mistralrs-core 0.8.1

Fast, flexible LLM inference.
Documentation 
1
2
3
4
5
6
7
8
9
10

use crate::{models::qwen2, vision_models::siglip};

#[derive(Debug, Clone, serde::Deserialize)]
pub struct MiniCpmOConfig {
    #[serde(flatten)]
    pub text_config: qwen2::Config,
    pub vision_config: siglip::SiglipVisionConfig,
    pub vision_batch_size: usize,
    pub query_num: usize,
}