aha/src/models/minicpm4/model.rs

use crate::{
    models::{
        base_modules::{AttentionNobias, MLPNoBias},
        minicpm4::config::MiniCPM4Config,
    },
    position_embed::rope::compute_default_rope_parameters,
    utils::tensor_utils::prepare_causal_attention_mask,
};
use anyhow::{Ok, Result};
use candle_core::{D, DType, Device, Tensor, Var};
use candle_nn::{embedding, rms_norm, Embedding, Linear, Module, RmsNorm, VarBuilder};

pub struct MiniCPMLongRoPE {
    head_dim: usize,
    rope_theta: f32,
    max_position_embeddings: usize,
    short_factor: Vec<f32>,
    long_factor: Vec<f32>,
    original_max_position_embeddings: usize,
    inv_freq: Tensor,
    cos_cached: Tensor,
    sin_cached: Tensor,
}
impl MiniCPMLongRoPE {
    pub fn new(cfg: &MiniCPM4Config, device: &Device) -> Result<Self> {
        let head_dim = cfg.hidden_size / cfg.num_attention_heads;
        let rope_theta = 10000.0;
        let max_position_embeddings = cfg.max_position_embeddings;
        let short_factor = cfg.rope_scaling.short_factor.clone();
        let long_factor = cfg.rope_scaling.short_factor.clone();
        let original_max_position_embeddings = cfg.rope_scaling.original_max_position_embeddings;
        let scale = max_position_embeddings / original_max_position_embeddings;
        let scaling_factor =
            (1.0 + (scale as f64).ln() + (original_max_position_embeddings as f64).ln()).sqrt();
        let inv_freq = compute_default_rope_parameters(head_dim, rope_theta);
        let inv_freq = Tensor::from_slice(&inv_freq, (1, inv_freq.len()), device)?;
        let t = Tensor::arange(0.0_f32, max_position_embeddings as f32, device)?
            .reshape((max_position_embeddings, 1))?;
        // short_factor.len() = 32
        // head_dim = 1024 / 16 = 64, inv_freq.len() = 32
        let ext_factors = Tensor::from_slice(&short_factor, (1, short_factor.len()), device)?;
        let ext_factors = Tensor::ones_like(&ext_factors)?.div(&ext_factors)?;
        // (seq_len, 1) matmul (1, 32) -> (seq_len, 32) * (1, 32)-> (seq_len, 32)
        let freqs = t.matmul(&ext_factors)?.broadcast_mul(&inv_freq)?;

        let emb = Tensor::cat(&[&freqs, &freqs], D::Minus1)?;
        let cos_cached = emb.cos()?.affine(scaling_factor, 0.0)?;
        let sin_cached = emb.sin()?.affine(scaling_factor, 0.0)?;
        Ok(Self {
            head_dim,
            rope_theta,
            max_position_embeddings,
            short_factor,
            long_factor,
            original_max_position_embeddings,
            inv_freq,
            cos_cached,
            sin_cached,
        })
    }
    pub fn update_cos_sin_cache(&mut self, seqlen: usize, device: &Device) -> Result<()> {
        let t = Tensor::arange(0.0_f32, seqlen as f32, device)?.reshape((seqlen, 1))?;
        let mut ext_factors =
            Tensor::from_slice(&self.short_factor, (1, self.short_factor.len()), device)?;
        if seqlen > self.original_max_position_embeddings {
            ext_factors =
                Tensor::from_slice(&self.long_factor, (1, self.long_factor.len()), device)?;
        }
        let ext_factors = Tensor::ones_like(&ext_factors)?.div(&ext_factors)?;
        let freqs = t.matmul(&ext_factors)?.broadcast_mul(&self.inv_freq)?;
        let emb = Tensor::cat(&[&freqs, &freqs], D::Minus1)?;
        let scale = seqlen / self.original_max_position_embeddings;
        let scaling_factor =
            (1.0 + (scale as f64).ln() + (self.original_max_position_embeddings as f64).ln())
                .sqrt();
        let cos_cached = emb.cos()?.affine(scaling_factor, 0.0)?;
        let sin_cached = emb.sin()?.affine(scaling_factor, 0.0)?;
        self.cos_cached = cos_cached;
        self.sin_cached = sin_cached;
        Ok(())
    }
    pub fn forward(&self, pos_offset: usize, seqlen: usize) -> Result<(Tensor, Tensor)> {
        let cos = self.cos_cached.narrow(0, pos_offset, seqlen)?;
        let sin = self.sin_cached.narrow(0, pos_offset, seqlen)?;
        Ok((cos, sin))
    }
}

pub struct MiniCPMDecoderLayer {
    self_attn: AttentionNobias,
    mlp: MLPNoBias,
    input_layernorm: RmsNorm,
    post_attention_layernorm: RmsNorm,
    scale_depth: f32,
    num_hidden_layers: usize,
}

impl MiniCPMDecoderLayer {
    pub fn new(vb: VarBuilder, cfg: &MiniCPM4Config) -> Result<Self> {
        let self_attn = AttentionNobias::new(
            vb.pp("self_attn"),
            cfg.hidden_size,
            cfg.num_attention_heads,
            cfg.num_key_value_heads,
        )?;
        let mlp = MLPNoBias::new(
            vb.pp("mlp"),
            cfg.hidden_size,
            cfg.intermediate_size,
            cfg.hidden_act,
        )?;
        let input_layernorm =
            rms_norm(cfg.hidden_size, cfg.rms_norm_eps, vb.pp("input_layernorm"))?;
        let post_attention_layernorm = rms_norm(
            cfg.hidden_size,
            cfg.rms_norm_eps,
            vb.pp("post_attention_layernorm"),
        )?;
        Ok(Self {
            self_attn,
            mlp,
            input_layernorm,
            post_attention_layernorm,
            scale_depth: cfg.scale_depth,
            num_hidden_layers: cfg.num_hidden_layers,
        })
    }

    pub fn forward(
        &self,
        xs: &Tensor,
        cos: &Tensor,
        sin: &Tensor,
        attention_mask: Option<&Tensor>,
    ) -> Result<Tensor> {
        let residual = xs;
        let xs = self.input_layernorm.forward(xs)?;
        let xs = self.self_attn.forward(&xs, cos, sin, attention_mask)?;
        let xs = (xs + residual)?;
        let residual = &xs;
        let xs = xs.apply(&self.post_attention_layernorm)?.apply(&self.mlp)?;
        let xs = (residual + xs)?;
        Ok(xs)
    }

    pub fn forward_step(
        &mut self,
        xs: &Tensor,
        cos: &Tensor,
        sin: &Tensor,
        attention_mask: Option<&Tensor>,
    ) -> Result<Tensor> {
        let residual = xs;
        let xs = self.input_layernorm.forward(xs)?;
        let xs = self.self_attn.forward_step(&xs, cos, sin, attention_mask)?;
        let xs = (xs + residual)?;
        let residual = &xs;
        let xs = xs.apply(&self.post_attention_layernorm)?.apply(&self.mlp)?;
        let xs = (residual + xs)?;
        Ok(xs)
    }
    pub fn clear_kv_cache(&mut self) {
        self.self_attn.clear_kv_cache();
    }
}

pub struct MiniCPMModel {
    cfg: MiniCPM4Config,
    embed_tokens: Embedding,
    layers: Vec<MiniCPMDecoderLayer>,
    norm: RmsNorm,
    rope_emb: MiniCPMLongRoPE,
    lm_head: Linear,
}

impl MiniCPMModel {
    pub fn new(vb: VarBuilder, cfg: MiniCPM4Config) -> Result<Self> {
        let embed_tokens = embedding(cfg.vocab_size, cfg.hidden_size, vb.pp("embed_tokens"))?;
        let mut layers = Vec::with_capacity(cfg.num_hidden_layers);
        let vb_layers = vb.pp("layers");
        for i in 0..cfg.num_hidden_layers {
            let layer = MiniCPMDecoderLayer::new(vb_layers.pp(i), &cfg)?;
            layers.push(layer);
        }
        let norm = rms_norm(cfg.hidden_size, cfg.rms_norm_eps, vb.pp("norm"))?;
        let rope_emb = MiniCPMLongRoPE::new(&cfg, vb.device())?;
        let lm_head = Linear::new(embed_tokens.embeddings().clone(), None);
        Ok(Self {
            cfg,
            embed_tokens,
            layers,
            norm,
            rope_emb,
            lm_head
        })
    }

    pub fn forward(&self, input_ids: &Tensor, position_id: usize) -> Result<Tensor> {
        let (bs, seq_len) = input_ids.dims2()?;
        let input_embeds = self.embed_tokens.forward(&input_ids)?;
        let attention_mask: Option<&Tensor> = {
            if seq_len <= 1 {
                None
            } else {
                Some(&prepare_causal_attention_mask(
                    bs,
                    seq_len,
                    position_id,
                    input_ids.device(),
                )?)
            }
        };

        let (cos, sin) = self.rope_emb.forward(position_id, seq_len)?;
        let mut hidden_states = input_embeds;
        for decode_layer in &self.layers {
            hidden_states = decode_layer.forward(&hidden_states, &cos, &sin, attention_mask)?;
        }
        hidden_states = self.norm.forward(&hidden_states)?;
        let hidden_state = hidden_states.narrow(1, seq_len - 1, 1)?;
        let logits = self.lm_head.forward(&hidden_state)?;
        Ok(logits)
    }

    pub fn forward_step(&mut self, input_ids: &Tensor, position_id: usize) -> Result<Tensor> {
        let (bs, seq_len) = input_ids.dims2()?;
        let input_embeds = self.embed_tokens.forward(&input_ids)?;
        let attention_mask: Option<&Tensor> = {
            if seq_len <= 1 {
                None
            } else {
                Some(&prepare_causal_attention_mask(
                    bs,
                    seq_len,
                    position_id,
                    input_ids.device(),
                )?)
            }
        };

        let (cos, sin) = self.rope_emb.forward(position_id, seq_len)?;
        let mut hidden_states = input_embeds;
        for decode_layer in &mut self.layers {
            hidden_states = decode_layer.forward_step(&hidden_states, &cos, &sin, attention_mask)?;
        }
        hidden_states = self.norm.forward(&hidden_states)?;
        let hidden_state = hidden_states.narrow(1, seq_len - 1, 1)?;
        let logits = self.lm_head.forward(&hidden_state)?;
        Ok(logits)
    }

    pub fn clear_kv_cache(&mut self) {
        for layer in self.layers.iter_mut() {
            layer.clear_kv_cache()
        }
    }
}