Introduction
ERNIE-4.5-0.3B-PT is a lightweight foundational LLM from Baidu Wenxin series.This release is fully adapted to Muxi C550 chip with FlagOS acceleration stack.Precompiled Triton cache, FlagGems operator logs and 4k context benchmark data are embedded in the model folder.It supports out-of-the-box vLLM inference via vLLM-plugin-FL, with obvious throughput improvement compared with native implementation on Muxi hardware.
Integrated Deployment
- Out-of-the-box inference scripts with pre-configured hardware and software parameters
- Released FlagOS-Metax container image supporting deployment within minutes
Consistency Validation
- Rigorously evaluated through benchmark testing: Performance and results from the FlagOS software stack are compared against native stacks on multiple public.
Evaluation Results
Benchmark Result
| Metrics | ERNIE-4.5-0.3B-PT-Nvidia-Origin | ERNIE-4.5-0.3B-PT-Metax-FlagOS |
|---|
| musr_generative | 0.377 | 0.3717 |
| mmlu_pro | 0.1737 | 0.1758 |
| gpqa_generative_cot | 0.25 | 0.2424 |
| livebench_new | 0.1495 | 0.1456 |
User Guide
Environment Setup
| Item | Version |
|---|
| Docker Version | Docker version 27.5.1, build 27.5.1-0ubuntu3~22.04.2 |
| Operating System | Ubuntu 22.04.5 LTS (Jammy Jellyfish) |
Operation Steps
Download FlagOS Image
1docker pull harbor.baai.ac.cn/external-cooperation/ernie-4.5-0.3b-pt-muxi-tree_0.5.1-gems_5.0.2-vllm_0.13.0-plugin_0.1.1-cx_none-python_3.12-torch_2.8.0-pcp_maca3.3.0.15-gpu_c550-arc_amd64-driver_3.3.12:2606161504
2
Download Open-source Model Weights
1pip install modelscope
2modelscope download --model FlagRelease/ERNIE-4.5-0.3B-PT-metax-FlagOS --local_dir /data/ERNIE-4.5-0.3B-PT-metax-FlagOS
Start the Container
1 docker run -d \
2 --name ernie-4.5-0.3b-flagos \
3 --privileged \
4 --net=host \
5 --ipc=host \
6 -v /usr/local/models:/usr/local/models \
7 -v /usr/local/dev:/usr/local/dev \
8 -v /data:/data \
9 harbor.baai.ac.cn/external-cooperation/ernie-4.5-0.3b-pt-muxi-tree_0.5.1-gems_5.0.2-vllm_0.13.0-plugin_0.1.1-cx_none-python_3.12-torch_2.8.0-pcp_maca3.3.0.15-gpu_c550-arc_amd64-driver_3.3.12:2606161504 \
10 sleep infinity
11docker exec -it ernie-4.5-0.3b-flagos /bin/bash
Start the Server
1export VLLM_PLUGINS=fl
2export TRITON_ALL_BLOCKS_PARALLEL=1
3export VLLM_USE_MODELSCOPE=true
4
5ulimit -n 2048 && nohup env VLLM_FL_FLAGOS_BLACKLIST="masked_fill,masked_fill_,mm,sort,sort_stable,cumsum,cumsum_out,gather,exponential_,arange_start,index" vllm serve /data/ERNIE-4.5-0.3B-PT-metax-FlagOS \
6 --served-model-name ernie-4.5-0.3b-flagos \
7 --port 9010 \
8 --tensor-parallel-size 1 \
9 --enforce-eager \
10 --trust-remote-code \
11 > ernie_flagos.log 2>&1 &
Service Invocation
Invocation Script
1curl http://localhost:9010/v1/chat/completions \
2 -H "Content-Type: application/json" \
3 -d '{
4 "model": "ernie-4.5-0.3b-flagos",
5 "messages": [{"role": "user", "content": "你好"}]
6 }'
AnythingLLM Integration Guide
1. Download & Install
- Visit the official site: https://anythingllm.com/
- Choose the appropriate version for your OS (Windows/macOS/Linux)
- Follow the installation wizard to complete the setup
2. Configuration
- Launch AnythingLLM
- Open settings (bottom left, fourth tab)
- Configure core LLM parameters
- Click "Save Settings" to apply changes
3. Model Interaction
- After model loading is complete:
- Click "New Conversation"
- Enter your question (e.g., “Explain the basics of quantum computing”)
- Click the send button to get a response
Technical Overview
FlagOS is a fully open-source system software stack designed to unify the "model–system–chip" layers and foster an open, collaborative ecosystem. It enables a “develop once, run anywhere” workflow across diverse AI accelerators, unlocking hardware performance, eliminating fragmentation among vendor-specific software stacks, and substantially lowering the cost of porting and maintaining AI workloads. With core technologies such as the FlagScale, together with vllm-plugin-fl, distributed training/inference framework, FlagGems universal operator library, FlagCX communication library, and FlagTree unified compiler, the FlagRelease platform leverages the FlagOS stack to automatically produce and release various combinations of <chip + open-source model>. This enables efficient and automated model migration across diverse chips, opening a new chapter for large model deployment and application.
FlagGems
FlagGems is a high-performance, generic operator libraryimplemented in
Triton language. It is built on a collection of backend-neutralkernels that aims to accelerate LLM (Large-Language Models) training and inference across diverse hardware platforms.
FlagTree
FlagTree is an open source, unified compiler for multipleAI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-repository multi-backend support. Forupstream model users, it provides unified compilation capabilities across multiple backends; for downstream chip manufacturers, it offers examples of Triton ecosystem integration.
FlagScale and vllm-plugin-fl
Flagscale is a comprehensive toolkit designed to supportthe entire lifecycle of large models. It builds on the strengths of several prominent open-source projects, including
Megatron-LM and
vLLM, to provide a robust, end-to-end solution for managing and scaling large models.
vllm-plugin-fl is a vLLM plugin built on the FlagOS unified multi-chip backend, to help flagscale support multi-chip on vllm framework.
FlagCX
FlagCX is a scalable and adaptive cross-chip communication library. It serves as a platform where developers, researchers, and AI engineers can collaborate on various projects, contribute to the development of cutting-edge AI solutions, and share their work with the global community.
FlagEval Evaluation Framework
FlagEval is a comprehensive evaluation system and open platform for large models launched in 2023. It aims to establish scientific, fair, and open benchmarks, methodologies, and tools to help researchers assess model and training algorithm performance. It features:
- Multi-dimensional Evaluation: Supports 800+ modelevaluations across NLP, CV, Audio, and Multimodal fields,covering 20+ downstream tasks including language understanding and image-text generation.
- Industry-Grade Use Cases: Has completed horizonta1 evaluations of mainstream large models, providing authoritative benchmarks for chip-model performance validation.
Contributing
We warmly welcome global developers to join us:
- Submit Issues to report problems
- Create Pull Requests to contribute code
- Improve technical documentation
- Expand hardware adaptation support
License
The model weights are derived from PaddlePaddle/ERNIE-4.5-0.3B-PT and are open‑sourced under the Apache License 2.0:
https://www.apache.org/licenses/LICENSE-2.0.txt
