Merge branch 'PaddlePaddle:develop' into develop

Author: wtmlon

.github/workflows/lint.yml

@@ -0,0 +1,40 @@
+name: Codestyle Check
+
+on: [push, pull_request]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
+  cancel-in-progress: true
+
+jobs:
+  Lint:
+    name: Lint
+    runs-on: ubuntu-latest
+    steps:
+      - name: Download Code
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Checkout Develop
+        run: |
+          if ! git show-ref --quiet refs/heads/develop; then \
+              echo "local develop branch is missing, creating local develop branch that tracks remote develop branch"
+              git fetch origin develop
+              git branch develop --track origin/develop
+          else
+            echo "local develop branch exist, skipping"
+          fi
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.10'
+          cache: 'pip' # caching pip dependencies
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install pre-commit
+          pre-commit install
+      - name: Test
+        run: make -f tools/codestyle/Makefile lint

.github/workflows/test.yml

@@ -0,0 +1,37 @@
+name: Test
+
+on: [push, pull_request]
+
+jobs:
+  Test:
+    name: Test
+    runs-on: [self-hosted, 4gpu]
+    steps:
+      - name: Start Docker 
+        run: |
+          if [ ! "$(docker ps -q -f name=ERNIE_CI_Container)" ]; then
+            echo "Starting container ERNIE_CI_Container"
+            docker start ERNIE_CI_Container
+          else
+            echo "ERNIE_CI_Container already running"
+          fi
+      - name: Cleanup Environment Cache
+        run: docker exec -t ERNIE_CI_Container /bin/bash -c '
+          rm -rf /workspace/* /workspace/.[!.]* /workspace/..?*'
+      - name: Get Repo
+        env:
+          ACTIONS_RUNNER_FORCED_INTERNAL_NODE_VERSION: node16
+          ACTIONS_RUNNER_FORCE_ACTIONS_NODE_VERSION: node16
+          ACTIONS_ALLOW_UNSECURE_NODE_VERSION: true
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+      - name: Prepare Model
+        run: docker exec -t ERNIE_CI_Container /bin/bash -c '
+            cp -r /home/ERNIE_Model/* /workspace/ '
+      - name: Install erniekit
+        run: docker exec -t ERNIE_CI_Container /bin/bash -c '
+          pip install -e.'
+      - name: Runing case
+        run: docker exec -t ERNIE_CI_Container /bin/bash -c '
+          make test'

examples/pre-training/.gitignore .gitignoreexamples/pre-training/.gitignore renamed to .gitignore

@@ -0,0 +1,47 @@
+repos:
+# For Python files
+-   repo: https://github.com/psf/black.git
+    rev: 25.1.0
+    hooks:
+    -   id: black
+        files: \.(py|pyi)$
+        additional_dependencies: [toml]
+-   repo: https://github.com/PyCQA/isort
+    rev: 6.0.1
+    hooks:
+    -   id: isort
+-   repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.12.1
+    hooks:
+    -   id: ruff
+        args: [--fix, --exit-non-zero-on-fix]
+    -   id: ruff-format
+-   repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v5.0.0
+    hooks:
+    -   id: check-merge-conflict
+    -   id: check-symlinks
+    -   id: detect-private-key
+        files: (?!.*paddle)^.*$
+    -   id: end-of-file-fixer
+        files: \.md$
+    -   id: trailing-whitespace
+        files: \.md$
+-   repo: https://github.com/Lucas-C/pre-commit-hooks
+    rev: v1.5.5
+    hooks:
+    -   id: forbid-crlf
+        files: \.md$
+    -   id: remove-crlf
+        files: \.md$
+    -   id: forbid-tabs
+        files: \.md$
+    -   id: remove-tabs
+        files: \.md$
+-   repo: local
+    hooks:
+    -   id: copyright_checker
+        name: copyright_checker
+        entry: python ./tools/codestyle/.copyright.hook
+        language: system
+        files: \.(c|cc|cxx|cpp|cu|h|hpp|hxx|proto|xpu|kps|py|sh)$

.pre-commit-config.yaml

@@ -0,0 +1,33 @@
+# Makefile For ERNIE
+# GitHb: https://github.com/PaddlePaddle/ERNIE
+
+
+.PHONY: all
+all : lint test
+check_dirs := cookbook data_processor ernie erniekit examples tools test requirements
+
+# # # # # # # # # # # # # # # Lint Block # # # # # # # # # # # # # # #
+.PHONY: lint
+lint:
+	$(eval modified_py_files := $(shell python tests/codestyle/get_modified_files.py $(check_dirs)))
+	@if test -n "$(modified_py_files)"; then \
+		echo ${modified_py_files}; \
+		pre-commit run --files ${modified_py_files}; \
+	else \
+		echo "No library .py files were modified"; \
+	fi
+
+# # # # # # # # # # # # # # # Install Requirements Block # # # # # # # # # # # # # # #
+.PHONY: install
+install:
+	pip uninstall paddlepaddle-gpu -y
+	pip install paddlepaddle-gpu==3.1.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/
+	pip install -r requirements/gpu/requirements.txt
+	pip install pytest
+	pip install allure-pytest
+	pip install -e.
+
+# # # # # # # # # # # # # # # Test Block # # # # # # # # # # # # # # #
+.PHONY: test
+test:
+	PYTHONPATH=$(shell pwd) pytest -s -v --alluredir=result tests/

Makefile

@@ -6,7 +6,9 @@
 
 [ERNIE Bot](https://ernie.baidu.com/) |  [🤗Hugging Face](https://huggingface.co/baidu) | [AI Studio](https://aistudio.baidu.com/modelsoverview) 
 
-📑 [Blog](https://yiyan.baidu.com/blog/posts/ernie4.5) | 📚 [Cookbook](./cookbook/) | 📑 [Paper](https://yiyan.baidu.com/blog/publication/)  | 🛠️ [Toolkit](./docs/erniekit.md)
+📑 [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)
+
+<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>
 
 </div>
 
@@ -102,7 +104,7 @@ ERNIE-4.5-300B-A47B-Base surpasses DeepSeek-V3-671B-A37B-Base on 22 out of 28 be
 
 ERNIE-4.5-300B-A47B, the post trained model, demonstrates significant strengths in instruction following and knowledge tasks, as evidenced by the state-of-the-art scores on benchmarks such as IFEval, Multi-IF, SimpleQA, and ChineseSimpleQA. The lightweight model ERNIE-4.5-21B-A3B achieves competitive performance compared to Qwen3-30B-A3B, despite having approximately 30% fewer total parameters.
 
-In the non-thinking mode, ERNIE-4.5-VL exhibits outstanding proficiency in visual perception, document and chart understanding, and visual knowledge, performing strongly across a range of established benchmarks. Under the thinking mode, ERNIE-4.5-VL not only demonstrates enhanced reasoning abilities compared to the non-thinking mode, but also retains the strong perception capabilities of the latter. ERNIE-4.5-VL-424B-A47B delivers consistently strong results across the various multimodal evaluation benchmarks. Its thinking mode offers a distinct advantage on challenging benchmarks such as MathVista, MMMU, and VisualPuzzle, while maintaining competitive performance on perception-focused datasets like CV-Bench and RealWorldQA. The lightweight vision-language model ERNIE-4.5-28B-A3B achieves competitive or even superior performance compared to Qwen2.5-VL-7B and Qwen2.5-VL-32B across most benchmarks, despite using significantly fewer activation parameters. Notably, our lightweight model also supports both thinking and non-thinking modes, offering functionalities consistent with ERNIE-4.5-VL-424B-A47B.
+In the non-thinking mode, ERNIE-4.5-VL exhibits outstanding proficiency in visual perception, document and chart understanding, and visual knowledge, performing strongly across a range of established benchmarks. Under the thinking mode, ERNIE-4.5-VL not only demonstrates enhanced reasoning abilities compared to the non-thinking mode, but also retains the strong perception capabilities of the latter. ERNIE-4.5-VL-424B-A47B delivers consistently strong results across the full multimodal evaluation suite. Its thinking mode provides a distinct advantage on reasoning-centric tasks, narrowing or even surpassing the gap to OpenAI-o1 on challenging benchmarks such as MathVista, MMMU, and VisualPuzzle, while maintaining competitive performance on perception-focused datasets like CV-Bench and RealWorldQA. The lightweight vision-language model ERNIE-4.5-VL-28B-A3B achieves competitive or even superior performance compared to Qwen2.5-VL-7B and Qwen2.5-VL-32B across most benchmarks, despite using significantly fewer activation parameters. Notably, our lightweight model also supports both thinking and non-thinking modes, offering functionalities consistent with ERNIE-4.5-VL-424B-A47B.
 
 ### Performace of ERNIE-4.5 pre-trained models
 
@@ -244,6 +246,12 @@ Discover best-practice guides showcasing ERNIE’s capabilities across multiple
 
 </br>
 
+## Community
+
+| PaddlePaddle WeChat official account |  Join the tech discussion group |
+| :---: | :---: |
+| <img src="https://github.com/user-attachments/assets/864a45ec-0773-44b2-a2f1-c0e21e157792" width="150"> | <img src="https://github.com/user-attachments/assets/52e05674-7143-4207-8b19-67247fe88f55" width="150"> |
+
 ## License
 
 The ERNIE 4.5 models are provided under the Apache License 2.0. This license permits commercial use, subject to its terms and conditions.