LLamaRecipes/research/long-context-llama/H2O/utils_llama.py

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

import pdb
import types
import torch
from torch import nn
import torch.utils.checkpoint
import torch.nn.functional as F

from transformers.models.llama.configuration_llama import LlamaConfig
from transformers.models.llama.modeling_llama import (
    LlamaAttention,
    rotate_half,
    apply_rotary_pos_emb,
    repeat_kv,
    LlamaRotaryEmbedding,
    LlamaForCausalLM,
)
from kv_cache import Cache, HHCache, StaticCache
from transformers.utils import logging
from transformers.modeling_outputs import BaseModelOutputWithPast

logger = logging.get_logger(__name__)

__all__ = ["H2OLlamaForCausalLM"]

def _make_causal_mask(
    bsz: int, tgt_len: int, past_key_values_length: int, dtype: torch.dtype, device: torch.device):
    """
    Make causal mask used for bi-directional self-attention.
    """
    mask = torch.full((tgt_len, tgt_len), torch.finfo(dtype).min, device=device)
    mask_cond = torch.arange(mask.size(-1), device=device)
    mask.masked_fill_(mask_cond < (mask_cond + 1).view(mask.size(-1), 1), 0)
    mask = mask.to(dtype)

    if past_key_values_length > 0:
        mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
    return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)

def apply_rotary_pos_emb_single(x, cos, sin, position_ids=None, unsqueeze_dim=1):

    cos = cos.unsqueeze(unsqueeze_dim)
    sin = sin.unsqueeze(unsqueeze_dim)
    x_embed = (x * cos) + (rotate_half(x) * sin)

    return x_embed

def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
    """
    This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
    num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
    """
    batch, num_key_value_heads, slen, head_dim = hidden_states.shape
    if n_rep == 1:
        return hidden_states
    hidden_states = hidden_states[:, :, None, :, :].expand(batch, num_key_value_heads, n_rep, slen, head_dim)
    return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)


class H2OLlamaAttention(nn.Module):
    """Multi-headed attention from 'Attention Is All You Need' paper"""

    def __init__(self, config: LlamaConfig, layer_idx: Optional[int] = None):
        super().__init__()
        self.config = config
        self.layer_idx = layer_idx
        if layer_idx is None:
            logger.warning_once(
                f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will "
                "lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` "
                "when creating this class."
            )

        self.attention_dropout = config.attention_dropout
        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.rope_theta = config.rope_theta
        self.is_causal = True
        self.positional_rolling = config.enable_position_rolling

        if (self.head_dim * self.num_heads) != self.hidden_size:
            raise ValueError(
                f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
                f" and `num_heads`: {self.num_heads})."
            )

        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=config.attention_bias)
        self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.attention_bias)
        self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.attention_bias)
        self.o_proj = nn.Linear(self.hidden_size, self.hidden_size, bias=config.attention_bias)
        self._init_rope()

    def _init_rope(self):
        if self.config.rope_scaling is None:
            self.rotary_emb = LlamaRotaryEmbedding(
                self.head_dim,
                max_position_embeddings=self.max_position_embeddings,
                base=self.rope_theta,
            )
        else:
            scaling_type = self.config.rope_scaling["type"]
            scaling_factor = self.config.rope_scaling["factor"]
            if scaling_type == "linear":
                self.rotary_emb = LlamaLinearScalingRotaryEmbedding(
                    self.head_dim,
                    max_position_embeddings=self.max_position_embeddings,
                    scaling_factor=scaling_factor,
                    base=self.rope_theta,
                )
            elif scaling_type == "dynamic":
                self.rotary_emb = LlamaDynamicNTKScalingRotaryEmbedding(
                    self.head_dim,
                    max_position_embeddings=self.max_position_embeddings,
                    scaling_factor=scaling_factor,
                    base=self.rope_theta,
                )
            else:
                raise ValueError(f"Unknown RoPE scaling type {scaling_type}")

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.Tensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_value: Optional[Cache] = None,
        output_attentions: bool = False,
        use_cache: bool = False,
        cache_position: Optional[torch.LongTensor] = None,
        **kwargs,
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        bsz, q_len, _ = hidden_states.size()

        if self.config.pretraining_tp > 1:
            key_value_slicing = (self.num_key_value_heads * self.head_dim) // self.config.pretraining_tp
            query_slices = self.q_proj.weight.split(
                (self.num_heads * self.head_dim) // self.config.pretraining_tp, dim=0
            )
            key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
            value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)

            query_states = [F.linear(hidden_states, query_slices[i]) for i in range(self.config.pretraining_tp)]
            query_states = torch.cat(query_states, dim=-1)

            key_states = [F.linear(hidden_states, key_slices[i]) for i in range(self.config.pretraining_tp)]
            key_states = torch.cat(key_states, dim=-1)

            value_states = [F.linear(hidden_states, value_slices[i]) for i in range(self.config.pretraining_tp)]
            value_states = torch.cat(value_states, dim=-1)

        else:
            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)

        past_key_value = getattr(self, "past_key_value", past_key_value)

        if not self.positional_rolling:
            cos, sin = self.rotary_emb(value_states, position_ids)
            query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
            if past_key_value is not None:
                # sin and cos are specific to RoPE models; cache_position needed for the static cache
                cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
                key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
        else:
            if past_key_value is not None:
                # sin and cos are specific to RoPE models; cache_position needed for the static cache
                key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx)

            kv_seq_len = past_key_value.get_seq_length(self.layer_idx) if past_key_value is not None else key_states.shape[-2]

            if not position_ids.nelement() > 1:
                # decoding stage
                key_position_ids = torch.arange(kv_seq_len, device=hidden_states.device).unsqueeze(0)
                query_position_ids = key_position_ids[:, -1].unsqueeze(0)
            else:
                query_position_ids = position_ids
                key_position_ids = position_ids

            key_cos, key_sin = self.rotary_emb(value_states, key_position_ids)
            query_cos, query_sin = self.rotary_emb(value_states, query_position_ids)

            if self.layer_idx == 0:
                print(kv_seq_len, query_position_ids, key_position_ids)

            query_states = apply_rotary_pos_emb_single(query_states, query_cos, query_sin)
            key_states = apply_rotary_pos_emb_single(key_states, key_cos, key_sin)

        key_states = repeat_kv(key_states, self.num_key_value_groups)
        value_states = repeat_kv(value_states, self.num_key_value_groups)

        attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)

        if attention_mask is not None:  # no matter the length, we just slice it
            causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
            attn_weights = attn_weights + causal_mask

        # upcast attention to fp32
        attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)

        # Update KV Cache based on Heavy-Hitter Oracle
        if past_key_value is not None:
            past_key_value.update_slimming(attn_weights, self.num_key_value_groups, self.layer_idx)

        attn_weights = nn.functional.dropout(attn_weights, p=self.attention_dropout, training=self.training)
        attn_output = torch.matmul(attn_weights, value_states)

        if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
            raise ValueError(
                f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
                f" {attn_output.size()}"
            )

        attn_output = attn_output.transpose(1, 2).contiguous()

        attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)

        if self.config.pretraining_tp > 1:
            attn_output = attn_output.split(self.hidden_size // self.config.pretraining_tp, dim=2)
            o_proj_slices = self.o_proj.weight.split(self.hidden_size // self.config.pretraining_tp, dim=1)
            attn_output = sum([F.linear(attn_output[i], o_proj_slices[i]) for i in range(self.config.pretraining_tp)])
        else:
            attn_output = self.o_proj(attn_output)

        if not output_attentions:
            attn_weights = None
        
        return attn_output, attn_weights, past_key_value


def enable_h2ocache_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,
    cache_position: Optional[torch.LongTensor] = None,
) -> Union[Tuple, BaseModelOutputWithPast]:
    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 None) ^ (inputs_embeds is not None):
        raise ValueError(
            "You cannot specify both input_ids and inputs_embeds at the same time, and must specify either one"
        )

    if self.gradient_checkpointing and self.training and use_cache:
        logger.warning_once(
            "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`."
        )
        use_cache = False

    if inputs_embeds is None:
        inputs_embeds = self.embed_tokens(input_ids)

    past_seen_tokens = 0
    if use_cache:  # kept for BC (cache positions)
        if not isinstance(past_key_values, StaticCache):
            past_key_values = HHCache.from_legacy_cache(self.num_window_length, self.num_heavy_hitter_tokens, past_key_values)
            past_seen_tokens = past_key_values.get_seq_length()

    if cache_position is None:
        if isinstance(past_key_values, StaticCache):
            raise ValueError("cache_position is a required argument when using StaticCache.")
        cache_position = torch.arange(
            past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
        )

    if position_ids is None:
        position_ids = cache_position.unsqueeze(0)

    causal_mask = self._update_causal_mask(attention_mask, inputs_embeds, cache_position)

    # embed positions
    hidden_states = inputs_embeds

    # decoder layers
    all_hidden_states = () if output_hidden_states else None
    all_self_attns = () if output_attentions else None
    next_decoder_cache = None

    for decoder_layer in self.layers:
        if output_hidden_states:
            all_hidden_states += (hidden_states,)

        if self.gradient_checkpointing and self.training:
            layer_outputs = self._gradient_checkpointing_func(
                decoder_layer.__call__,
                hidden_states,
                causal_mask,
                position_ids,
                past_key_values,
                output_attentions,
                use_cache,
                cache_position,
            )
        else:
            layer_outputs = decoder_layer(
                hidden_states,
                attention_mask=causal_mask,
                position_ids=position_ids,
                past_key_value=past_key_values,
                output_attentions=output_attentions,
                use_cache=use_cache,
                cache_position=cache_position,
            )

        hidden_states = layer_outputs[0]

        if use_cache:
            next_decoder_cache = layer_outputs[2 if output_attentions else 1]

        if output_attentions:
            all_self_attns += (layer_outputs[1],)

    hidden_states = self.norm(hidden_states)

    # add hidden states from the last decoder layer
    if output_hidden_states:
        all_hidden_states += (hidden_states,)

    next_cache = None
    if use_cache:
        next_cache = (
            next_decoder_cache.to_legacy_cache() if isinstance(next_decoder_cache, Cache) else next_decoder_cache
        )
    if not return_dict:
        return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
    
    return BaseModelOutputWithPast(
        last_hidden_state=hidden_states,
        past_key_values=next_cache,
        hidden_states=all_hidden_states,
        attentions=all_self_attns,
    )

class H2OLlamaForCausalLM(LlamaForCausalLM):
    def __init__(self, config):
        super().__init__(config)
        num_layers = len(self.model.layers)
        for layer_idx in range(num_layers):
            self.model.layers[layer_idx].self_attn = H2OLlamaAttention(config, layer_idx)

        self.model.forward = types.MethodType(enable_h2ocache_forward, self.model)
        self.model.num_heavy_hitter_tokens = config.num_heavy_hitter_tokens
        self.model.num_window_length = config.num_window_length
    
    def prepare_inputs_for_generation(
        self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, cache_position=None, **kwargs
    ):
        # With static cache, the `past_key_values` is None
        # TODO joao: standardize interface for the different Cache classes and remove of this if

        has_static_cache = False
        if past_key_values is None:
            past_key_values = getattr(self.model.layers[0].self_attn, "past_key_value", None)
            has_static_cache = past_key_values is not None

        past_length = 0
        if past_key_values is not None:
            if isinstance(past_key_values, Cache):
                past_length = cache_position[0]
                max_cache_length = (
                    torch.tensor(past_key_values.get_max_length(), device=input_ids.device)
                    if past_key_values.get_max_length() is not None
                    else None
                )
                cache_length = past_key_values.get_seq_length()

            # TODO joao: remove this `else` after `generate` prioritizes `Cache` objects
            else:
                past_length = cache_position[0]
                cache_length = past_key_values[0].shape[2] # length = num_layers * 3 (3 -> key, value, score)
                max_cache_length = None

            # Keep only the unprocessed tokens:
            # 1 - If the length of the attention_mask exceeds the length of input_ids, then we are in a setting where
            # some of the inputs are exclusively passed as part of the cache (e.g. when passing input_embeds as
            # input)
            if attention_mask is not None and attention_mask.shape[1] > input_ids.shape[1]:
                input_ids = input_ids[:, -(attention_mask.shape[1] - past_length) :]
            # 2 - If the past_length is smaller than input_ids', then input_ids holds all input tokens. We can discard
            # input_ids based on the past_length.
            elif past_length < input_ids.shape[1]:
                input_ids = input_ids[:, past_length:]
            # 3 - Otherwise (past_length >= input_ids.shape[1]), let's assume input_ids only has unprocessed tokens.

            # If we are about to go beyond the maximum cache length, we need to crop the input attention mask.
            if (
                max_cache_length is not None
                and attention_mask is not None
                and cache_length + input_ids.shape[1] > max_cache_length
            ):
                attention_mask = attention_mask[:, -max_cache_length:]

        position_ids = kwargs.get("position_ids", None)
        if attention_mask is not None and position_ids is None:
            # create position_ids on the fly for batch generation
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            if past_key_values:
                position_ids = position_ids[:, -input_ids.shape[1] :]

        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
        if inputs_embeds is not None and past_key_values is None:
            model_inputs = {"inputs_embeds": inputs_embeds}
        else:
            # The `contiguous()` here is necessary to have a static stride during decoding. torchdynamo otherwise
            # recompiles graphs as the stride of the inputs is a guard. Ref: https://github.com/huggingface/transformers/pull/29114
            # TODO: use `next_tokens` directly instead.
            model_inputs = {"input_ids": input_ids.contiguous()}

        input_length = position_ids.shape[-1] if position_ids is not None else input_ids.shape[-1]
        if cache_position is None:
            cache_position = torch.arange(past_length, past_length + input_length, device=input_ids.device)
        else:
            cache_position = cache_position[-input_length:]

        if has_static_cache:
            past_key_values = None

        model_inputs.update(
            {
                "position_ids": position_ids,
                "cache_position": cache_position,
                "past_key_values": past_key_values,
                "use_cache": kwargs.get("use_cache"),
                "attention_mask": attention_mask,
            }
        )
        return model_inputs