modeling_twinkel_llm.py

"""
Twinkel LLM Model Implementation
Creator: Kunal Pandey
"""

import torch
import torch.nn as nn
import torch.nn.functional as F
from transformers import PreTrainedModel
from transformers.modeling_outputs import CausalLMOutputWithPast
from typing import Optional, Tuple, Union
from .configuration_twinkel_llm import TwinkelLLMConfig


class RMSNorm(nn.Module):
    def __init__(self, dim, eps=1e-6):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(dim))
        self.eps = eps

    def forward(self, x):
        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps) * self.weight


class RotaryEmbedding(nn.Module):
    def __init__(self, dim, max_seq_len=2048):
        super().__init__()
        inv_freq = 1.0 / (10000 ** (torch.arange(0, dim, 2).float() / dim))
        self.register_buffer("inv_freq", inv_freq, persistent=False)
        self.max_seq_len = max_seq_len
        self._set_cos_sin_cache(max_seq_len)

    def _set_cos_sin_cache(self, seq_len):
        self.max_seq_len = seq_len
        t = torch.arange(seq_len, dtype=torch.float32)
        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
        emb = torch.cat((freqs, freqs), dim=-1)
        self.register_buffer("cos_cached", emb.cos(), persistent=False)
        self.register_buffer("sin_cached", emb.sin(), persistent=False)

    def forward(self, x, seq_len):
        if seq_len > self.max_seq_len:
            self._set_cos_sin_cache(seq_len)
        return self.cos_cached[:seq_len].to(x.device), self.sin_cached[:seq_len].to(x.device)


def apply_rotary_pos_emb(q, k, cos, sin):
    def rotate_half(x):
        x1, x2 = x.chunk(2, dim=-1)
        return torch.cat((-x2, x1), dim=-1)
    
    cos = cos.unsqueeze(1)
    sin = sin.unsqueeze(1)
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


class GroupedQueryAttention(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.n_heads = config.num_attention_heads
        self.n_kv_heads = config.num_key_value_heads
        self.head_dim = config.hidden_size // self.n_heads
        self.hidden_size = config.hidden_size
        self.n_rep = self.n_heads // self.n_kv_heads

        self.q_proj = nn.Linear(self.hidden_size, self.n_heads * self.head_dim, bias=False)
        self.k_proj = nn.Linear(self.hidden_size, self.n_kv_heads * self.head_dim, bias=False)
        self.v_proj = nn.Linear(self.hidden_size, self.n_kv_heads * self.head_dim, bias=False)
        self.o_proj = nn.Linear(self.n_heads * self.head_dim, self.hidden_size, bias=False)

        self.rope = RotaryEmbedding(self.head_dim)

    def forward(self, x):
        B, T, C = x.shape

        q = self.q_proj(x).view(B, T, self.n_heads, self.head_dim).transpose(1, 2)
        k = self.k_proj(x).view(B, T, self.n_kv_heads, self.head_dim).transpose(1, 2)
        v = self.v_proj(x).view(B, T, self.n_kv_heads, self.head_dim).transpose(1, 2)

        cos, sin = self.rope(x, T)
        q = q.transpose(1, 2)
        k = k.transpose(1, 2)
        q, k = apply_rotary_pos_emb(q, k, cos, sin)
        q = q.transpose(1, 2)
        k = k.transpose(1, 2)

        if self.n_rep > 1:
            k = k.repeat_interleave(self.n_rep, dim=1)
            v = v.repeat_interleave(self.n_rep, dim=1)

        attn_weights = torch.matmul(q, k.transpose(-2, -1)) / (self.head_dim ** 0.5)
        causal_mask = torch.triu(torch.ones(T, T, device=x.device, dtype=torch.bool), diagonal=1)
        attn_weights = attn_weights.masked_fill(causal_mask, float('-inf'))
        attn_weights = F.softmax(attn_weights, dim=-1, dtype=torch.float32).to(x.dtype)
        attn_output = torch.matmul(attn_weights, v)

        attn_output = attn_output.transpose(1, 2).contiguous().view(B, T, -1)
        return self.o_proj(attn_output)


class SwiGLU(nn.Module):
    def __init__(self, config):
        super().__init__()
        hidden = config.hidden_size
        intermediate = config.intermediate_size
        
        self.gate_proj = nn.Linear(hidden, intermediate, bias=False)
        self.up_proj = nn.Linear(hidden, intermediate, bias=False)
        self.down_proj = nn.Linear(intermediate, hidden, bias=False)

    def forward(self, x):
        return self.down_proj(F.silu(self.gate_proj(x)) * self.up_proj(x))


class TransformerBlock(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.attn_norm = RMSNorm(config.hidden_size)
        self.ffn_norm = RMSNorm(config.hidden_size)
        self.attn = GroupedQueryAttention(config)
        self.ffn = SwiGLU(config)
        self.dropout = nn.Dropout(config.dropout)

    def forward(self, x):
        x = x + self.dropout(self.attn(self.attn_norm(x)))
        x = x + self.dropout(self.ffn(self.ffn_norm(x)))
        return x


class TwinkelLLMPreTrainedModel(PreTrainedModel):
    config_class = TwinkelLLMConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["TransformerBlock"]

    def _init_weights(self, module):
        if isinstance(module, nn.Linear):
            torch.nn.init.normal_(module.weight, mean=0.0, std=0.02)
        elif isinstance(module, nn.Embedding):
            torch.nn.init.normal_(module.weight, mean=0.0, std=0.02)


class TwinkelLLMForCausalLM(TwinkelLLMPreTrainedModel):
    """
    Twinkel LLM Model for Causal Language Modeling
    Creator: Kunal Pandey
    """
    
    def __init__(self, config):
        super().__init__(config)
        self.config = config
        
        self.token_embedding = nn.Embedding(config.vocab_size, config.hidden_size)
        self.layers = nn.ModuleList([TransformerBlock(config) for _ in range(config.num_hidden_layers)])
        self.norm = RMSNorm(config.hidden_size)
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
        
        # Weight tying
        self.lm_head.weight = self.token_embedding.weight
        
        self.post_init()

    def get_input_embeddings(self):
        return self.token_embedding

    def set_input_embeddings(self, value):
        self.token_embedding = value

    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        labels: Optional[torch.LongTensor] = None,
        **kwargs
    ) -> Union[Tuple, CausalLMOutputWithPast]:
        
        x = self.token_embedding(input_ids)
        
        for layer in self.layers:
            x = layer(x)
        
        x = self.norm(x)
        logits = self.lm_head(x)

        loss = None
        if labels is not None:
            shift_logits = logits[..., :-1, :].contiguous()
            shift_labels = labels[..., 1:].contiguous()
            loss = F.cross_entropy(
                shift_logits.view(-1, self.config.vocab_size),
                shift_labels.view(-1),
                ignore_index=-100,
                reduction='mean'
            )
        
        return CausalLMOutputWithPast(
            loss=loss,
            logits=logits,
        )

    def prepare_inputs_for_generation(self, input_ids, **kwargs):
        return {"input_ids": input_ids}