refactor(项目结构) 重构为 workspace 多成员项目

- 将项目拆分为三个 crate：minicpm-core（核心模型）、minicpm-convert（转换功能）、minicpm-inference（推理功能）
- 添加两个示例：minimal-inference（最小推理）和 convert（模型转换）
- 转换后自动拷贝 config.json 和 tokenizer.json 到 model 目录
- 更新 README 说明 workspace 结构和使用方式

examples/convert/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "convert"
+version = "0.1.0"
+edition = "2021"
+
+[dependencies]
+minicpm-convert = { path = "../../crates/minicpm-convert" }
+burn = { version = "0.21", features = ["std", "wgpu"] }
+anyhow = "1.0"

examples/convert/src/main.rs

@@ -0,0 +1,24 @@
+use burn::backend::Wgpu;
+use minicpm_convert::export_model;
+use std::path::Path;
+
+fn main() -> anyhow::Result<()> {
+    let device = Default::default();
+
+    let safetensors_path = Path::new("MiniCPM5-1B/model-00000-of-00001.safetensors");
+    let config_path = Path::new("MiniCPM5-1B/config.json");
+    let tokenizer_path = Path::new("MiniCPM5-1B/tokenizer.json");
+    let output_dir = Path::new("model");
+
+    export_model::<Wgpu>(
+        safetensors_path,
+        config_path,
+        tokenizer_path,
+        output_dir,
+        &device,
+    )?;
+
+    println!("模型转换完成！");
+
+    Ok(())
+}

examples/minimal-inference/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "minimal-inference"
+version = "0.1.0"
+edition = "2021"
+
+[dependencies]
+minicpm-inference = { path = "../../crates/minicpm-inference" }
+burn = { version = "0.21", features = ["std", "wgpu"] }
+anyhow = "1.0"

examples/minimal-inference/src/main.rs

@@ -0,0 +1,46 @@
+// 运行方式：
+// 1. 确保已在项目根目录
+// 2. 运行：cargo run --release -p minimal-inference
+// 3. 确保模型文件存在：
+//    - model/model.mpk (模型权重)
+//    - MiniCPM5-1B/config.json (模型配置)
+//    - MiniCPM5-1B/tokenizer.json (分词器)
+
+use burn::backend::Wgpu;
+use minicpm_inference::{GenerationConfig, MiniCPM};
+use std::time::Instant;
+
+fn main() -> anyhow::Result<()> {
+    let device = Default::default();
+
+    println!("正在加载模型...");
+    let start = Instant::now();
+
+    let model = MiniCPM::<Wgpu>::load(
+        "model/model",
+        "MiniCPM5-1B/config.json",
+        "MiniCPM5-1B/tokenizer.json",
+        &device,
+    )?;
+
+    println!("模型加载完成，耗时: {:.2?}", start.elapsed());
+
+    let prompt = "你知道今天的天气不";
+    println!("\n用户: {}", prompt);
+    println!("Assistant: ");
+
+    let config = GenerationConfig {
+        max_new_tokens: Some(200),
+        temperature: 0.7,
+        top_p: 0.95,
+    };
+
+    let start = Instant::now();
+    let response = model.generate(prompt, false, &config)?;
+    let elapsed = start.elapsed();
+
+    println!("{}", response);
+    println!("\n生成耗时: {:.2?}", elapsed);
+
+    Ok(())
+}