modeling_helion.py

"""PyTorch Helion model implementation."""

import math
from typing import Optional, Tuple, Union, List

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import CrossEntropyLoss

from transformers import PreTrainedModel
from transformers.modeling_outputs import CausalLMOutputWithPast
from transformers.utils import logging

from .configuration_helion import HelionConfig

logger = logging.get_logger(__name__)


class HelionRMSNorm(nn.Module):
    """Root Mean Square Layer Normalization."""
    
    def __init__(self, hidden_size: int, eps: float = 1e-6):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(hidden_size))
        self.variance_epsilon = eps

    def forward(self, hidden_states):
        input_dtype = hidden_states.dtype
        hidden_states = hidden_states.to(torch.float32)
        variance = hidden_states.pow(2).mean(-1, keepdim=True)
        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
        return self.weight * hidden_states.to(input_dtype)


class HelionRotaryEmbedding(nn.Module):
    """Rotary Position Embedding (RoPE)."""
    
    def __init__(self, dim: int, max_position_embeddings: int = 8192, base: int = 10000):
        super().__init__()
        self.dim = dim
        self.max_position_embeddings = max_position_embeddings
        self.base = base
        
        inv_freq = 1.0 / (self.base ** (torch.arange(0, self.dim, 2).float() / self.dim))
        self.register_buffer("inv_freq", inv_freq, persistent=False)
        
    def forward(self, x, seq_len: int):
        # Generate position indices
        t = torch.arange(seq_len, device=x.device, dtype=self.inv_freq.dtype)
        freqs = torch.outer(t, self.inv_freq)
        emb = torch.cat((freqs, freqs), dim=-1)
        return emb.cos().to(x.dtype), emb.sin().to(x.dtype)


def rotate_half(x):
    """Rotates half the hidden dims of the input."""
    x1 = x[..., : x.shape[-1] // 2]
    x2 = x[..., x.shape[-1] // 2 :]
    return torch.cat((-x2, x1), dim=-1)


def apply_rotary_pos_emb(q, k, cos, sin, position_ids):
    """Apply rotary position embedding to queries and keys."""
    cos = cos[position_ids].unsqueeze(1)
    sin = sin[position_ids].unsqueeze(1)
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


class HelionAttention(nn.Module):
    """Multi-head attention with Grouped Query Attention (GQA)."""
    
    def __init__(self, config: HelionConfig):
        super().__init__()
        self.config = config
        self.hidden_size = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.hidden_size // self.num_heads
        self.num_key_value_heads = config.num_key_value_heads
        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
        self.max_position_embeddings = config.max_position_embeddings
        
        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
        self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
        self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
        self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
        
        self.rotary_emb = HelionRotaryEmbedding(
            self.head_dim,
            max_position_embeddings=self.max_position_embeddings,
            base=config.rope_theta,
        )
        
    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Tuple[torch.Tensor]] = None,
        output_attentions: bool = False,
        use_cache: bool = False,
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        bsz, q_len, _ = hidden_states.size()
        
        query_states = self.q_proj(hidden_states)
        key_states = self.k_proj(hidden_states)
        value_states = self.v_proj(hidden_states)
        
        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
        
        kv_seq_len = key_states.shape[-2]
        if past_key_value is not None:
            kv_seq_len += past_key_value[0].shape[-2]
            
        cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
        
        if past_key_value is not None:
            key_states = torch.cat([past_key_value[0], key_states], dim=2)
            value_states = torch.cat([past_key_value[1], value_states], dim=2)
            
        past_key_value = (key_states, value_states) if use_cache else None
        
        # Repeat k/v heads for GQA
        key_states = key_states.repeat_interleave(self.num_key_value_groups, dim=1)
        value_states = value_states.repeat_interleave(self.num_key_value_groups, dim=1)
        
        attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
        
        if attention_mask is not None:
            attn_weights = attn_weights + attention_mask
            
        attn_weights = F.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
        attn_output = torch.matmul(attn_weights, value_states)
        
        attn_output = attn_output.transpose(1, 2).contiguous()
        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
        attn_output = self.o_proj(attn_output)
        
        return attn_output, attn_weights if output_attentions else None, past_key_value


class HelionMLP(nn.Module):
    """Feed-forward network with SwiGLU activation."""
    
    def __init__(self, config: HelionConfig):
        super().__init__()
        self.hidden_size = config.hidden_size
        self.intermediate_size = config.intermediate_size
        
        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
        self.act_fn = nn.SiLU()
        
    def forward(self, x):
        return self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))


class HelionDecoderLayer(nn.Module):
    """Single transformer decoder layer."""
    
    def __init__(self, config: HelionConfig):
        super().__init__()
        self.hidden_size = config.hidden_size
        self.self_attn = HelionAttention(config)
        self.mlp = HelionMLP(config)
        self.input_layernorm = HelionRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.post_attention_layernorm = HelionRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        
    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Tuple[torch.Tensor]] = None,
        output_attentions: Optional[bool] = False,
        use_cache: Optional[bool] = False,
    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
        residual = hidden_states
        hidden_states = self.input_layernorm(hidden_states)
        
        # Self Attention
        hidden_states, self_attn_weights, present_key_value = self.self_attn(
            hidden_states=hidden_states,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_value=past_key_value,
            output_attentions=output_attentions,
            use_cache=use_cache,
        )
        hidden_states = residual + hidden_states
        
        # MLP
        residual = hidden_states
        hidden_states = self.post_attention_layernorm(hidden_states)
        hidden_states = self.mlp(hidden_states)
        hidden_states = residual + hidden_states
        
        outputs = (hidden_states,)
        if output_attentions:
            outputs += (self_attn_weights,)
        if use_cache:
            outputs += (present_key_value,)
            
        return outputs


class HelionPreTrainedModel(PreTrainedModel):
    """Helion pretrained model base class."""
    
    config_class = HelionConfig
    base_model_prefix = "model"
    supports_gradient_checkpointing = True
    _no_split_modules = ["HelionDecoderLayer"]
    
    def _init_weights(self, module):
        std = self.config.initializer_range
        if isinstance(module, nn.Linear):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=std)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()


class HelionModel(HelionPreTrainedModel):
    """Helion transformer model."""
    
    def __init__(self, config: HelionConfig):
        super().__init__(config)
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size
        
        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
        self.layers = nn.ModuleList([HelionDecoderLayer(config) for _ in range(config.num_hidden_layers)])
        self.norm = HelionRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        
        self.gradient_checkpointing = False
        self.post_init()
        
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ):
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        use_cache = use_cache if use_cache is not None else self.config.use_cache
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        
        if input_ids is not None and inputs_embeds is not None:
            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
        elif input_ids is not None:
            batch_size, seq_length = input_ids.shape
        elif inputs_embeds is not None:
            batch_size, seq_length, _ = inputs_embeds.shape
        else:
            raise ValueError("You have to specify either input_ids or inputs_embeds")
            
        if position_ids is None:
            position_ids = torch.arange(seq_length, dtype=torch.long, device=input_ids.device if input_ids is not None else inputs_embeds.device)
            position_ids = position_ids.unsqueeze(0).view(-1, seq_length)
            
        if inputs_embeds is None:
            inputs_embeds = self.embed_tokens(input_ids)
            
        hidden_states = inputs_embeds
        
        for idx, decoder_layer in enumerate(self.layers):
            past_key_value = past_key_values[idx] if past_key_values is not None else None
            
            layer_outputs = decoder_layer(
                hidden_states,
                attention_mask=attention_mask,
                position_ids=position_ids,
                past_key_value=past_key_value,
                output_attentions=output_attentions,
                use_cache=use_cache,
            )
            
            hidden_states = layer_outputs[0]
            
        hidden_states = self.norm(hidden_states)
        
        return hidden_states


class HelionForCausalLM(HelionPreTrainedModel):
    """Helion model with language modeling head."""
    
    def __init__(self, config):
        super().__init__(config)
        self.model = HelionModel(config)
        self.vocab_size = config.vocab_size
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
        
        self.post_init()
        
    def get_input_embeddings(self):
        return self.model.embed_tokens
        
    def set_input_embeddings(self, value):
        self.model.embed_tokens = value
        
    def get_output_embeddings(self):
        return self.lm_head
        
    def set_output_embeddings(self, new_embeddings):
        self.lm_head = new_embeddings
        
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[List[torch.FloatTensor]] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple, CausalLMOutputWithPast]:
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        
        hidden_states = self.model(
            input_ids=input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            past_key_values=past_key_values,
            inputs_embeds=inputs_embeds,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )
        
        logits = self.lm_head(hidden_states)
        logits = logits.float()
        
        loss = None
        if labels is not None:
            shift_logits = logits[..., :-1, :].contiguous()
            shift_labels = labels[..., 1:].contiguous()
            loss_fct = CrossEntropyLoss()
            shift_logits = shift_logits.view(-1, self.config.vocab_size)
            shift_labels = shift_labels.view(-1)
            shift_labels = shift_labels.to(shift_logits.device)
            loss = loss_fct(shift_logits, shift_labels)
            
        if not return_dict:
            output = (logits,)
            return (loss,) + output if loss is not None else output
            
        return CausalLMOutputWithPast(
            loss=loss,
            logits=logits,
            past_key_values=past_key_values,
            hidden_states=None,
            attentions=None,
        )
        
    def prepare_inputs_for_generation(
        self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
    ):
        if past_key_values:
            input_ids = input_ids[:, -1:]
            
        position_ids = kwargs.get("position_ids", None)
        if attention_mask is not None and position_ids is None:
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            if past_key_values:
                position_ids = position_ids[:, -1].unsqueeze(-1)
                
        if inputs_embeds is not None and past_key_values is None:
            model_inputs = {"inputs_embeds": inputs_embeds}
        else:
            model_inputs = {"input_ids": input_ids}
            
        model_inputs.update({
            "position_ids": position_ids,
            "past_key_values": past_key_values,
            "use_cache": kwargs.get("use_cache"),
            "attention_mask": attention_mask,
        })
        
        return model_inputs