tllama 0.1.1

Lightweight Local LLM Inference Engine
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70

use clap::{Parser, Subcommand};

#[derive(Parser, Debug)]
#[command(version, about, long_about = None)]
pub struct Cli {
    #[command(subcommand)]
    pub command: Commands,
}

#[derive(Subcommand, Debug)]
pub enum Commands {
    Infer(InferArgs),
    Discover(DiscoverArgs),
    List,
    #[cfg(feature = "chat")]
    Chat(ChatArgs),
    Serve(ServeArgs),
}

#[derive(Parser, Debug)]
pub struct InferArgs {
    pub model: String,
    pub prompt: String,
    #[arg(short, long)]
    /// Number of tokens to generate
    pub n_len: Option<u32>,
    #[arg(short, long)]
    /// Sampling temperature
    /// Typical values are between 0.1 and 1.0
    pub temperature: Option<f32>,
    #[arg(short = 'k', long)]
    /// Top-k sampling
    /// Typical values are between 1 and 100
    pub top_k: Option<i32>,
    #[arg(short = 'p', long)]
    /// Top-p (nucleus) sampling
    /// Typical values are between 0.5 and 1.0
    pub top_p: Option<f32>,
    #[arg(short = 'r', long)]
    /// Repeat penalty
    /// Typical values are between 1.0 and 2.0
    pub repeat_penalty: Option<f32>,
    #[arg(short = 'c', long)]
    /// Context size
    /// Typical values are 512, 1024, 2048, etc.
    pub n_ctx: Option<i32>,
}

#[derive(Parser, Debug)]
pub struct DiscoverArgs {
    #[arg(short, long)]
    /// Scan all paths
    pub all: bool,
}

#[cfg(feature = "chat")]
#[derive(Parser, Debug)]
pub struct ChatArgs {
    pub model: String,
}

#[derive(Parser, Debug)]
pub struct ServeArgs {
    #[arg(long, default_value = "0.0.0.0")]
    pub host: String,
    #[arg(short, long, default_value_t = 8080)]
    pub port: u16,
    #[arg(long)]
    pub system_prompt: Option<String>,
}