phi2_model.py

import torch
import torch.nn as nn
from transformers import PreTrainedModel
from transformers.modeling_outputs import CausalLMOutputWithPast
from typing import Any, cast

from .attention import ParallelAttentionBlock, KVCache
from .phi2_configuration import Phi2Config


class Phi2PreTrainedModel(PreTrainedModel):
    config_class = Phi2Config  # not necessary unless you want to register model with auto classes
    supports_gradient_checkpointing = False
    # _no_split_modules = ["ParallelAttentionBlock"]

    def __init__(self, config: Phi2Config):
        super().__init__(config)
        self.config = config

    def _init_weights(self, module: nn.Module) -> None:
        # initialize weights - will get overwritten by saved weights in from_pretrained() if they exist
        if isinstance(module, (nn.Linear,)):
            module.weight.data.normal_(mean=0.0, std=self.config.weight_initialization_range)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=self.config.weight_initialization_range)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()
        elif isinstance(module, nn.LayerNorm):
            if module.bias is not None:
                module.bias.data.zero_()
            module.weight.data.fill_(1.0)

    def prepare_inputs_for_generation(
        self,
        input_ids: torch.LongTensor,  # dim: (batch_size, seq_len)
        past_key_values: KVCache | None = None,  # has to be named this
        key_padding_mask: torch.LongTensor | torch.BoolTensor | None = None,
        **kwargs,  # has to be here
    ) -> dict[str, Any]:
        kv_cache = past_key_values
        if not kv_cache:
            kv_cache = KVCache(
                max_seqlen=self.config.initial_cos_sin_cache_len,
                max_batch_size=input_ids.shape[0],
                seqlen_offset=0,
                batch_size_offset=0,
                kv_block_map={},
                lengths_per_sample=None,
            )
        else:
            # assume that `kv_cache` has cached all tokens up to the last token in `input_ids`
            kv_cache.seqlen_offset = input_ids.shape[1] - 1
            input_ids = cast(torch.LongTensor, input_ids[:, -1].unsqueeze(-1))

        return {  # to be passed to forward()
            "input_ids": input_ids,
            "kv_cache": kv_cache,
            "key_padding_mask": key_padding_mask,
        }


class Embedding(nn.Module):
    """Token embedding with dropout."""

    def __init__(
        self,
        vocab_size: int,
        d_embedding: int,
        embd_pdrop: float,
    ) -> None:
        super().__init__()
        self.embeddings = nn.Embedding(vocab_size, d_embedding)
        self.dropout = nn.Dropout(embd_pdrop)

    def forward(
        self,
        input_ids: torch.LongTensor,  # dim: (batch_size, seq_len)
    ) -> torch.FloatTensor:
        x = self.embeddings(  # dim: (batch_size, seq_len, d_embedding)
            input_ids.view(-1, input_ids.size()[-1])
        )
        x = self.dropout(x)
        return x


class Phi2Model(Phi2PreTrainedModel):
    def __init__(self, config: Phi2Config) -> None:
        super().__init__(config)
        self.embedding = Embedding(
            vocab_size=config.vocab_size,
            d_embedding=config.d_embedding,
            embd_pdrop=config.embd_pdrop,
        )
        self.parallel_blocks = nn.ModuleList([
            ParallelAttentionBlock(
                resid_pdrop=config.resid_pdrop,
                layer_norm_epsilon=config.layer_norm_epsilon,
                d_embedding=config.d_embedding,
                n_attn_heads=config.n_attn_heads,
                block_n=i,
                initial_cos_sin_cache_len=config.initial_cos_sin_cache_len,
                attn_pdrop=config.attn_pdrop,
                use_flash_rotary=config.use_flash_rotary,
                use_flash_attn=config.use_flash_attn,
                use_fused_dense=config.use_fused_dense,
                checkpointing=config.checkpointing,
            )
            for i in range(config.n_attn_blocks)
        ])
        self.gradient_checkpointing_disable()  # https://github.com/cybertronai/gradient-checkpointing - I think this is turned off due to flash attention?
        self.post_init()  # calls self._init_weights() for all modules

    """
    def get_input_embeddings(self) -> nn.Embedding:
        return self.embedding.embeddings

    def set_input_embeddings(self, new_embeddings: nn.Embedding) -> None:
        self.embedding.embeddings = new_embeddings
    """

    def forward(
        self,
        input_ids: torch.LongTensor,
        kv_cache: KVCache | None = None,
        key_padding_mask: torch.BoolTensor | None = None,
    ) -> torch.FloatTensor:
        x = self.embedding(input_ids)
        for block in self.parallel_blocks:
            x = block(
                x,
                kv_cache=kv_cache,
                key_padding_mask=key_padding_mask,
            )
        return x


class Phi2ModelForCausalLM(Phi2PreTrainedModel):
    def __init__(self, config: Phi2Config) -> None:
        super().__init__(config)
        self.model = Phi2Model(config)
        self.lm_head_layer_norm = nn.LayerNorm(config.d_embedding, eps=config.layer_norm_epsilon)
        self.lm_head_linear = nn.Linear(config.d_embedding, config.vocab_size)
        self.loss_fn = nn.CrossEntropyLoss()
        self.post_init()  # calls self._init_weights() for all modules

    def forward(
        self,
        input_ids: torch.LongTensor,
        kv_cache: KVCache | None = None,
        key_padding_mask: torch.BoolTensor | None = None,
        labels: torch.LongTensor | None = None,
        **kwargs,  # has to be here
    ) -> CausalLMOutputWithPast:
        x = self.model(input_ids, kv_cache=kv_cache, key_padding_mask=key_padding_mask)
        x = self.lm_head_layer_norm(x)
        logits = self.lm_head_linear(x).to(torch.float32)
        loss = (
            self.loss_fn(logits.view(-1, logits.size(-1)), labels.view(-1))
            if labels is not None
            else None
        )
        return CausalLMOutputWithPast(loss=loss, logits=logits, past_key_values=kv_cache)