JGZYES
28535b1c74
- pmai unified CLI (generate/chat/interactive/http/inspect/output/config) - GPT+MoE transformer, .pap (f32/fp16/q8) + GGUF loader (order+version adaptive) - llama/Mixtral arch: RoPE+GQA+SwiGLU+MoE (C==torch verified) - C llama BPE tokenizer (validated vs llama-cpp-python) - training framework + 0.1B/0.22B MoE models; quantization fp16/q8 - build artifacts to output/; HTTP API; config.yaml; scripts; openapi
52 行
1.5 KiB
YAML
52 行
1.5 KiB
YAML
openapi: 3.0.0
|
|
info:
|
|
title: ParlzMAI API
|
|
version: 1.0.0
|
|
description: pmai HTTP 服务（--serve，需 -DBUILD_HTTP=ON）。MoE 推理生成。
|
|
servers:
|
|
- url: http://127.0.0.1:11434
|
|
paths:
|
|
/health:
|
|
get:
|
|
summary: 健康检查
|
|
responses:
|
|
'200': { description: ok }
|
|
/api/models:
|
|
get:
|
|
summary: 已加载模型信息
|
|
responses:
|
|
'200':
|
|
description: JSON 模型信息
|
|
content:
|
|
application/json:
|
|
schema:
|
|
type: object
|
|
properties:
|
|
model: { type: string }
|
|
params: { type: integer }
|
|
arch: { type: string }
|
|
n_experts: { type: integer }
|
|
/api/generate:
|
|
post:
|
|
summary: 生成（SSE 流式）
|
|
requestBody:
|
|
required: true
|
|
content:
|
|
application/json:
|
|
schema:
|
|
type: object
|
|
properties:
|
|
prompt: { type: string, example: 'ROMEO:' }
|
|
options:
|
|
type: object
|
|
properties:
|
|
max_tokens: { type: integer, example: 60 }
|
|
temperature: { type: number, example: 0.8 }
|
|
top_k: { type: integer, example: 40 }
|
|
responses:
|
|
'200':
|
|
description: SSE text/event-stream，每事件 data: {"response":"..."}
|
|
content:
|
|
text/event-stream:
|
|
schema: { type: string }
|