Step3 Seamlessness is achieved by:
- native support OpenAI/Anthropic Compatible REST APIs Interface protocols, request formats are identical
- providing a standardized
/v1/completionsEndpoints that support mixed text/image input in JSON format
Implementation Steps:
- pass (a bill or inspection etc)
git clone https://github.com/stepfun-ai/Step3.gitGet Code - utilization
python -m vllm.entrypoints.api_serverStart the service (default port 8000) - Refer to the sample code to modify the original OpenAI client's
base_urllocal address
Tests have shown that the average transformation time to migrate existing applications can be kept under 2 hours.
This answer comes from the articleStep3: Efficient generation of open source big models for multimodal contentThe
May not be reproduced without permission:AI productivity tools " How to quickly implement multimodal services that are compatible with commercial APIs such as OpenAI
English 
简体中文 
日本語 
Deutsch 
Português do Brasil