Jack Zhou 85f3d7cbe7 | 1 year ago | |
---|---|---|
.. | ||
cpp | 1 year ago | |
python | 1 year ago | |
serving | 1 year ago | |
simple_serving | 1 year ago | |
README.md | 1 year ago |
⚡️FastDeploy 是一款全场景、易用灵活、极致高效的 AI 推理部署工具,满足开发者多硬件、多平台的产业部署需求。开发者可以基于 FastDeploy 将训练好的预测模型在不同的硬件、不同的推理引擎后端上进行部署。目前 FastDeploy 提供多种编程语言的 SDK,包括 C++、Python 以及 Java SDK。
在部署 ERNIE 3.0 模型前,需要安装 FastDeploy SDK,可参考 FastDeploy SDK安装文档确认部署环境是否满足 FastDeploy 环境要求,并按照介绍安装相应的 SDK。
目前,ERNIE 3.0 模型支持在如下的硬件以及推理引擎进行部署。
符号说明: (1) ✅: 已经支持; (2) ❔: 正在进行中; (3) N/A: 暂不支持;
硬件 | 可用的推理引擎 | 是否支持 Paddle 新格式量化模型 | 是否支持 FP16 模式 |
CPU | Paddle Inference | ✅ | N/A |
ONNX Runtime | ✅ | N/A | |
OpenVINO | ✅ | N/A | |
GPU | Paddle Inference | ✅ | N/A |
ONNX Runtime | ✅ | ❔ | |
Paddle TensorRT | ✅ | ✅ | |
TensorRT | ✅ | ✅ | |
昆仑芯 XPU | Paddle Lite | N/A | ✅ |
华为 昇腾 | Paddle Lite | ❔ | ✅ |
Graphcore IPU | Paddle Inference | ❔ | N/A |
符号说明: (1) ✅: 已经支持; (2) ❔: 正在进行中; (3) N/A: 暂不支持;
任务 Task | 部署方式 | 是否支持 |
文本分类 | Python | ✅ |
C++ | ✅ | |
Serving | ✅ | |
序列标注 | Python | ✅ |
C++ | ✅ | |
Serving | ✅ |
ERNIE 3.0 模型支持 Python、C++ 部署以及 Serving 服务化部署。以下是详细文档。
No Description
Python C++ Cuda Text Markdown other
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》