型。同时提供了对 ERNIE-4.5、ERNIE-4.5-"},{"ref":"P12","kind":"page","title":"PaddlePaddle/PaddleFleetX repository metadata","date":"2026-06-11T03:57:59.326475+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleFleetX","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleFleetX\n\nDescription: 飞桨大模型开发套件，提供大语言模型、跨模态大模型、生物计算大模型等领域的全流程开发工具链。\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 480\n\nForks: 165\n\nOpen issues: 42\n\nCreated: 2018-12-12T15:45:00Z\n\nPushed: 2024-05-24T01:10:17Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"./paddlefleetx-logo.png\" align=\"middle\" width=\"350\" />\n</p>\n\n------------------------------------------------------------------------------------------\n\n<p align=\"center\">\n<a href=\"./LICENSE\"><img src=\"https://img.shields.io/badge/license-Apache%202-dfd.svg\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleFleetX/releases\"><img src=\"https://img.shields.io/github/v/release/PaddlePaddle/PaddleFleetX?color=ffa\"></a>\n<a href=\"\"><img src=\"https://img.shields.io/badge/python-3.7+-aff.svg\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleFleetX/graphs/contributors\"><img src=\"https://img.shields.io/github/contributors/PaddlePaddle/PaddleFleetX?color=9ea\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleFleetX/issues\"><img src=\"https://img.shields.io/github/issues/PaddlePaddle/PaddleFleetX?color=9cc\"></a>\n<a href=\"https://github.com/PaddlePaddle/PaddleFleetX/stargazers\"><img src=\"https://img.shields.io/github/stars/PaddlePaddle/PaddleFleetX?color=ccf\"></a>\n</p>\n\n## 简介\n\nPaddleFleetX是基于飞桨深度学习框架开发的大模型套件，旨在提供高性能、灵活易用的大模型全流程应用能力，在**开发**、**训练**、**精调**、**压推**、**推理**、**部署**六大环节提供端到端全流程优化。\n\n<p align=\"center\">\n<img width=\"1000\" alt=\"飞桨大模型套件\" src=\"https://github.com/PaddlePaddle/PaddleFleetX/assets/1371212/ab5e87cc-df52-48cb-9968-8951d3b164ba\">\n</p>\n\n## 特色介绍\n\n### 大模型开发：动静统一开发模式，4D混合并行策略灵活配置\n\n<p align=\"center\">\n<img width=\"771\" alt=\"大模型开发\" src=\"https://github.com/PaddlePaddle/PaddleFleetX/assets/1371212/95d1c0e8-df92-489b-8472-0a8b438cbfcf\">\n</p>\n\n基于飞桨动静统一的开发模式，大模型套件全面使用动态图开发，在Generate API中可自动完成算子融合具备静态图的调试性能。全场景统一训练器Trainer可以轻松完成4D混合并行的配置，在预训练与精调环节皆可使用。\n\n### 大模型训练：发挥基础计算潜能、全面提升分布式效率\n\n飞桨针对大模型训练，对数据读取、混合精度计算策略、高性能算子库、并行策略自动寻优、流水线调度的整个全流程实现优化，助力文心大模型训练速度提升3倍。\n\n<p align=\"center\"> \n<img width=\"1000\" alt=\"飞桨支持大模型训练\" src=\"https://github.com/PaddlePaddle/PaddleFleetX/assets/1371212/3874440d-0b0c-4730-bbcb-f9b87900d75f\">\n</p>\n\n### 大模型精调：主流精调算法实现性能全面领先\n\n提供了主流的精调算法，包括SFT、Pr"},{"ref":"P13","kind":"page","title":"PaddlePaddle/X2Paddle repository metadata","date":"2026-06-11T03:57:59.058006+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/X2Paddle","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/X2Paddle\n\nDescription: Deep learning model converter for PaddlePaddle. (『飞桨』深度学习模型转换工具)\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 772\n\nForks: 162\n\nOpen issues: 273\n\nCreated: 2019-01-09T05:46:03Z\n\nPushed: 2025-10-22T02:49:23Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n# X2Paddle\n\n[![PyPI - X2Paddle Version](https://img.shields.io/pypi/v/x2paddle.svg?label=pip&logo=PyPI&logoColor=white)](https://pypi.org/project/x2paddle/)\n[![PyPI Status](https://pepy.tech/badge/x2paddle/month)](https://pepy.tech/project/x2paddle)\n[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)\n[![Version](https://img.shields.io/github/release/PaddlePaddle/X2Paddle.svg)](https://github.com/PaddlePaddle/X2Paddle/releases)\n![python version](https://img.shields.io/badge/python-3.8+-orange.svg)\n\n## 简介\n\nX2Paddle是飞桨生态下的模型转换工具，致力于帮助其它深度学习框架用户快速迁移至飞桨框架。目前支持**推理模型的框架转换**与**PyTorch训练代码迁移**，我们还提供了详细的不同框架间API对比文档，降低开发者将模型迁移到飞桨的时间成本。\n\n## 特性\n\n- **支持主流深度学习框架**\n\n- 目前已经支持Caffe/TensorFlow/ONNX/PyTorch四大框架的预测模型的转换，PyTorch训练项目的转换，涵盖了目前市面主流深度学习框架，详见 ***[支持模型列表](./docs/introduction/x2paddle_model_zoo.md)*** 和 ***[测试 demo ](./test_benchmark)***\n\n- **支持的模型丰富**\n\n- 在主流的CV和NLP模型上支持大部分模型转换，目前X2Paddle支持130+ PyTorch OP，90+ ONNX OP，90+ TensorFlow OP 以及 30+ Caffe OP，详见 ***[支持列表](./docs/inference_model_convertor/op_list.md)***\n\n- **简洁易用**\n\n- 一条命令行或者一个API即可完成模型转换\n\n## 能力\n\n- **预测模型转换**\n\n- 支持Caffe/TensorFlow/ONNX/PyTorch的模型一键转为飞桨的预测模型，并使用PaddleInference/PaddleLite进行CPU/GPU/Arm等设备的部署\n\n- **PyTorch训练项目转换**\n\n- 支持PyTorch项目Python代码（包括训练、预测）一键转为基于飞桨框架的项目代码，帮助开发者快速迁移项目，并可享受[AIStudio平台](https://aistudio.baidu.com/)对于飞桨框架提供的海量免费计算资源[**【新功能，试一下！】**](/docs/pytorch_project_convertor/README.md)\n\n- **API对应文档**\n\n- 详细的API文档对比分析，帮助开发者快速从PyTorch框架的使用迁移至飞桨框架的使用，大大降低学习成本 [**【新内容，了解一下！】**](docs/pytorch_project_convertor/API_docs/README.md)\n\n## 安装\n\n### 环境依赖\n- python >= 3.8\n- paddlepaddle >= 2.2.2 (官方验证到 `3.0.0beta1`)\n- tensorflow == 1.14 (如需转换TensorFlow模型。其中 `test_benchmark` 模型已在 `2.16.1` 测试通过)\n- onnx >= 1.6.0 (如需转换ONNX模型，其中 `test_benchmark` 模型已在 `1.17.0` 测试通过)\n- torch >= 1.5.0 (如需转换PyTorch模型，其中 `test_benchmark` 模型已在 `2.4.1` 测试通过)\n- paddlelite >= 2.9.0 (如需一键转换成Paddle-Lite支持格式,推荐最新版本)\n\n> 说明："},{"ref":"P14","kind":"page","title":"PaddlePaddle/AutoDL repository metadata","date":"2026-06-11T03:57:58.898966+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/AutoDL","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/AutoDL\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 157\n\nForks: 30\n\nOpen issues: 6\n\nCreated: 2019-02-21T06:16:48Z\n\nPushed: 2020-01-07T07:54:57Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Introduction to AutoDL Design\n\n## Content\n- [Installation](#Installation)\n- [Introduction](#Introduction)\n- [Data Preparation](#Data-Preparation)\n- [Model Training](#Model-Training)\n\n## Installation\nRunning demo code in the current directory requires PadddlePaddle Fluid v.1.3.0 or above. If your runtime environment does not meet this requirement, please update PaddlePaddle according to the documents.\n* Install Python2.7\n* Install dependencies [PARL](https://github.com/PaddlePaddle/PARL) framework and [absl-py](https://github.com/abseil/abseil-py/tree/master/absl) library，as follows:\n```\npip install parl\npip install absl-py\n```\n\n## Introduction\n[AutoDL](http://www.paddlepaddle.org/paddle/ModelAutoDL) is an efficient automated neural architecture design method. It designs quality customized neural architecture via reinforcement learning. The system consists of two components: an encoder of the neural architecture, and a critic of the model performance. The encoder encodes neural architecture using a recurrent neural network, and the critic evaluates the sampled architecture in terms of accuracy, number of model parameters, etc., which are fed back to the encoder. The encoder updates its parameters accordingly, and samples a new batch of architectures. After several iterations, the encoder is trained to converge and finds a quality architecture. The open-sourced AutoDl Design is one implementation of AutoDL technique. Section 2 presents the usage of AutoDL. Section 3 presents the framework and examples.\n\n## Data Preparation\n* Clone [PaddlePaddle/AutoDL](https://github.com/PaddlePaddle/AutoDL.git) to local machine，and enter the path of AutoDL Design. \n* Download [CIFAR-10](https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz) training data, unzip to AutoDL Design/cifar, and generate a dataset of 10 classes and 100 images per class using `dataset_maker.py`\n```\ntar zxf cifar-10-python.tar.gz\npython dataset_maker.py\n```\n\n## Model Training\nIn the tra"},{"ref":"P15","kind":"page","title":"PaddlePaddle/benchmark repository metadata","date":"2026-06-11T03:57:58.879769+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/benchmark","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/benchmark\n\nLanguage: Python\n\nStars: 82\n\nForks: 157\n\nOpen issues: 64\n\nCreated: 2019-02-28T03:14:16Z\n\nPushed: 2026-04-07T02:25:20Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# PaddlePaddle Benchmark\n\n我们对PaddlePaddle的最新版本[v1.5.0](https://github.com/PaddlePaddle/Paddle/tree/v1.5.0)，在训练性能和显存占用方面进行了基准测试。\n\n## 目录\n* [测试环境](#测试环境)\n* [智能视觉（PaddleCV）](#PaddleCV)\n* [SE-ResNeXt50](#SE-ResNeXt50)\n* [Mask-RCNN](#Mask-RCNN)\n* [YOLOv3](#YOLOv3)\n* [DeepLab V3+](#deepLab-v3)\n* [Cycle-GAN](#Cycle-GAN)\n* [智能文本处理（PaddleNLP）](#PaddleNLP)\n* [PaddingRNN](#PaddingRNN)\n* [BERT](#BERT)\n* [Transformer](#Transformer)\n* [强化学习（PARL）](#PARL)\n* [DDPG](#DDPG)\n\n## 测试环境\n- 测试对象\n- 本次测试[PaddlePaddle v1.5.0](https://github.com/PaddlePaddle/Paddle/tree/v1.5.0)，具体commit是：`401c03fc20478f5cc067440422fc3a7b306d0e32`\n- 基准测试程序[benchmark](https://github.com/PaddlePaddle/benchmark)，具体commit是：`3c34ed6b166f6b77e759b4c54e8854652ad3d776`\n\n- Docker镜像\n- Paddle编译镜像\n- CUDA 9.0，`paddlepaddle/paddle_manylinux_devel:cuda9.0_cudnn7`\n- CUDA 10.0，`paddlepaddle/paddle_manylinux_devel:cuda10.0_cudnn7`\n- Paddle测试镜像\n- CUDA 9.0，`paddlepaddle/paddle:latest-gpu-cuda9.0-cudnn7`\n- CUDA 10.0，`paddlepaddle/paddle:latest-gpu-cuda10.0-cudnn7`\n- TensorFlow测试镜像\n- CUDA 9.0，`tensorflow/tensorflow:1.12.0-gpu`\n- CUDA 10.0，`tensorflow/tensorflow:1.14.0-gpu`\n- PyTorch\n- CUDA 9.0，\n- CUDA 10.0，\n\n- GPU服务器参数\n- GPU型号：Nvidia Tesla V100-SXM2，显存16 GB\n- CPU型号：Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz，38核\n- Driver Version: 418.39\n- CUDA Version：9.0.176，10.0.130\n- NCCL Version：2.4.2\n- cuDNN Version：7.4.2.24，7.5.0.56\n\n注意：测试所用GPU服务器为虚拟机，跟相同配置的物理机测试结果可能会有一定的差别。\n\n- CPU服务器参数\n- CPU型号：Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz，24核\n- 指令集：AVX2\n\n## PaddleCV\n\n| 方向 | 模型 | Paddle | TensorFlow | PyTorch | MXNet | 数据集 | batch_size(单卡) |\n|---|---|---|---|---|---|---|---|\n| 图像分类 | SE-ResNeXt50 | [PaddleCV/image_classification](https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification) | - | [SENet-PyTorch](https://github.com/miraclewkf/SENet-PyTorch) | - | ILSVRC2012 | 32 |\n| 目标检测 | Mask-RCNN | [PaddleCV/rcnn](https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/rcnn) | - | [maskrcnn-benchmark](https://github.com/fa"},{"ref":"P16","kind":"page","title":"PaddlePaddle/ERNIE repository metadata","date":"2026-06-11T03:57:58.629954+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/ERNIE","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/ERNIE\n\nDescription: The official repository for ERNIE 4.5 and ERNIEKit – its industrial-grade development toolkit based on PaddlePaddle.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 7717\n\nForks: 1442\n\nOpen issues: 83\n\nCreated: 2019-03-03T07:31:29Z\n\nPushed: 2026-01-04T12:29:14Z\n\nDefault branch: release/v1.5\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"https://github.com/user-attachments/assets/9ad1ffce-2310-4f80-a3cd-7a117bfb4f17\" width=\"300px\"></a>\n</p>\n\n<div align=\"center\">\n\n[ERNIE Bot](https://ernie.baidu.com/) | [🤗Hugging Face](https://huggingface.co/baidu) | [AI Studio](https://aistudio.baidu.com/modelsoverview)\n\n📑 [Blog](https://yiyan.baidu.com/blog/posts/ernie4.5) | 📚 [Cookbook](./cookbook/) | 📑 [Paper](https://yiyan.baidu.com/blog/publication/) | 🛠️ [Training](./docs/erniekit.md) | ⚡️ [Deploy](https://github.com/PaddlePaddle/FastDeploy)\n\n<a href=\"https://trendshift.io/repositories/14169\" target=\"_blank\"><img src=\"https://trendshift.io/api/badge/repositories/14169\" alt=\"PaddlePaddle%2FERNIE | Trendshift\" style=\"width: 250px; height: 55px;\" width=\"250\" height=\"55\"/></a>\n\n</div>\n\n## 📣 Recent updates\n\n**[2025-11] 🔥 Released ERNIEKit v1.5:**\n\n- **New Features**\n- [ERNIE-4.5-VL-28B-A3B-Thinking] Supports SFT training and function call training for ERNIE-4.5-VL-28B-A3B-Thinking (https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-Thinking).\n\n**[2025-10] 🔥 Released ERNIEKit v1.4:**\n\n- **New Features**\n- VL Model Training: Support SFT for [PaddleOCR-VL-0.9B]((https://huggingface.co/PaddlePaddle/PaddleOCR-VL/tree/main/PaddleOCR-VL-0.9B)) model. More details in [PaddleOCR-VL-0.9B SFT](./docs/paddleocr_vl_sft.md).\n- Dataflow : Support padding-free startegy.\n- Packing data within a batch into a sequence to avoid padding, thereby reducing GPU memory usage and accelerating training.\n\n**[2025-09] 🔥 Released ERNIEKit v1.3:**\n\n- **New Features**\n- [ERNIE-4.5-21B-A3B-Thinking] Supports SFT training and function call training for ERNIE-4.5-21B-A3B-Thinking (https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Thinking).\n\n- **Bug Fixes:**\n- [VL Model Training] Optimization of multimodal video data processing speed (#1266).\n\n**[2025-09] 🔥 Re"},{"ref":"P17","kind":"page","title":"PaddlePaddle/Serving repository metadata","date":"2026-06-11T03:57:58.551302+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/Serving","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/Serving\n\nDescription: A flexible, high-performance carrier for machine learning models（『飞桨』服务化部署框架）\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 920\n\nForks: 246\n\nOpen issues: 16\n\nCreated: 2019-03-31T12:36:25Z\n\nPushed: 2026-02-20T19:16:36Z\n\nDefault branch: v0.9.0\n\nFork: no\n\nArchived: no\n\nREADME:\n(简体中文|[English](./README.md))\n\n<p align=\"center\">\n<br>\n<img src='doc/images/serving_logo.png' width = \"600\" height = \"130\">\n<br>\n<p>\n\n<p align=\"center\">\n<br>\n<a href=\"https://travis-ci.com/PaddlePaddle/Serving\">\n<img alt=\"Build Status\" src=\"https://img.shields.io/travis/com/PaddlePaddle/Serving/develop?style=flat-square\">\n<img alt=\"Docs\" src=\"https://img.shields.io/badge/docs-中文文档-brightgreen?style=flat-square\">\n<img alt=\"Release\" src=\"https://img.shields.io/badge/release-0.9.0-blue?style=flat-square\">\n<img alt=\"Python\" src=\"https://img.shields.io/badge/python-3.6/3.7/3.8/3.9-blue?style=flat-square\">\n<img alt=\"License\" src=\"https://img.shields.io/github/license/PaddlePaddle/Serving?color=blue&style=flat-square\">\n<img alt=\"Forks\" src=\"https://img.shields.io/github/forks/PaddlePaddle/Serving?color=yellow&style=flat-square\">\n<img alt=\"Issues\" src=\"https://img.shields.io/github/issues/PaddlePaddle/Serving?color=yellow&style=flat-square\">\n<img alt=\"Contributors\" src=\"https://img.shields.io/github/contributors/PaddlePaddle/Serving?color=orange&style=flat-square\">\n<img alt=\"Community\" src=\"https://img.shields.io/badge/join-Wechat,QQ-orange?style=flat-square\">\n</a>\n<br>\n<p>\n\n***\n\n**【更新说明】**\n我们在新开源项目FastDeploy里面，基于Triton Inference Server，集成FastDeploy Runtime(包括Paddle Inference、ONNX Runtime、TensorRT以及OpenVINO等)，可支持飞桨模型的高性能服务化部署，对服务化部署有需求的开发者，可以参考如下文档进行使用，有任何问题，欢迎在FastDeploy开源项目里通过issue反馈。\n- [FastDeploy服务化部署](https://github.com/PaddlePaddle/FastDeploy/blob/develop/serving/README_CN.md)\n\nPaddle Serving 依托深度学习框架 PaddlePaddle 旨在帮助深度学习开发者和企业提供高性能、灵活易用的工业级在线推理服务。Paddle Serving 支持 RESTful、gRPC、bRPC 等多种协议，提供多种异构硬件和多种操作系统环境下推理解决方案，和多种经典预训练模型示例。核心特性如下：\n\n- 集成高性能服务端推理引擎 [Paddle Inference](https://paddleinference.paddlepaddle.org.cn/product_introduction/inference_intro.html) 和端侧引擎 [Paddle Lite](https://paddlelite.paddlepaddle.org.cn/introduction/tech_highlights.html)，其他机器学习平台（Caff"},{"ref":"P18","kind":"page","title":"PaddlePaddle/any repository metadata","date":"2026-06-11T03:57:58.190982+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/any","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/any\n\nDescription: Legacy Repo only for PaddlePaddle with version <= 1.3\n\nLanguage: C++\n\nLicense: BSL-1.0\n\nStars: 5\n\nForks: 6\n\nOpen issues: 0\n\nCreated: 2019-05-30T03:28:02Z\n\nPushed: 2019-05-30T03:33:14Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Any\n\nThis is a implementation of [N4562](http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4562.html) std::experimental::any (merged into C++17) for C++11 compilers.\n\nIt contains a small object optimization for objects with a size of up to 2 words (such as `int`, `float` and `std::shared_ptr`). Storing those objects in the container will not trigger a dynamic allocation.\n\nFor a easy to understand documentation, see [cppreference](http://en.cppreference.com/w/cpp/experimental/any), except our namespace is `linb`.\n\n## Defines\n\nYou may additionally define the following preprocessor symbols (making the implementation non-standard):\n\n+ `ANY_IMPL_ANY_CAST_MOVEABLE`: This implements a fix proposed in [LWG Defect 2509](https://cplusplus.github.io/LWG/lwg-active.html#2509). This will cause the expressions `T x = any_cast<T>(any(T()))` and `T x = any_cast<T&&>(any(T()))`to perform a move into `x` instead of a copy.\n+ `ANY_IMPL_FAST_TYPE_INFO_COMPARE`: When checking if two `typeid` are the same, performs just a pointer comparision instead of the actual `type_info::operator==` comparision. Be aware this isn't recommended unless you know what you're doing."},{"ref":"P19","kind":"page","title":"PaddlePaddle/PGL repository metadata","date":"2026-06-11T03:57:58.174544+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PGL","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PGL\n\nDescription: Paddle Graph Learning (PGL) is an efficient and flexible graph learning framework based on PaddlePaddle\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 1587\n\nForks: 309\n\nOpen issues: 63\n\nCreated: 2019-06-11T03:23:28Z\n\nPushed: 2023-12-11T05:15:14Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<img src=\"./docs/source/_static/logo.png\" alt=\"The logo of Paddle Graph Learning (PGL)\" width=\"320\">\n\n[![PyPi Latest Release](https://img.shields.io/pypi/v/pgl.svg)](https://pypi.org/project/pgl/)\n[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](./LICENSE)\n\n[DOC](https://pgl.readthedocs.io/en/latest/) | [Quick Start](https://pgl.readthedocs.io/en/latest/quick_start/instruction.html) | [中文](./README.zh.md)\n\n## Breaking News !!\n\n**One amazing paper about knowledge representation learning was accepted!** (2022.05.06)\n\n- Simple and Effective Relation-based Embedding Propagation for Knowledge Representation Learning, to appear in **IJCAI2022**. Code can be found [here](./apps/Graph4KG/examples/REP).\n\nPGL v2.2 2021.12.20\n\n- **Graph4Rec**: We released a universal and large-scale toolkit with graph neural networks for recommender systems. Details can be found [here](./apps/Graph4Rec).\n\n- **Graph4KG**: We released a flexible framework named Graph4KG to learn embeddings of entities and relations in KGs, which supports training on massive KGs. Details can be found [here](./apps/Graph4KG).\n\n- **GNNAutoScale**: PGL now supports GNNAutoScale framework, which can scale arbitrary message-passing GNNs to large graphs. Details can be found [here](./apps/GNNAutoScale).\n\n&#x1F525; &#x1F525; &#x1F525; **OGB-LSC KDD CUP 2021 winners announced!!** (2021.06.17)\n\nSuper excited to announce our PGL team won <font color=Red>**TWO FIRST**</font> place and <font color=Red>**ONE SECOND**</font> place in a total of three track in OGB-LSC KDD CUP 2021.\nLeaderboards can be found [here](https://ogb.stanford.edu/kddcup2021/results/).\n\n- **First place in MAG240M-LSC track**: Code and Technical Report can be found [here](./examples/kddcup2021/MAG240M/r_unimp).\n\n- **First place in WikiKG90M-LSC track**: Code and Technical Report can be found [here]("},{"ref":"P20","kind":"page","title":"PaddlePaddle/MetaGym repository metadata","date":"2026-06-11T03:57:58.037487+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/MetaGym","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/MetaGym\n\nDescription: Collection of Reinforcement Learning / Meta Reinforcement Learning Environments.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 302\n\nForks: 61\n\nOpen issues: 13\n\nCreated: 2019-07-13T16:37:35Z\n\nPushed: 2024-07-15T17:29:11Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# MetaGym\n\nMetaGym provides abundant environments for benchmarking Reinforcement Learning and Meta Reinforcement Learning\n\n# Environments Updating\n\n- [LiftSim](metagym/liftsim)：Simulator for Evelvator Dispatching (Sep, 2019)\n\n- [Quadrotor](metagym/quadrotor): 3D Quadrotor simulator for different tasks (Mar, 2020)\n\n- [Quadrupedal](metagym/quadrupedal): Quadrupedal robot adapting to different terrains (Seq, 2021)\n\n- [MetaMaze](metagym/metamaze): 2D/3D maze generators for task generalization (Oct, 2021)\n\n- [MetaLocomotion](metagym/metalocomotion): Locomotion simulator with diverse geometries (June, 2022)\n\n- [MetaLM](metagym/metalm): Meta language model dataset (Dec, 2022)\n\n- [Bandits](metagym/bandits): Bandits task generalization (Dec, 2022)"},{"ref":"P21","kind":"page","title":"PaddlePaddle/Paddle-Lite-Demo repository metadata","date":"2026-06-11T03:57:57.864656+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/Paddle-Lite-Demo","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/Paddle-Lite-Demo\n\nDescription: lib, demo, model, data\n\nLanguage: C++\n\nLicense: Apache-2.0\n\nStars: 797\n\nForks: 305\n\nOpen issues: 27\n\nCreated: 2019-08-21T05:40:51Z\n\nPushed: 2025-09-17T06:05:56Z\n\nDefault branch: develop\n\nFork: no\n\nArchived: no\n\nREADME:\n# Paddle-Lite-Demo\n\nPaddle-Lite 提供了多个应用场景的 demo，并支持 Android、iOS 和 ArmLinux 三个平台：\n* 图像分类\n* 基于 [mobilenet_v1](https://paddlelite-demo.bj.bcebos.com/models/mobilenet_v1_fp32_224.tar.gz) 模型\n* [Android 示例](./image_classification/android/)\n* [iOS 示例](./image_classification/ios/)\n* [ArmLinux 示例](./image_classification/armlinux/)\n* 目标检测\n* 基于 [ssd_mobilenetv1](https://paddlelite-demo.bj.bcebos.com/demo/object_detection/models/ssd_mobilenet_v1_pascalvoc_fp32_300_fluid.tar.gz) 模型\n* [Android 示例](./object_detection/android/app/cxx/ssd_mobilenetv1_detection_demo/)\n* [iOS 示例](./object_detection/ios/ssd_mobilenetv1_demo/)\n* 基于 [yolov3_mobilenet_v3](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/lite/yolov3_mobilenet_v3.tar) 模型\n* [Android 示例](./object_detection/android/app/cxx/yolo_detection_demo/)\n* [iOS 示例](./object_detection/ios/yolov3_mobilenet_v3_demo/)\n* 基于 [yolov5](https://paddlelite-demo.bj.bcebos.com/models/yolov5n/yolov5n.zip) 模型\n* [Android 示例](./object_detection/android/app/cxx/yolov5n_detection_demo/)\n* iOS 示例\n* 基于 [pp_picodet](https://paddlelite-demo.bj.bcebos.com/demo/object_detection/models/picodet_s_320_coco_for_cpu.tar.gz) 模型\n* [Android 示例](./object_detection/android/app/cxx/picodet_detection_demo/)\n* [iOS 示例](./object_detection/ios/picodet_demo/)\n* 文字识别\n* 基于 [pp_ocr_det](https://paddleocr.bj.bcebos.com/dygraph_v2.0/slim/ch_ppocr_mobile_v2.0_det_slim_infer.tar)、[pp_ocr_rec](https://paddleocr.bj.bcebos.com/dygraph_v2.0/slim/ch_ppocr_mobile_v2.0_rec_slim_infer.tar) 和 [pp_ocr_cls](https://paddleocr.bj.bcebos.com/dygraph_v2.0/slim/ch_ppocr_mobile_v2.0_cls_slim_infer.tar) 模型\n* [Android 示例](./ocr/android/)\n* [iOS 示例](./ocr/ios/)\n* 人脸检测\n* 基于 [face-detection](https://paddlelite-demo.bj.bcebos.com/models/facedetection_fp32_240_430_fluid.tar.gz) 模型\n* [Android 示例](./face_detection/android/)\n* [iOS 示例](./face_detection/ios/face_detection/)\n* 人脸关键点检测\n* 基于 [face-detection](https://paddlelite-demo."},{"ref":"P22","kind":"page","title":"PaddlePaddle/epep repository metadata","date":"2026-06-11T03:57:57.786594+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/epep","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/epep\n\nDescription: Easy & Effective Application Framework for PaddlePaddle\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 34\n\nForks: 8\n\nOpen issues: 1\n\nCreated: 2019-08-22T07:16:21Z\n\nPushed: 2020-07-04T06:44:41Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# Easy Paddle, Effective Paddle\n\n[![Build Status](https://travis-ci.org/PaddlePaddle/epep.svg?branch=master)](https://travis-ci.org/PaddlePaddle/epep)\n[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](https://github.com/PaddlePaddle/epep)\n[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)\n\n**EPEP** is an Application Framework for PaddlePaddle, to make everyone can easily learn and use. 目前已经被广泛应用在百度内部业务，显著提升单机CPU, 单机GPU, 多机多卡的模型迭代效率\n\n## 目录\n* [环境搭建](#环境搭建)\n* [框架说明](#框架说明)\n* [使用说明](#使用说明)\n\n## 环境搭建\n\n1. Linux CentOS 6.3, Python 2.7, 获取PaddlePaddle v1.6.1版本以上, 请参考[安装指南](http://www.paddlepaddle.org/#quick-start)进行安装\n\n2. 配置修改conf/var_sys.conf\n```\nfluid_bin=/home/epep/tools/paddle_release_home/python/bin/python\n\n#gpu训练配置\ncuda_lib_path=/home/epep/tools/cuda-9.0/lib64:/home/epep/tools/cudnn/cudnn_v7.3/cuda/lib64:/home/epep/tools/nccl-2.2_cuda-8.0/lib:$LD_LIBRARY_PATH\n```\n\n## 框架说明\n\n### 整体框架\n![EPEP Frame Overview](docs/frame.png)\n\n![EPEP Train Overview](docs/train_diff.png)\n\n![EPEP Pred Overview](docs/pred_diff.png)\n\n## 使用说明\n\n框架提供了一些NLP的例子，主要包括分类，回归，匹配，标注，翻译，生成等\n\n这里以LR为例，用户只要写20行相关代码即可完成，全是业务模型相关，通过epep轻松一键CPU->GPU, GPU多卡，多机多卡(TODO with Easy-DL)\n\n### 1. 定义输入\n\n```python\nclass LinearRegression(BaseDataset):\ndef __init__(self, flags):\nsuper(LinearRegression, self).__init__(flags)\n\n#输入的定义\ndef parse_context(self, inputs):\n\"\"\"\nset inputs_kv: please set key as the same as layer.data.name\nnotice:\n(1)\nIf user defined \"inputs key\" is different from layer.data.name,\nthe frame will rewrite \"inputs key\" with layer.data.name\n(2)\nThe param \"inputs\" will be passed to user defined nets class through\nthe nets class interface function : net(self, FLAGS, inputs), \n\"\"\"\ninputs['x'] = fluid.layers.data(name=\"x\", shape=[self._flags.input_size], dtype=\"float32\")\ninputs['y'] = fluid.layers.data(name=\"y\", shape=[1], dtype=\"float32\")\n\ncontext = {\"inpu"},{"ref":"P23","kind":"page","title":"PaddlePaddle/examples repository metadata","date":"2026-06-11T03:57:57.491327+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/examples","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/examples\n\nLanguage: Python\n\nStars: 17\n\nForks: 13\n\nOpen issues: 0\n\nCreated: 2019-08-23T01:15:57Z\n\nPushed: 2020-03-30T15:59:25Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# Examples\n\n## example guide\n- api_examples: examples for each public api in paddle fluid\n- community_examples: community examples contributed by third party developers"},{"ref":"P24","kind":"page","title":"PaddlePaddle/PaddleSeg repository metadata","date":"2026-06-11T03:57:57.425045+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleSeg","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleSeg\n\nDescription: Easy-to-use image segmentation library with awesome pre-trained model zoo, supporting wide-range of practical tasks in Semantic Segmentation, Interactive Segmentation, Panoptic Segmentation, Image Matting, 3D Segmentation, etc.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 9340\n\nForks: 1710\n\nOpen issues: 28\n\nCreated: 2019-08-26T02:32:22Z\n\nPushed: 2026-02-05T16:49:17Z\n\nDefault branch: release/2.10\n\nFork: no\n\nArchived: no\n\nREADME:\nREADME_CN.md"},{"ref":"P25","kind":"page","title":"PaddlePaddle/PALM repository metadata","date":"2026-06-11T03:57:57.307504+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PALM","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PALM\n\nDescription: a Fast, Flexible, Extensible and Easy-to-use NLP Large-scale Pretraining and Multi-task Learning Framework.\n\nLanguage: Python\n\nStars: 185\n\nForks: 30\n\nOpen issues: 11\n\nCreated: 2019-09-24T05:37:22Z\n\nPushed: 2021-03-29T11:40:54Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# PaddlePALM\n\nEnglish | [简体中文](./README_zh.md)\n\nPaddlePALM (PArallel Learning from Multi-tasks) is a fast, flexible, extensible and easy-to-use NLP large-scale pretraining and multi-task learning framework. PaddlePALM is a high level framework **aiming at fastly developing high-performance NLP models**. \n\nWith PaddlePALM, it is easy to achieve effecient exploration of robust learning of NLP models with multiple auxilary tasks. For example, based on PaddlePALM, the produced robust MRC model, [D-Net](https://github.com/PaddlePaddle/models/tree/develop/PaddleNLP/Research/MRQA2019-D-NET), has achieved **the 1st place** in [EMNLP2019 MRQA](https://mrqa.github.io) track.\n\n<p align=\"center\">\n<img src=\"https://tva1.sinaimg.cn/large/006tNbRwly1gbjkuuwrmlj30hs0hzdh2.jpg\" alt=\"Sample\" width=\"300\" height=\"333\">\n<p align=\"center\">\n<em>MRQA2019 Leaderboard</em>\n</p>\n</p>\n\nBeyond the research scope, PaddlePALM has been applied on **Baidu Search Engine** to seek for more accurate user query understanding and answer mining, which implies the high reliability and performance of PaddlePALM.\n\n#### Features:\n\n- **Easy-to-use:** with PALM, *8 steps* to achieve a typical NLP task. Moreover, all basic components (e.g., the model backbone, dataset reader, task output head, optimizer...) have been decoupled, which allows the replacement of any component to other candidates with quite minor changes of your code. \n- **Built-in Popular NLP Backbones and Pre-trained models:** multiple state-of-the-art general purpose model architectures and pretrained models (e.g., BERT,ERNIE,RoBERTa,...) are built-in. \n- **Easy to play Multi-task Learning:** only one API is needed for jointly training of several tasks with parameters reusement.\n- **Support train/eval with Multi-GPUs:** automatically recognize and adapt to multiple gpus mode to accelerate training and inference.\n- **Pre-training"},{"ref":"P26","kind":"page","title":"PaddlePaddle/PaddleFL repository metadata","date":"2026-06-11T03:57:57.170381+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleFL","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleFL\n\nDescription: Federated Deep Learning in PaddlePaddle\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 515\n\nForks: 120\n\nOpen issues: 56\n\nCreated: 2019-09-25T15:01:39Z\n\nPushed: 2023-07-26T22:30:56Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n<img src='https://github.com/PaddlePaddle/PaddleFL/blob/master/docs/source/_static/FL-logo.png' width = \"400\" height = \"160\">\n\n[DOC](https://paddlefl.readthedocs.io/en/latest/) | [Quick Start](https://paddlefl.readthedocs.io/en/latest/compile_and_intall.html) | [中文](./README_cn.md)\n\n[![Release](https://img.shields.io/github/release/PaddlePaddle/PaddleFL.svg)](https://github.com/PaddlePaddle/PaddleFL/releases)\n[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)\n\nPaddleFL is an open source federated learning framework based on PaddlePaddle. Researchers can easily replicate and compare different federated learning algorithms with PaddleFL. Developers can also benefit from PaddleFL in that it is easy to deploy a federated learning system in large scale distributed clusters. In PaddleFL, several federated learning strategies will be provided with application in computer vision, natural language processing, recommendation and so on. Application of traditional machine learning training strategies such as Multi-task learning, Transfer Learning in Federated Learning settings will be provided. Based on PaddlePaddle's large scale distributed training and elastic scheduling of training job on Kubernetes, PaddleFL can be easily deployed based on full-stack open sourced software.\n\n## Overview of PaddleFL\n\nData is becoming more and more expensive nowadays, and sharing of raw data is very hard across organizations. Federated Learning aims to solve the problem of data isolation and secure sharing of data knowledge among organizations. The concept of federated learning is proposed by researchers in Google [1, 2, 3]. PaddleFL implements federated learning based on the PaddlePaddle framework. Application demonstrations in natural language processing, computer vision and recommendation will be provided in PaddleFL. PaddleFL supports the current two main federated learning strategies[4"},{"ref":"P27","kind":"page","title":"PaddlePaddle/PaddleDetection repository metadata","date":"2026-06-11T03:57:57.037978+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleDetection","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleDetection\n\nDescription: Object Detection toolkit based on PaddlePaddle. It supports object detection, instance segmentation, multiple object tracking and real-time multi-person keypoint detection.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 14241\n\nForks: 3020\n\nOpen issues: 936\n\nCreated: 2019-10-25T07:21:14Z\n\nPushed: 2026-05-28T07:28:31Z\n\nDefault branch: release/2.9\n\nFork: no\n\nArchived: no\n\nREADME:\nREADME_cn.md"},{"ref":"P28","kind":"page","title":"PaddlePaddle/PaddleCraft repository metadata","date":"2026-06-11T03:57:56.700373+00:00","date_source":null,"source_url":"https://github.com/PaddlePaddle/PaddleCraft","signal_url":null,"signal_json_url":null,"text":"# PaddlePaddle/PaddleCraft\n\nDescription: Take neural networks as APIs for human-like AI.\n\nLanguage: Python\n\nStars: 20\n\nForks: 5\n\nOpen issues: 0\n\nCreated: 2019-11-28T09:54:14Z\n\nPushed: 2019-12-04T03:14:09Z\n\nDefault branch: master\n\nFork: no\n\nArchived: yes\n\nREADME:\n# PaddleCraft\n\nA lib of models for human-like AI.\n\n### 什么是PaddleCraft\n\nPaddleCraft是一套基于模型的API，使用PaddleCraft，您可以直接声明并加载各种CV，Speech，NLP等经典人工智能领域的模型（及其预训练参数），并可以通过各种预置功能，对不同模型进行链接，对抗等操作。\n\nPaddleCraft完全基于PaddlePaddle平台开发，暂时计划以插件形式存在，最终稳定后融入paddlepaddle框架。\n\n### 为什么要提出PaddleCraft\n\n借助于算力，大数据，模型工程的不断提升，经典领域的NLP，CV，Speech模型已经逐渐收敛到一些固定结构和范式。虽然已有可以解决特定领域的问题，然而往往仍需要针对特点问题进行加工打磨才能达到商用的目的；另外，由于模型工程的收敛，解决经典场景任务（如文本分类，图像识别）的门槛越来越低，造成竞争白热化。\n\n为了更好的发展，我们需要能够解决更复杂场景，多模态下的AI问题，因此，我们创造了PaddleCraft。\n\n### PaddleCraft能解决什么问题\n\nPaddleCraft为开发者提供了各领域常见且典型的神经网络模型。不同于大部分模型库，PaddleCraft中的模型都以API的形式存在；我们针对各种行业标准化的模型做了细致的整理和加工，使得开发者可以无需了解模型细节，就可以快速使用，解决80%的常见问题。\n\n除此以外，PaddleCraft统一了各个行业解决标准模型的使用方式和数据流，使得不同领域的模型可以相互协作，对抗，亦或者组成一个新的模型。\n\n使用PaddleCraft可以显著减少开发多模态，多智能体等不同场景的AI代码，加快对解决复杂AI任务的研发周期。\n\n### 如何使用PaddleCraft\n\n下面我们以MNIST图像识别模型（image_mlp_encoder）为例，讲解如何使用PaddleCraft：\n\n假设我们现在的行业标准模型为： \n\n![Image result for MLP mnist](https://corochann.com/wp-content/uploads/2017/02/mlp-800x367.png)\n\n如上图所示，这是一个MLP模型，假设两个隐层的size都为784；那么我们如何使用这个模型并完成一些操作呢？在PaddleCraft中，每个模型（以及模型中的op，param）是通过2个名字来控制的：\n\n1. 逻辑名字：所谓逻辑名字，就是这个模型在算法描述中的客观名字，比如在上述模型中，我们用 hidden_layer_1, hidden_layer_2，prediction等名字来描述这个模型中的不同layer。\n2. 物理名字：所谓物理名字，就是这个模型的op在真是硬件环境中的名字，比如paddle.model1.hidden_layer_1.fc_0.w_0。\n\n过去我们大部分人在使用paddle过程中的主要痛苦，源于需要操作物理名字，而缺乏直接手段找到，并通过物理名字操作op和var，这就导致二次开发成本很高。因此，PaddleCraft在设计的时候，设计了一套直接通过逻辑名字就可以操作模型的API，但是仍然返回，并支持使用物理名字对program进行操作；\n\n我们以下面的例子来说明如何使用PaddleCraft\n\n首先安装PaddleCraft：\n\n```shell\npip install paddlecraft\n```\n\n##### case1：常规训练\n\n```python\nimport os\nimport sys\nimport paddle\nimport paddle.fluid as fluid\n\nimport paddlecraft\n\nfrom paddlecraft.image_mlp_encoder import ImageMLPEncoder\n\nif __name__ == \"__main__\":\n\n# build dataset\nmnist_dataset_train = paddle.dataset.mnist.train()\nmnist_dataset_test = paddle.dataset.mnist.test()\n\n# configure the network\n\nimage_encoder1 = ImageMLPEncoder('model1')\nimage_encoder1.build()\n\nloss = fluid.layers.cross_entropy(input"},{"ref":"E1","kind":"event","title":"baidu/Qianfan-OCR","date":"2026-03-18T07:48:43+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/Qianfan-OCR","signal_url":"https://onlylabs.fyi/signals/cbc56756-49f1-4e67-897a-be3b2c3db134","signal_json_url":"https://onlylabs.fyi/signals/cbc56756-49f1-4e67-897a-be3b2c3db134/signal.json","text":"model_released · baidu/Qianfan-OCR · signal_desk=releases · occurred_at=2026-03-18T07:48:43+00:00 · url=https://huggingface.co/baidu/Qianfan-OCR · hf_downloads=174673 · hf_likes=1186 · hf_params=4741408256 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E2","kind":"event","title":"baidu/ERNIE-4.5-21B-A3B-Thinking","date":"2025-09-08T14:18:31+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Thinking","signal_url":"https://onlylabs.fyi/signals/06853c57-f518-4efb-8d74-5c801298e7e6","signal_json_url":"https://onlylabs.fyi/signals/06853c57-f518-4efb-8d74-5c801298e7e6/signal.json","text":"model_released · baidu/ERNIE-4.5-21B-A3B-Thinking · signal_desk=releases · occurred_at=2025-09-08T14:18:31+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Thinking · hf_downloads=12372 · hf_likes=786 · hf_params=21825437888 · pipeline=text-generation · license=apache-2.0"},{"ref":"E3","kind":"event","title":"baidu/ERNIE-Image","date":"2026-04-07T07:26:56+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-Image","signal_url":"https://onlylabs.fyi/signals/a0cd5468-4410-4a1e-8393-2a64fd36a871","signal_json_url":"https://onlylabs.fyi/signals/a0cd5468-4410-4a1e-8393-2a64fd36a871/signal.json","text":"model_released · baidu/ERNIE-Image · signal_desk=releases · occurred_at=2026-04-07T07:26:56+00:00 · url=https://huggingface.co/baidu/ERNIE-Image · hf_downloads=38151 · hf_likes=643 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E4","kind":"event","title":"baidu/ERNIE-4.5-VL-28B-A3B-Thinking","date":"2025-11-07T02:48:32+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-Thinking","signal_url":"https://onlylabs.fyi/signals/947f7039-998c-4ed5-a77b-d6e52f466350","signal_json_url":"https://onlylabs.fyi/signals/947f7039-998c-4ed5-a77b-d6e52f466350/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-28B-A3B-Thinking · signal_desk=releases · occurred_at=2025-11-07T02:48:32+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-Thinking · hf_downloads=613 · hf_likes=540 · hf_params=29664059776 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E5","kind":"event","title":"baidu/ERNIE-Image-Turbo","date":"2026-04-02T10:57:06+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-Image-Turbo","signal_url":"https://onlylabs.fyi/signals/f65a6dd3-1181-4afc-91d0-b5e120ef267a","signal_json_url":"https://onlylabs.fyi/signals/f65a6dd3-1181-4afc-91d0-b5e120ef267a/signal.json","text":"model_released · baidu/ERNIE-Image-Turbo · signal_desk=releases · occurred_at=2026-04-02T10:57:06+00:00 · url=https://huggingface.co/baidu/ERNIE-Image-Turbo · hf_downloads=4533 · hf_likes=393 · pipeline=text-to-image · license=apache-2.0"},{"ref":"E6","kind":"event","title":"baidu/ERNIE-4.5-21B-A3B-PT","date":"2025-06-28T06:13:30+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-PT","signal_url":"https://onlylabs.fyi/signals/e2381804-d403-4275-9d5f-4be74063e9fd","signal_json_url":"https://onlylabs.fyi/signals/e2381804-d403-4275-9d5f-4be74063e9fd/signal.json","text":"model_released · baidu/ERNIE-4.5-21B-A3B-PT · signal_desk=releases · occurred_at=2025-06-28T06:13:30+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-PT · hf_downloads=60949 · hf_likes=175 · hf_params=21948655808 · pipeline=text-generation · license=apache-2.0"},{"ref":"E7","kind":"event","title":"baidu/NAVA","date":"2026-05-29T02:17:22+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/NAVA","signal_url":"https://onlylabs.fyi/signals/e734d8f2-4cd6-4efd-be3a-564ac456a931","signal_json_url":"https://onlylabs.fyi/signals/e734d8f2-4cd6-4efd-be3a-564ac456a931/signal.json","text":"model_released · baidu/NAVA · signal_desk=releases · occurred_at=2026-05-29T02:17:22+00:00 · url=https://huggingface.co/baidu/NAVA · hf_downloads=332 · hf_likes=108 · pipeline=text-to-video · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E8","kind":"event","title":"baidu/ERNIE-4.5-0.3B-PT","date":"2025-06-28T06:11:42+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-0.3B-PT","signal_url":"https://onlylabs.fyi/signals/5781c65a-ad63-4e79-932c-7b84b1bdb5c3","signal_json_url":"https://onlylabs.fyi/signals/5781c65a-ad63-4e79-932c-7b84b1bdb5c3/signal.json","text":"model_released · baidu/ERNIE-4.5-0.3B-PT · signal_desk=releases · occurred_at=2025-06-28T06:11:42+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-0.3B-PT · hf_downloads=26657 · hf_likes=111 · hf_params=360748032 · pipeline=text-generation · license=apache-2.0"},{"ref":"E9","kind":"event","title":"baidu/ERNIE-4.5-VL-28B-A3B-PT","date":"2025-06-28T05:50:33+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-PT","signal_url":"https://onlylabs.fyi/signals/0fd00e30-f9a5-4fb5-afda-b9d7010b538b","signal_json_url":"https://onlylabs.fyi/signals/0fd00e30-f9a5-4fb5-afda-b9d7010b538b/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-28B-A3B-PT · signal_desk=releases · occurred_at=2025-06-28T05:50:33+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-PT · hf_downloads=186476 · hf_likes=103 · hf_params=29399294336 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E10","kind":"event","title":"baidu/ERNIE-4.5-VL-424B-A47B-PT","date":"2025-06-28T05:49:55+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-PT","signal_url":"https://onlylabs.fyi/signals/dcc08054-f1dc-47ac-a90f-61bfb6ba3b29","signal_json_url":"https://onlylabs.fyi/signals/dcc08054-f1dc-47ac-a90f-61bfb6ba3b29/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-424B-A47B-PT · signal_desk=releases · occurred_at=2025-06-28T05:49:55+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-PT · hf_downloads=101 · hf_likes=106 · hf_params=423526285184 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E11","kind":"event","title":"baidu/ERNIE-4.5-VL-424B-A47B-Base-PT","date":"2025-06-28T05:49:36+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Base-PT","signal_url":"https://onlylabs.fyi/signals/e3dc0ffa-7613-4a00-a463-652388d393c7","signal_json_url":"https://onlylabs.fyi/signals/e3dc0ffa-7613-4a00-a463-652388d393c7/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-424B-A47B-Base-PT · signal_desk=releases · occurred_at=2025-06-28T05:49:36+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Base-PT · hf_downloads=137 · hf_likes=83 · hf_params=423526285184 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E12","kind":"event","title":"baidu/ERNIE-4.5-300B-A47B-PT","date":"2025-06-28T05:38:53+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-PT","signal_url":"https://onlylabs.fyi/signals/e835ceef-608b-41ed-8f81-c71937725063","signal_json_url":"https://onlylabs.fyi/signals/e835ceef-608b-41ed-8f81-c71937725063/signal.json","text":"model_released · baidu/ERNIE-4.5-300B-A47B-PT · signal_desk=releases · occurred_at=2025-06-28T05:38:53+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-PT · hf_downloads=1385 · hf_likes=78 · hf_params=300474051776 · pipeline=text-generation · license=apache-2.0"},{"ref":"E13","kind":"event","title":"baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle","date":"2025-06-28T16:05:05+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle","signal_url":"https://onlylabs.fyi/signals/d7d3a9ae-128e-443c-b81a-470891692230","signal_json_url":"https://onlylabs.fyi/signals/d7d3a9ae-128e-443c-b81a-470891692230/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle · signal_desk=releases · occurred_at=2025-06-28T16:05:05+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle · hf_downloads=33 · hf_likes=68 · hf_params=423526285184 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E14","kind":"event","title":"baidu/ERNIE-Image-Aes","date":"2026-05-18T09:12:17+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-Image-Aes","signal_url":"https://onlylabs.fyi/signals/b6335999-d77b-4243-abca-5368af446b9d","signal_json_url":"https://onlylabs.fyi/signals/b6335999-d77b-4243-abca-5368af446b9d/signal.json","text":"model_released · baidu/ERNIE-Image-Aes · signal_desk=releases · occurred_at=2026-05-18T09:12:17+00:00 · url=https://huggingface.co/baidu/ERNIE-Image-Aes · hf_downloads=4388 · hf_likes=11 · hf_params=7944373760 · license=apache-2.0"},{"ref":"E15","kind":"event","title":"baidu/ERNIE-4.5-300B-A47B-Base-PT","date":"2025-06-28T05:37:06+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-Base-PT","signal_url":"https://onlylabs.fyi/signals/3d2178c7-d1d2-4dd7-81cd-b48a5bce1f7d","signal_json_url":"https://onlylabs.fyi/signals/3d2178c7-d1d2-4dd7-81cd-b48a5bce1f7d/signal.json","text":"model_released · baidu/ERNIE-4.5-300B-A47B-Base-PT · signal_desk=releases · occurred_at=2025-06-28T05:37:06+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-Base-PT · hf_downloads=94 · hf_likes=42 · hf_params=299484163264 · pipeline=text-generation · license=apache-2.0"},{"ref":"E16","kind":"event","title":"baidu/ERNIE-4.5-VL-28B-A3B-Base-PT","date":"2025-06-28T07:24:07+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-Base-PT","signal_url":"https://onlylabs.fyi/signals/ce347a38-aed8-4fe6-86f1-2a67dc6c18ad","signal_json_url":"https://onlylabs.fyi/signals/ce347a38-aed8-4fe6-86f1-2a67dc6c18ad/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-28B-A3B-Base-PT · signal_desk=releases · occurred_at=2025-06-28T07:24:07+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-Base-PT · hf_downloads=73 · hf_likes=42 · hf_params=29399294336 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E17","kind":"event","title":"baidu/Qianfan-VL-8B","date":"2025-09-16T14:18:23+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/Qianfan-VL-8B","signal_url":"https://onlylabs.fyi/signals/be94b3ca-4ee6-42a3-82e9-456aa7b6243c","signal_json_url":"https://onlylabs.fyi/signals/be94b3ca-4ee6-42a3-82e9-456aa7b6243c/signal.json","text":"model_released · baidu/Qianfan-VL-8B · signal_desk=releases · occurred_at=2025-09-16T14:18:23+00:00 · url=https://huggingface.co/baidu/Qianfan-VL-8B · hf_downloads=7707 · hf_likes=41 · hf_params=8808322048 · pipeline=image-text-to-text · license=other"},{"ref":"E18","kind":"event","title":"baidu/Qianfan-VL-70B","date":"2025-09-16T14:18:26+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/Qianfan-VL-70B","signal_url":"https://onlylabs.fyi/signals/e628a6da-2cbb-41fa-a430-47a0039e12b9","signal_json_url":"https://onlylabs.fyi/signals/e628a6da-2cbb-41fa-a430-47a0039e12b9/signal.json","text":"model_released · baidu/Qianfan-VL-70B · signal_desk=releases · occurred_at=2025-09-16T14:18:26+00:00 · url=https://huggingface.co/baidu/Qianfan-VL-70B · hf_downloads=47 · hf_likes=39 · hf_params=71839363072 · pipeline=image-text-to-text · license=mit"},{"ref":"E19","kind":"event","title":"Advertising Sales Associate (Contractor)","date":"2026-05-30T00:56:12+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7826971","signal_url":"https://onlylabs.fyi/signals/80d2f6b4-69fa-41c8-8744-a41e57d469e3","signal_json_url":"https://onlylabs.fyi/signals/80d2f6b4-69fa-41c8-8744-a41e57d469e3/signal.json","text":"job_opened · Advertising Sales Associate (Contractor) · signal_desk=hiring · occurred_at=2026-05-30T00:56:12+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7826971 · data_radar_lanes=Product and customer · data_radar_terms=sales · data_radar_reason=Baidu (ERNIE) has a job signal matching product and customer. · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E20","kind":"event","title":"GTM Strategy & Operations (AI Desktop & Mobile App)","date":"2026-05-28T18:53:51+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7964093","signal_url":"https://onlylabs.fyi/signals/32a08b6f-001e-4ec0-ae3a-d4b87cd8f200","signal_json_url":"https://onlylabs.fyi/signals/32a08b6f-001e-4ec0-ae3a-d4b87cd8f200/signal.json","text":"job_opened · GTM Strategy & Operations (AI Desktop & Mobile App) · signal_desk=hiring · occurred_at=2026-05-28T18:53:51+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7964093 · data_radar_lanes=Product and customer · data_radar_terms=gtm · data_radar_reason=Baidu (ERNIE) has a job signal matching product and customer. · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E21","kind":"event","title":"PaddlePaddle/PaddleOCR v3.6.0","date":"2026-05-28T11:42:29+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/PaddleOCR/releases/tag/v3.6.0","signal_url":"https://onlylabs.fyi/signals/fef9275e-0e82-41f1-bba1-2435d0ea78fe","signal_json_url":"https://onlylabs.fyi/signals/fef9275e-0e82-41f1-bba1-2435d0ea78fe/signal.json","text":"release · PaddlePaddle/PaddleOCR v3.6.0 · signal_desk=releases · occurred_at=2026-05-28T11:42:29+00:00 · url=https://github.com/PaddlePaddle/PaddleOCR/releases/tag/v3.6.0 · raw={\"repo\":\"PaddlePaddle/PaddleOCR\"}"},{"ref":"E22","kind":"event","title":"PaddlePaddle/PaddleX v3.6.1","date":"2026-05-27T14:36:14+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.6.1","signal_url":"https://onlylabs.fyi/signals/fa062841-8907-4a3a-a678-e88ec59bd01f","signal_json_url":"https://onlylabs.fyi/signals/fa062841-8907-4a3a-a678-e88ec59bd01f/signal.json","text":"release · PaddlePaddle/PaddleX v3.6.1 · signal_desk=releases · occurred_at=2026-05-27T14:36:14+00:00 · url=https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.6.1 · raw={\"repo\":\"PaddlePaddle/PaddleX\"}"},{"ref":"E23","kind":"event","title":"PaddlePaddle/PaddleX v3.6.0","date":"2026-05-27T13:51:59+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.6.0","signal_url":"https://onlylabs.fyi/signals/f3a247db-4dde-4b63-87a2-81a1074042c0","signal_json_url":"https://onlylabs.fyi/signals/f3a247db-4dde-4b63-87a2-81a1074042c0/signal.json","text":"release · PaddlePaddle/PaddleX v3.6.0 · signal_desk=releases · occurred_at=2026-05-27T13:51:59+00:00 · url=https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.6.0 · raw={\"repo\":\"PaddlePaddle/PaddleX\"}"},{"ref":"E24","kind":"event","title":"Senior Business Development & Partnerships Manager","date":"2026-05-26T19:03:13+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7952151","signal_url":"https://onlylabs.fyi/signals/bdca6ce7-fc13-4e2a-baaf-a696fd3c7e12","signal_json_url":"https://onlylabs.fyi/signals/bdca6ce7-fc13-4e2a-baaf-a696fd3c7e12/signal.json","text":"job_opened · Senior Business Development & Partnerships Manager · signal_desk=hiring · occurred_at=2026-05-26T19:03:13+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7952151 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E25","kind":"event","title":"baidu/ERNIE-4.5-21B-A3B-Base-PT","date":"2025-06-28T06:12:54+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Base-PT","signal_url":"https://onlylabs.fyi/signals/79bc879c-352f-4ecb-9420-92ea7e548a17","signal_json_url":"https://onlylabs.fyi/signals/79bc879c-352f-4ecb-9420-92ea7e548a17/signal.json","text":"model_released · baidu/ERNIE-4.5-21B-A3B-Base-PT · signal_desk=releases · occurred_at=2025-06-28T06:12:54+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Base-PT · hf_downloads=438 · hf_likes=36 · hf_params=21825437888 · pipeline=text-generation · license=apache-2.0"},{"ref":"E26","kind":"event","title":"Account Manager","date":"2026-05-21T21:53:11+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7951123","signal_url":"https://onlylabs.fyi/signals/002a8e51-386e-491a-b818-9a2992d9e39b","signal_json_url":"https://onlylabs.fyi/signals/002a8e51-386e-491a-b818-9a2992d9e39b/signal.json","text":"job_opened · Account Manager · signal_desk=hiring · occurred_at=2026-05-21T21:53:11+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7951123 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E27","kind":"event","title":"PaddlePaddle/Paddle-Metax","date":"2026-05-14T07:06:27+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/Paddle-Metax","signal_url":"https://onlylabs.fyi/signals/19133a46-a10c-4d39-bf33-e17104e0737b","signal_json_url":"https://onlylabs.fyi/signals/19133a46-a10c-4d39-bf33-e17104e0737b/signal.json","text":"repo_new · PaddlePaddle/Paddle-Metax · signal_desk=repos · occurred_at=2026-05-14T07:06:27+00:00 · url=https://github.com/PaddlePaddle/Paddle-Metax · stars=1 · raw={\"repo\":\"PaddlePaddle/Paddle-Metax\",\"language\":\"Python\"}"},{"ref":"E28","kind":"event","title":"PaddlePaddle/PaddleX v3.5.2","date":"2026-05-11T04:56:45+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.5.2","signal_url":"https://onlylabs.fyi/signals/f557ec7a-087a-451a-b48a-acb237277400","signal_json_url":"https://onlylabs.fyi/signals/f557ec7a-087a-451a-b48a-acb237277400/signal.json","text":"release · PaddlePaddle/PaddleX v3.5.2 · signal_desk=releases · occurred_at=2026-05-11T04:56:45+00:00 · url=https://github.com/PaddlePaddle/PaddleX/releases/tag/v3.5.2 · raw={\"repo\":\"PaddlePaddle/PaddleX\"}"},{"ref":"E29","kind":"event","title":"ERNIE 5.1 Officially Released! Topping Multiple Leaderboards — A Model That Writes Better and Understands You More","date":"2026-05-09T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.1-0508-release/","signal_url":"https://onlylabs.fyi/signals/7756917e-15b8-4913-a84f-118b03151d58","signal_json_url":"https://onlylabs.fyi/signals/7756917e-15b8-4913-a84f-118b03151d58/signal.json","text":"post_published · ERNIE 5.1 Officially Released! Topping Multiple Leaderboards — A Model That Writes Better and Understands You More · signal_desk=talking · occurred_at=2026-05-09T00:00:00+00:00 · url=/blog/posts/ernie-5.1-0508-release/ · data_radar_lanes=Infrastructure · data_radar_terms=training · data_radar_reason=Baidu (ERNIE) has a writing signal matching infrastructure. · raw={\"excerpt\":\"ERNIE 5.1 is officially released, achieving leading performance at only 6% of the pre-training cost of comparable models. Powered by disaggregated fully-asynchronous reinforcement learning and scaled agentic post-training, ERNIE 5.1 delivers comprehensive upgrades across Agent, reasoning, and creative capabilities, ranking 1st in China on the Arena Search Arena.\"}"},{"ref":"E30","kind":"event","title":"baidu/Qianfan-VL-3B","date":"2025-09-16T14:18:12+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/Qianfan-VL-3B","signal_url":"https://onlylabs.fyi/signals/5523cd05-a017-4d88-a0b9-0030429b281c","signal_json_url":"https://onlylabs.fyi/signals/5523cd05-a017-4d88-a0b9-0030429b281c/signal.json","text":"model_released · baidu/Qianfan-VL-3B · signal_desk=releases · occurred_at=2025-09-16T14:18:12+00:00 · url=https://huggingface.co/baidu/Qianfan-VL-3B · hf_downloads=114 · hf_likes=28 · hf_params=3713711104 · pipeline=image-text-to-text · license=other"},{"ref":"E31","kind":"event","title":"baidu/ERNIE-4.5-0.3B-Paddle","date":"2025-06-29T07:24:14+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-0.3B-Paddle","signal_url":"https://onlylabs.fyi/signals/d58e262d-dfc9-4d0b-a0b6-86ecef2aa7da","signal_json_url":"https://onlylabs.fyi/signals/d58e262d-dfc9-4d0b-a0b6-86ecef2aa7da/signal.json","text":"model_released · baidu/ERNIE-4.5-0.3B-Paddle · signal_desk=releases · occurred_at=2025-06-29T07:24:14+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-0.3B-Paddle · hf_downloads=107 · hf_likes=28 · hf_params=360748032 · pipeline=text-generation · license=apache-2.0"},{"ref":"E32","kind":"event","title":"baidu/ERNIE-4.5-0.3B-Base-PT","date":"2025-06-28T06:10:12+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-0.3B-Base-PT","signal_url":"https://onlylabs.fyi/signals/77483efc-e0c5-405f-bf7d-d6462ce52105","signal_json_url":"https://onlylabs.fyi/signals/77483efc-e0c5-405f-bf7d-d6462ce52105/signal.json","text":"model_released · baidu/ERNIE-4.5-0.3B-Base-PT · signal_desk=releases · occurred_at=2025-06-28T06:10:12+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-0.3B-Base-PT · hf_downloads=1308 · hf_likes=27 · hf_params=360748032 · pipeline=text-generation · license=apache-2.0"},{"ref":"E33","kind":"event","title":"baidu/ERNIE-4.5-VL-424B-A47B-Paddle","date":"2025-06-28T15:56:13+00:00","date_source":"source","source_url":"https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Paddle","signal_url":"https://onlylabs.fyi/signals/dbed864e-42d5-4d3d-864b-e96e9db5c19e","signal_json_url":"https://onlylabs.fyi/signals/dbed864e-42d5-4d3d-864b-e96e9db5c19e/signal.json","text":"model_released · baidu/ERNIE-4.5-VL-424B-A47B-Paddle · signal_desk=releases · occurred_at=2025-06-28T15:56:13+00:00 · url=https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-Paddle · hf_downloads=34 · hf_likes=27 · hf_params=423526285184 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E34","kind":"event","title":"Business Development & Partnerships Specialist","date":"2026-05-04T18:26:39+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7521178","signal_url":"https://onlylabs.fyi/signals/859e3e5c-d4a0-4550-a95e-82809b7b86ac","signal_json_url":"https://onlylabs.fyi/signals/859e3e5c-d4a0-4550-a95e-82809b7b86ac/signal.json","text":"job_opened · Business Development & Partnerships Specialist · signal_desk=hiring · occurred_at=2026-05-04T18:26:39+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7521178 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E35","kind":"event","title":"ERNIE-5.1-Preview Tops LMArena Text Leaderboard as No.1 Chinese Model!","date":"2026-04-30T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.1-preview-0430-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/431263ef-3d52-497c-934d-732c0d42bb52","signal_json_url":"https://onlylabs.fyi/signals/431263ef-3d52-497c-934d-732c0d42bb52/signal.json","text":"post_published · ERNIE-5.1-Preview Tops LMArena Text Leaderboard as No.1 Chinese Model! · signal_desk=talking · occurred_at=2026-04-30T00:00:00+00:00 · url=/blog/posts/ernie-5.1-preview-0430-release-on-lmarena/ · raw={\"excerpt\":\"On April 30, LMArena released its latest rankings. ERNIE-5.1-Preview ranked No. 1 among Chinese models and No. 13 globally on the LMArena Text Arena, placing in the global top 10 across multiple category leaderboards.\"}"},{"ref":"E36","kind":"event","title":"Advertising Sales Manager","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7392194","signal_url":"https://onlylabs.fyi/signals/82aa0bae-bdc3-4004-b28f-0338b2301284","signal_json_url":"https://onlylabs.fyi/signals/82aa0bae-bdc3-4004-b28f-0338b2301284/signal.json","text":"job_opened · Advertising Sales Manager · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7392194 · data_radar_lanes=Product and customer · data_radar_terms=sales · data_radar_reason=Baidu (ERNIE) has a job signal matching product and customer. · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E37","kind":"event","title":"Account Manager","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7767326","signal_url":"https://onlylabs.fyi/signals/1dafc866-ad9e-4ba6-a750-e512c5988d49","signal_json_url":"https://onlylabs.fyi/signals/1dafc866-ad9e-4ba6-a750-e512c5988d49/signal.json","text":"job_opened · Account Manager · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7767326 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E38","kind":"event","title":"Machine Learning System Hardware Architect","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/3226162","signal_url":"https://onlylabs.fyi/signals/58425602-a9d1-4ef1-b176-f5e1eb8866bb","signal_json_url":"https://onlylabs.fyi/signals/58425602-a9d1-4ef1-b176-f5e1eb8866bb/signal.json","text":"job_opened · Machine Learning System Hardware Architect · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/3226162 · raw={\"location\":\"Sunnyvale,CA\",\"ats\":\"greenhouse\"}"},{"ref":"E39","kind":"event","title":"Associate Campaign Manager ","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7521209","signal_url":"https://onlylabs.fyi/signals/80062b19-b3b1-41b1-be0d-83ecac304b0f","signal_json_url":"https://onlylabs.fyi/signals/80062b19-b3b1-41b1-be0d-83ecac304b0f/signal.json","text":"job_opened · Associate Campaign Manager  · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7521209 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E40","kind":"event","title":"CPU Digital Design Engineer","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/4499622","signal_url":"https://onlylabs.fyi/signals/56b43e55-5de8-426e-9ff1-8e64a2b354f9","signal_json_url":"https://onlylabs.fyi/signals/56b43e55-5de8-426e-9ff1-8e64a2b354f9/signal.json","text":"job_opened · CPU Digital Design Engineer · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/4499622 · raw={\"location\":\"Sunnyvale, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E41","kind":"event","title":"CPU/GPU/Processor Hardware Architect","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/3226157","signal_url":"https://onlylabs.fyi/signals/bd68b609-c50f-4128-98f9-75f6796c7b11","signal_json_url":"https://onlylabs.fyi/signals/bd68b609-c50f-4128-98f9-75f6796c7b11/signal.json","text":"job_opened · CPU/GPU/Processor Hardware Architect · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/3226157 · data_radar_lanes=Infrastructure · data_radar_terms=gpu · data_radar_reason=Baidu (ERNIE) has a job signal matching infrastructure. · raw={\"location\":\"Sunnyvale,CA\",\"ats\":\"greenhouse\"}"},{"ref":"E42","kind":"event","title":"Design Verification Engineer","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/3234390","signal_url":"https://onlylabs.fyi/signals/a423d2f5-3690-4d5f-a7ca-684b59d69cd4","signal_json_url":"https://onlylabs.fyi/signals/a423d2f5-3690-4d5f-a7ca-684b59d69cd4/signal.json","text":"job_opened · Design Verification Engineer · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/3234390 · raw={\"location\":\"Sunnyvale,CA\",\"ats\":\"greenhouse\"}"},{"ref":"E43","kind":"event","title":"Senior RTL Design Engineer","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/4499691","signal_url":"https://onlylabs.fyi/signals/864deb78-15c2-4838-bbbf-8e36c00efe23","signal_json_url":"https://onlylabs.fyi/signals/864deb78-15c2-4838-bbbf-8e36c00efe23/signal.json","text":"job_opened · Senior RTL Design Engineer · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/4499691 · raw={\"location\":\"Sunnyvale, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E44","kind":"event","title":"Senior Manager - Global Business Unit, Baidu USA","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7689681","signal_url":"https://onlylabs.fyi/signals/e2d97a44-f9cb-46be-b1a0-cfc1f6d5b1ed","signal_json_url":"https://onlylabs.fyi/signals/e2d97a44-f9cb-46be-b1a0-cfc1f6d5b1ed/signal.json","text":"job_opened · Senior Manager - Global Business Unit, Baidu USA · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7689681 · raw={\"location\":\"Mountain View, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E45","kind":"event","title":"SoC Memory Subsystem Architect","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/3234364","signal_url":"https://onlylabs.fyi/signals/bf824b10-0491-41d0-b967-db1640f25825","signal_json_url":"https://onlylabs.fyi/signals/bf824b10-0491-41d0-b967-db1640f25825/signal.json","text":"job_opened · SoC Memory Subsystem Architect · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/3234364 · raw={\"location\":\"Sunnyvale,CA\",\"ats\":\"greenhouse\"}"},{"ref":"E46","kind":"event","title":"SoC System Hardware Architect","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/3234382","signal_url":"https://onlylabs.fyi/signals/bf1124f4-30b1-4e00-b313-5b13f37270e3","signal_json_url":"https://onlylabs.fyi/signals/bf1124f4-30b1-4e00-b313-5b13f37270e3/signal.json","text":"job_opened · SoC System Hardware Architect · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/3234382 · raw={\"location\":\"Sunnyvale,CA\",\"ats\":\"greenhouse\"}"},{"ref":"E47","kind":"event","title":"CPU Design Architect / Principal Digital Design Engineer","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/4499671","signal_url":"https://onlylabs.fyi/signals/69c8ec4d-4dbd-4e0a-8d79-f9c1ef36ab09","signal_json_url":"https://onlylabs.fyi/signals/69c8ec4d-4dbd-4e0a-8d79-f9c1ef36ab09/signal.json","text":"job_opened · CPU Design Architect / Principal Digital Design Engineer · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/4499671 · raw={\"location\":\"Sunnyvale, CA\",\"ats\":\"greenhouse\"}"},{"ref":"E48","kind":"event","title":"Client Manager - Canada","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7521232","signal_url":"https://onlylabs.fyi/signals/3f890239-4908-43c9-97ce-482502e91d6e","signal_json_url":"https://onlylabs.fyi/signals/3f890239-4908-43c9-97ce-482502e91d6e/signal.json","text":"job_opened · Client Manager - Canada · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7521232 · raw={\"location\":\"Toronto, ON\",\"ats\":\"greenhouse\"}"},{"ref":"E49","kind":"event","title":"Business Development & Partnerships Specialist - Canada","date":"2026-04-28T18:27:49+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/baidu/jobs/7521228","signal_url":"https://onlylabs.fyi/signals/ac2660aa-a164-4c06-8935-5d62d83895b5","signal_json_url":"https://onlylabs.fyi/signals/ac2660aa-a164-4c06-8935-5d62d83895b5/signal.json","text":"job_opened · Business Development & Partnerships Specialist - Canada · signal_desk=hiring · occurred_at=2026-04-28T18:27:49+00:00 · url=https://job-boards.greenhouse.io/baidu/jobs/7521228 · raw={\"location\":\"Toronto, ON\",\"ats\":\"greenhouse\"}"},{"ref":"E50","kind":"event","title":"Introducing ERNIE-Image","date":"2026-04-15T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-image/","signal_url":"https://onlylabs.fyi/signals/0a3da723-711b-483f-a36d-5b99a5ee6507","signal_json_url":"https://onlylabs.fyi/signals/0a3da723-711b-483f-a36d-5b99a5ee6507/signal.json","text":"post_published · Introducing ERNIE-Image · signal_desk=talking · occurred_at=2026-04-15T00:00:00+00:00 · url=/blog/posts/ernie-image/ · raw={\"excerpt\":\"ERNIE-Image is a text-to-image generation model built on a single-stream Diffusion Transformer (DiT) with 8B DiT parameters, achieving leading performance among open-weights models.\"}"},{"ref":"E51","kind":"event","title":"ERNIE 5.0: A 2.4 Trillion-Parameter Unified Multimodal Foundation Model","date":"2026-02-06T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie5.0/","signal_url":"https://onlylabs.fyi/signals/3dd0d29e-ba0e-44be-8612-2cba4d88c1a5","signal_json_url":"https://onlylabs.fyi/signals/3dd0d29e-ba0e-44be-8612-2cba4d88c1a5/signal.json","text":"post_published · ERNIE 5.0: A 2.4 Trillion-Parameter Unified Multimodal Foundation Model · signal_desk=talking · occurred_at=2026-02-06T00:00:00+00:00 · url=/blog/posts/ernie5.0/ · raw={\"excerpt\":\"We introduce ERNIE 5.0: a 2.4 trillion-parameter Unified Multimodal Model trained from scratch. Integrating text, image, video, and audio into a single autoregressive framework, it overcomes the limitations of late-fusion architectures to achieve seamless cross-modal understanding and generation.\"}"},{"ref":"E52","kind":"event","title":"PaddleOCR-VL-1.5: Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing","date":"2026-01-29T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/paddleocr-vl-1.5/","signal_url":"https://onlylabs.fyi/signals/df8beb77-39a4-4f31-b8d9-36800a4c2602","signal_json_url":"https://onlylabs.fyi/signals/df8beb77-39a4-4f31-b8d9-36800a4c2602/signal.json","text":"post_published · PaddleOCR-VL-1.5: Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing · signal_desk=talking · occurred_at=2026-01-29T00:00:00+00:00 · url=/blog/posts/paddleocr-vl-1.5/ · raw={\"excerpt\":\"🚀 We release PaddleOCR-VL-1.5, an upgraded model achieving a new state-of-the-art (SOTA)accuracy of 94.5% on OmniDocBench v1.5.\"}"},{"ref":"E53","kind":"event","title":"ERNIE-5.0 Tops LMArena Text Leaderboard as No.1 Chinese Model!","date":"2026-01-15T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-0110-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/d120fc07-69ea-45b9-9a6d-2ed473b59be8","signal_json_url":"https://onlylabs.fyi/signals/d120fc07-69ea-45b9-9a6d-2ed473b59be8/signal.json","text":"post_published · ERNIE-5.0 Tops LMArena Text Leaderboard as No.1 Chinese Model! · signal_desk=talking · occurred_at=2026-01-15T00:00:00+00:00 · url=/blog/posts/ernie-5.0-0110-release-on-lmarena/ · raw={\"excerpt\":\"On January 15, LMArena released its latest rankings. ERNIE-5.0-0110 achieved a score of 1,460, ranking No. 1 among Chinese models and No. 8 globally on the LMArena Text Arena.\"}"},{"ref":"E54","kind":"event","title":"ERNIE-5.0-Preview-1220 Becomes the Sole Chinese Model in LMArena Vision Top 10!","date":"2026-01-08T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-preview-1220-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/e6540cd7-b324-44a1-8f69-e6e60666db7c","signal_json_url":"https://onlylabs.fyi/signals/e6540cd7-b324-44a1-8f69-e6e60666db7c/signal.json","text":"post_published · ERNIE-5.0-Preview-1220 Becomes the Sole Chinese Model in LMArena Vision Top 10! · signal_desk=talking · occurred_at=2026-01-08T00:00:00+00:00 · url=/blog/posts/ernie-5.0-preview-1220-release-on-lmarena/ · raw={\"excerpt\":\"On January 8, LMArena released its latest rankings. ERNIE-5.0-Preview-1220 achieved a score of 1226, ranking No. 1 in China and No. 8 globally on the LMArena Vision Arena.\"}"},{"ref":"E55","kind":"event","title":"Best Text model from China in LMArena is now ERNIE-5.0-Preview-1203!","date":"2025-12-23T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-preview-1203-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/6a8af492-922e-4c7f-9aca-c28e326d2b62","signal_json_url":"https://onlylabs.fyi/signals/6a8af492-922e-4c7f-9aca-c28e326d2b62/signal.json","text":"post_published · Best Text model from China in LMArena is now ERNIE-5.0-Preview-1203! · signal_desk=talking · occurred_at=2025-12-23T00:00:00+00:00 · url=/blog/posts/ernie-5.0-preview-1203-release-on-lmarena/ · raw={\"excerpt\":\"Just now, LMArena released its latest rankings. Baidu’s ERNIE-5.0-Preview-1203 scored an impressive 1,451 points.\"}"},{"ref":"E56","kind":"event","title":"ERNIE-5.0-Preview-1103 landed on the LMArena Text Leaderboard!","date":"2025-12-09T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-preview-1103-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/69871318-5bad-4c78-9db8-c032aba97d51","signal_json_url":"https://onlylabs.fyi/signals/69871318-5bad-4c78-9db8-c032aba97d51/signal.json","text":"post_published · ERNIE-5.0-Preview-1103 landed on the LMArena Text Leaderboard! · signal_desk=talking · occurred_at=2025-12-09T00:00:00+00:00 · url=/blog/posts/ernie-5.0-preview-1103-release-on-lmarena/ · raw={\"excerpt\":\"We’ve just refreshed our standings with the latest ERNIE-5.0-Preview-1103 on LMArena. 🚀 ERNIE-5.0-Preview-1103 holds the top 20 in the most competitive Arena.\"}"},{"ref":"E57","kind":"event","title":"ERNIE-5.0-Preview-1120, ready for testing in LMArena!","date":"2025-11-21T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-preview-1120-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/43f7f638-4810-4729-b12a-79b8b8a0f88f","signal_json_url":"https://onlylabs.fyi/signals/43f7f638-4810-4729-b12a-79b8b8a0f88f/signal.json","text":"post_published · ERNIE-5.0-Preview-1120, ready for testing in LMArena! · signal_desk=talking · occurred_at=2025-11-21T00:00:00+00:00 · url=/blog/posts/ernie-5.0-preview-1120-release-on-lmarena/ · data_radar_lanes=Evals and quality · data_radar_terms=testing · data_radar_reason=Baidu (ERNIE) has a writing signal matching evals and quality. · raw={\"excerpt\":\"ERNIE-5.0-Preview-1120 now ranks #1 in domestic on the LMArena Vision leaderboard\"}"},{"ref":"E58","kind":"event","title":"ERNIE-4.5-VL-28B-A3B-Thinking: A Breakthrough in Multimodal AI","date":"2025-11-11T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-4.5-vl-28b-a3b-thinking/","signal_url":"https://onlylabs.fyi/signals/6e0a9115-3185-4f50-bde0-b8e017e09897","signal_json_url":"https://onlylabs.fyi/signals/6e0a9115-3185-4f50-bde0-b8e017e09897/signal.json","text":"post_published · ERNIE-4.5-VL-28B-A3B-Thinking: A Breakthrough in Multimodal AI · signal_desk=talking · occurred_at=2025-11-11T00:00:00+00:00 · url=/blog/posts/ernie-4.5-vl-28b-a3b-thinking/ · raw={\"excerpt\":\"We release ERNIE-4.5-VL-28B-A3B-Thinking, a multimodal reasoning model that achieves SOTA performance while activating only 3B parameters.\"}"},{"ref":"E59","kind":"event","title":"ERNIE-5.0-Preview-1022, ready for testing in LMArena!","date":"2025-11-07T00:00:00+00:00","date_source":"rss.item_date","source_url":"/blog/posts/ernie-5.0-preview-1022-release-on-lmarena/","signal_url":"https://onlylabs.fyi/signals/c4feae91-ed43-4242-9377-c3ef92ec574c","signal_json_url":"https://onlylabs.fyi/signals/c4feae91-ed43-4242-9377-c3ef92ec574c/signal.json","text":"post_published · ERNIE-5.0-Preview-1022, ready for testing in LMArena! · signal_desk=talking · occurred_at=2025-11-07T00:00:00+00:00 · url=/blog/posts/ernie-5.0-preview-1022-release-on-lmarena/ · data_radar_lanes=Evals and quality · data_radar_terms=testing · data_radar_reason=Baidu (ERNIE) has a writing signal matching evals and quality. · raw={\"excerpt\":\"ERNIE-5.0-Preview-1022 now ranks #2 globally on the LMArena Text leaderboard\"}"},{"ref":"E60","kind":"event","title":"PaddlePaddle/flux","date":"2024-12-13T06:45:26+00:00","date_source":"source","source_url":"https://github.com/PaddlePaddle/flux","signal_url":"https://onlylabs.fyi/signals/ee2e8c12-b455-4f27-84ad-01ae685f77c8","signal_json_url":"https://onlylabs.fyi/signals/ee2e8c12-b455-4f27-84ad-01ae685f77c8/signal.json","text":"repo_forked · PaddlePaddle/flux · signal_desk=forks · occurred_at=2024-12-13T06:45:26+00:00 · url=https://github.com/PaddlePaddle/flux · raw={\"repo\":\"PaddlePaddle/flux\",\"parent\":\"bytedance/flux\"}"}]}