ultravox_model.py

import logging
import re
from typing import Any, Dict, Generator, Optional, Set, Tuple, TypeVar, Union

import accelerate
import peft
import torch
import torch.nn as nn
import torch.nn.functional as F
import transformers
import transformers.activations
import transformers.modeling_outputs
import transformers.models
from transformers.generation.utils import GenerationMixin
from transformers.models.whisper import modeling_whisper as whisper

# We must use relative import in this directory to allow uploading to HF Hub
# Even "from . import X" pattern doesn't work (undocumented and unclear why)
from .ultravox_config import LossConfig
from .ultravox_config import LossFunction
from .ultravox_config import UltravoxConfig

FROM_PRETRAINED_KWARGS = {}
SHARED_PRETRAINED_KWARGS = [
    "tp_plan",
    "device_map",
    "torch_dtype",
    "attn_implementation",
    "use_flash_attention_2",
]


class UltravoxModel(transformers.LlamaPreTrainedModel, GenerationMixin):
    """
    The Ultravox model which consists of an audio encoder and a language model.

    Audio input is processed by the audio encoder, then every `stack_factor` frames are stacked together and
    projected to the language model's embedding space using a few linear layers.
    The text is embedded by the language model as usual and then the audio and text embeddings are merged together.

    A special token `<|audio|>` is used to indicate the start of the audio embeddings in the merged embeddings.

    Parameters:
        config: Model configuration class with all the parameters of the model.
    """

    config_class = UltravoxConfig
    config: UltravoxConfig  # for type hinting
    # Usually we load encoder and LLM weights from a pretrained model separately, so they are allowed to be missing
    _keys_to_ignore_on_load_missing = ["audio_tower.*", "language_model.*"]
    # Since we have kwargs in forward, we need to set this to False, otherwise grad_accum_steps will cause incorrect train loss to be reported
    # see https://github.com/huggingface/transformers/issues/35856 and https://github.com/huggingface/trl/pull/2615/files
    accepts_loss_kwargs = False

    def __init__(self, config: UltravoxConfig):
        super().__init__(config)
        self._register_load_state_dict_pre_hook(self._pre_load_state_dict_hook)

        self.keep_params: Set[str] = set()
        self.vocab_size = config.vocab_size

        if not config.llm_only_training:
            self.audio_tower = self._create_audio_tower(config)
            self.multi_modal_projector = self._create_multi_modal_projector(config)
            self.audio_tower_context_length = self.audio_tower.max_context_length

        self.language_model = self._create_language_model(config)

        if self.language_model._tied_weights_keys is not None:
            self._tied_weights_keys = [
                f"language_model.{k}" for k in self.language_model._tied_weights_keys
            ]

        # Determine no_split_modules dynamically to use with FSDP auto_wrap policy.
        # This would be something like ["LlamaDecoderLayer"] as we don't split audio encoder layers.
        # FSDP throws an error if some of the layer types are not found in the model, and they need to be filted out.
        # 1. Get the names the language model *wants* to keep intact
        candidate_names = set(
            getattr(self.language_model, "_no_split_modules", []) or []
        )
        # 2. Names that actually exist in the current model
        present_names = {m.__class__.__name__ for m in self.modules()}
        # 3. Keep only those that are both requested and present
        self._no_split_modules = list(candidate_names & present_names)

        self.loss_config = LossConfig()
        self.post_init()

    def _init_weights(self, module):
        if module is self:
            if self.config.text_model_id is not None:
                self.language_model = self._create_language_model(self.config)
            if self.config.audio_model_id is not None:
                self.audio_tower = self._create_audio_tower(self.config)
        elif module in self.language_model.modules():
            pass
        elif module in self.audio_tower.modules():
            pass
        else:
            super()._init_weights(module)

    @classmethod
    def from_pretrained(cls, *args, **kwargs):
        global FROM_PRETRAINED_KWARGS
        FROM_PRETRAINED_KWARGS = {
            k: v for k, v in kwargs.items() if k in SHARED_PRETRAINED_KWARGS
        }
        model = super().from_pretrained(*args, **kwargs)
        FROM_PRETRAINED_KWARGS = {}
        return model

    def get_input_embeddings(self):
        return self.language_model.get_input_embeddings()

    def set_input_embeddings(self, value):
        self.language_model.set_input_embeddings(value)

    def get_output_embeddings(self):
        return self.language_model.get_output_embeddings()

    def set_output_embeddings(self, new_embeddings):
        self.language_model.set_output_embeddings(new_embeddings)

    def set_decoder(self, decoder):
        self.language_model.set_decoder(decoder)

    def get_decoder(self):
        return self.language_model.get_decoder()

    def tie_weights(self):
        return self.language_model.tie_weights()

    def set_loss_config(self, loss_config: LossConfig):
        self.loss_config = loss_config

    def _setup_cache(
        self, cache_cls, max_batch_size: int, max_cache_len: Optional[int] = None
    ):
        self.language_model._setup_cache(cache_cls, max_batch_size, max_cache_len)

    def _reorder_cache(self, past_key_values, beam_idx):
        return self.language_model._reorder_cache(past_key_values, beam_idx)

    def resize_token_embeddings(
        self,
        new_num_tokens: Optional[int] = None,
        pad_to_multiple_of: Optional[int] = None,
    ) -> nn.Embedding:
        model_embeds = self.language_model.resize_token_embeddings(
            new_num_tokens, pad_to_multiple_of
        )
        # update vocab size
        self.config.text_config.vocab_size = model_embeds.num_embeddings
        self.config.vocab_size = model_embeds.num_embeddings
        self.vocab_size = model_embeds.num_embeddings
        return model_embeds

    def _get_prediction_mask(
        self, labels: Optional[torch.Tensor]
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        """Get boolean masks for positions where we want to compute KL divergence.

        For each label position, we want the position before it since that's where
        the model makes the prediction for that label.

        Additionally, we want to identify the position right before the EOT token
        (the last token with label != -100).

        Args:
            labels: Tensor of shape (B, T) where B is batch size and T is sequence length,
                   with -100 for masked positions and token ids for label positions

        Returns:
            Tuple containing:
            - pred_mask: Boolean tensor of shape (B, T) that's True for positions where we want to compute KL divergence
            - eot_mask: Boolean tensor of shape (B, T) that's True only for the last prediction position in each sequence
        """
        if labels is None:
            raise ValueError("labels must be provided")

        # Shift the label mask right by 1 along the sequence dimension
        # This gives us positions where we make predictions for the next token
        label_mask = labels != -100
        pred_mask = torch.zeros_like(label_mask)
        pred_mask[:, :-1] = label_mask[
            :, 1:
        ]  # shift right by 1 along sequence dimension

        # Create EOT mask - identify only the last prediction position in each sequence
        eot_mask = torch.zeros_like(pred_mask)
        batch_size = labels.shape[0]

        for i in range(batch_size):
            # Find positions where we make predictions
            pred_positions = torch.where(pred_mask[i])[0]
            if len(pred_positions) > 0:
                # Only mark the last prediction position
                eot_mask[i, pred_positions[-1]] = True

        return pred_mask, eot_mask

    def _compute_kl_loss(
        self,
        lm_output: transformers.modeling_outputs.CausalLMOutputWithPast,
        labels: Optional[torch.Tensor] = None,
        past_key_values: Optional[Union[Tuple, transformers.cache_utils.Cache]] = None,
        alt_input_ids: Optional[torch.Tensor] = None,
        alt_attention_mask: Optional[torch.Tensor] = None,
        alt_labels: Optional[torch.Tensor] = None,
        **kwargs,
    ):
        # disable gradient computation for the teacher model
        with torch.no_grad():
            # compute the teacher (text-only) model's distribution
            alt_inputs_embeds = self.get_input_embeddings().forward(alt_input_ids)
            alt_lm_output = self.language_model.forward(
                inputs_embeds=alt_inputs_embeds,
                labels=alt_labels,
                attention_mask=alt_attention_mask,
                past_key_values=past_key_values,
                **kwargs,
            )

        # Get prediction masks for regular tokens and EOT tokens
        pred_mask, eot_mask = self._get_prediction_mask(labels)
        alt_pred_mask, alt_eot_mask = self._get_prediction_mask(alt_labels)

        # compute the KL divergence loss between the two models for regular tokens
        kl_loss = F.kl_div(
            F.log_softmax(
                lm_output.logits[pred_mask] / self.loss_config.kl_temperature,
                dim=-1,
            ),
            F.softmax(
                alt_lm_output.logits[alt_pred_mask] / self.loss_config.kl_temperature,
                dim=-1,
            ),
            reduction="batchmean",
        )

        # Compute the KL divergence loss for EOT token positions if any exist
        if self.loss_config.eot_loss_weight > 0:
            eot_loss = F.kl_div(
                F.log_softmax(
                    lm_output.logits[eot_mask] / self.loss_config.kl_temperature,
                    dim=-1,
                ),
                F.softmax(
                    alt_lm_output.logits[alt_eot_mask]
                    / self.loss_config.kl_temperature,
                    dim=-1,
                ),
                reduction="batchmean",
            )
            kl_loss += self.loss_config.eot_loss_weight * eot_loss

        return kl_loss

    def _audio_iter(
        self, audio_batch_size: torch.Tensor
    ) -> Generator[Tuple[int, int], None, None]:
        """
        Iterate over the audio batch size and yield the batch index and audio index of each audio item.

        Args:
            audio_batch_size: A tensor of shape (B,) where B is the batch size.

        Returns:
            A generator that yields a tuple of (start index, length) for each audio item.
        """
        audio_index = 0
        for i_b, batch_count in enumerate(audio_batch_size):
            for _ in range(batch_count):
                yield i_b, audio_index
                audio_index += 1

    def forward(
        self,
        input_ids: torch.Tensor,
        audio_values: Optional[torch.Tensor] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
        labels: Optional[torch.Tensor] = None,
        attention_mask: Optional[torch.Tensor] = None,
        audio_token_start_idx: Optional[torch.Tensor] = None,
        audio_lens: Optional[torch.Tensor] = None,
        audio_token_len: Optional[torch.Tensor] = None,
        audio_batch_size: Optional[torch.Tensor] = None,
        past_key_values: Optional[Union[Tuple, transformers.cache_utils.Cache]] = None,
        # the alt_* fields are needed for KL divergence loss
        alt_input_ids: Optional[torch.Tensor] = None,
        alt_attention_mask: Optional[torch.Tensor] = None,
        alt_labels: Optional[torch.Tensor] = None,
        **kwargs,
    ) -> transformers.modeling_outputs.CausalLMOutputWithPast:
        """
        Forward pass for the Ultravox model.

        `input_ids` are the tokenized text input. They are embedded by the language model as usual.
        `audio_values` are processed by the audio encoder and then every `stack_factor` frames are stacked together and
        projected to the language model's embedding space using a few linear layers.
        The audio and text embeddings are merged together. A special token `<|audio|>` is used to indicate the start
        of the audio embeddings in the merged embeddings.

        Args:
            input_ids: The tokenized text input.
            audio_values: The processed audio values.
            inputs_embeds: The embeddings for the input tokens.
            labels: The tokenized text labels.
            attention_mask: The attention mask for the input.
            position_ids: The position ids for the input.
            past_key_values: The past key value cache for the language model attention layers.
            **kwargs: Additional keyword arguments. Passed directly to the language model.
        """
        if inputs_embeds is None:
            # B x T  ->  B x T x D
            inputs_embeds = self.get_input_embeddings().forward(input_ids)

        if audio_values is not None and len(audio_values) > 0:
            inputs_embeds = self._prepare_audio_embeds(
                inputs_embeds=inputs_embeds,
                audio_values=audio_values,
                audio_token_start_idx=audio_token_start_idx,
                audio_lens=audio_lens,
                audio_token_len=audio_token_len,
                audio_batch_size=audio_batch_size,
            )

        lm_output = self.language_model.forward(
            inputs_embeds=inputs_embeds,
            labels=labels,
            attention_mask=attention_mask,
            past_key_values=past_key_values,
            **kwargs,
        )
        if self.training:
            if self.loss_config.loss_function == LossFunction.CrossEntropy:
                pass
            elif self.loss_config.loss_function == LossFunction.KL_Divergence:
                lm_output.loss = self._compute_kl_loss(
                    lm_output=lm_output,
                    labels=labels,
                    past_key_values=past_key_values,
                    alt_input_ids=alt_input_ids,
                    alt_attention_mask=alt_attention_mask,
                    alt_labels=alt_labels,
                    **kwargs,
                )
            else:
                raise ValueError(
                    f"Unsupported loss function: {self.loss_config.loss_function}"
                )
        return lm_output

    def _prepare_audio_embeds(
        self,
        inputs_embeds: Optional[torch.Tensor] = None,
        audio_values: Optional[torch.Tensor] = None,
        audio_token_start_idx: Optional[torch.Tensor] = None,
        audio_lens: Optional[torch.Tensor] = None,
        audio_token_len: Optional[torch.Tensor] = None,
        audio_batch_size: Optional[torch.Tensor] = None,
    ) -> torch.Tensor:
        assert (
            inputs_embeds is not None
            and audio_values is not None
            and audio_token_start_idx is not None
            and audio_token_len is not None
            and audio_lens is not None
            and audio_batch_size is not None
        ), "inputs_embeds/audio_values/audio_token_start_idx/audio_token_len/audio_lens/audio_batch_size must be provided."
        assert (
            len(audio_token_start_idx)
            == len(audio_token_len)
            == len(audio_lens)
            == len(audio_values)
        ), "audio_token_start_idx/audio_token_len/audio_lens/audio_values must have the same batch size."
        assert len(audio_batch_size) == len(
            inputs_embeds
        ), "audio_batch_size and inputs_embeds must have the same batch size."

        # B x A/3200 x (D=max-audio-length-in-batch)
        audio_tower_output = self.audio_tower.forward(
            audio_values.to(self.audio_tower.dtype),
            audio_len=audio_lens,
        ).last_hidden_state
        audio_tower_output = audio_tower_output.to(inputs_embeds.dtype)
        audio_embeds = self.multi_modal_projector.forward(audio_tower_output)

        # combine audio and text embeddings
        for i_b, i_a in self._audio_iter(audio_batch_size):
            start_idx = audio_token_start_idx[i_a]
            token_len = audio_token_len[i_a]
            item_embedding = audio_embeds[i_a][:token_len]
            inputs_embeds[i_b][start_idx : start_idx + token_len] = item_embedding

        return inputs_embeds

    def generate(
        self,
        input_ids: torch.Tensor,
        audio_values: Optional[torch.Tensor] = None,
        inputs_embeds: Optional[torch.Tensor] = None,
        audio_token_start_idx: Optional[torch.Tensor] = None,
        audio_lens: Optional[torch.Tensor] = None,
        audio_token_len: Optional[torch.Tensor] = None,
        audio_batch_size: Optional[torch.Tensor] = None,
        **kwargs,
    ) -> torch.Tensor:
        if inputs_embeds is None:
            inputs_embeds = self.get_input_embeddings().forward(input_ids)

        if audio_values is not None and len(audio_values) > 0:
            inputs_embeds = self._prepare_audio_embeds(
                inputs_embeds=inputs_embeds,
                audio_values=audio_values,
                audio_token_start_idx=audio_token_start_idx,
                audio_lens=audio_lens,
                audio_token_len=audio_token_len,
                audio_batch_size=audio_batch_size,
            )

        return self.language_model.generate(
            input_ids=input_ids,
            inputs_embeds=inputs_embeds,
            **kwargs,
        )

    @classmethod
    def _create_multi_modal_projector(
        cls, config: UltravoxConfig
    ) -> "UltravoxProjector":
        projector = UltravoxProjector(config)
        dtype = config.torch_dtype
        if isinstance(dtype, str):
            dtype = getattr(torch, dtype)
        projector.to(dtype)
        return projector

    @classmethod
    def _create_audio_tower(
        cls, config: UltravoxConfig
    ) -> Union[transformers.Wav2Vec2Model, "ModifiedWhisperEncoder"]:
        # We probably don't want to pass tp_plan or device_map to the audio tower
        # But potentially other kwargs can be passed in. TODO
        kwargs = {"torch_dtype": config.torch_dtype}
        if (
            transformers.modeling_utils._init_weights
            and config.audio_model_id is not None
        ):
            if "whisper" in config.audio_model_id.lower():
                audio_tower = ModifiedWhisperEncoder.from_pretrained(
                    config.audio_model_id, **kwargs
                )
                audio_tower.init_latency_mask(
                    config.audio_latency_block_size, dtype=config.torch_dtype
                )
            else:
                assert config.audio_latency_block_size in (
                    None,
                    0,
                ), "only whisper audio tower supports audio latency masking, got non-zero value for 'audio_latency_block_size'"
                audio_tower = transformers.AutoModel.from_pretrained(
                    config.audio_model_id, **kwargs
                )
        else:
            with accelerate.init_empty_weights():
                if "whisper" in config.audio_config._name_or_path.lower():
                    audio_tower = ModifiedWhisperEncoder(config.audio_config)
                    audio_tower.init_latency_mask(
                        config.audio_latency_block_size,
                        dtype=config.torch_dtype,
                    )
                else:
                    assert config.audio_latency_block_size in (
                        None,
                        0,
                    ), "only whisper audio tower supports audio latency masking, got non-zero value for 'audio_latency_block_size'"
                    # we only ever use from_config if the weights are retrained, hence initializing is not
                    # required. This makes the model quite creation faster since init on CPU is quite slow.
                    audio_tower = transformers.AutoModel.from_config(
                        config.audio_config, **kwargs
                    )

        if isinstance(
            audio_tower,
            (transformers.Wav2Vec2BertModel, transformers.WhisperModel),
        ):
            # For these models we only need the encoder part
            # Wav2Vec2BertModel -> Wav2Vec2BertEncoder
            # WhisperModel -> WhisperEncoder
            audio_tower = audio_tower.encoder

        audio_tower = apply_lora(audio_tower, config.audio_model_lora_config)
        return audio_tower

    @classmethod
    def _create_language_model(
        cls, config: UltravoxConfig
    ) -> transformers.LlamaForCausalLM:
        if (
            transformers.modeling_utils._init_weights
            and config.text_model_id is not None
        ):
            language_model = transformers.AutoModelForCausalLM.from_pretrained(
                config.text_model_id,
                **{
                    "attn_implementation": config.text_config._attn_implementation,
                    "torch_dtype": config.torch_dtype,
                    **FROM_PRETRAINED_KWARGS,
                },
            )
        else:
            with accelerate.init_empty_weights():
                # we only ever use from_config if the weights are retrained, hence initializing is not
                # required. This makes the model quite creation faster since init on CPU is quite slow.
                language_model = transformers.AutoModelForCausalLM.from_config(
                    config.text_config,
                    attn_implementation=config.text_config._attn_implementation,
                    torch_dtype=config.torch_dtype,
                )

        language_model = apply_lora(language_model, config.text_model_lora_config)
        return language_model

    def merge_and_unload(self):
        if isinstance(self.language_model, peft.PeftModel):
            self.language_model = self.language_model.merge_and_unload()
            # no need to download base language model weights anymore, so we can remove the id
            self.config.text_model_id = None
            self.keep_params.update(
                set(
                    [
                        f"language_model.{name}"
                        for name, _ in self.language_model.named_parameters()
                    ]
                )
            )

        if hasattr(self, "audio_tower") and isinstance(
            self.audio_tower, peft.PeftModel
        ):
            self.audio_tower = self.audio_tower.merge_and_unload()
            # no need to download base audio model weights anymore, so we can remove the id
            self.config.audio_model_id = None
            self.keep_params.update(
                set(
                    [
                        f"audio_tower.{name}"
                        for name, _ in self.audio_tower.named_parameters()
                    ]
                )
            )

        for param in ["text_model_lora_config", "audio_model_lora_config"]:
            if hasattr(self.config, param):
                delattr(self.config, param)

    def push_to_hub(self, *args, **kwargs):
        self.merge_and_unload()
        return super().push_to_hub(*args, **kwargs)

    def diff_state_dict(
        self, state_dict: Optional[Dict[str, Any]] = None
    ) -> Dict[str, Any]:
        if state_dict is None:
            state_dict = super().state_dict()

        trainable_params = {k for k, v in self.named_parameters() if v.requires_grad}
        # normalize the keys to match the original model
        # Example: audio_tower.base_model.model.layers.0._fsdp_wrapped_module.self_attn.k_proj.lora_B.default.weight
        trainable_params = {
            k.replace("_fsdp_wrapped_module.", "") for k in trainable_params
        }

        state_dict = {
            k: v
            for k, v in state_dict.items()
            if k in self.keep_params or k in trainable_params
        }

        return state_dict

    def save_pretrained(
        self, *args, state_dict: Optional[Dict[str, Any]] = None, **kwargs
    ):
        state_dict = self.diff_state_dict(state_dict)

        super().save_pretrained(*args, state_dict=state_dict, **kwargs)

    def _pre_load_state_dict_hook(self, state_dict: Dict[str, Any], *args, **kwargs):
        self.keep_params.update(set(state_dict.keys()))

    def print_trainable_parameters(self):
        """
        Prints the number of trainable parameters in the model (reuses Peft model's method)
        """
        count_params = peft.peft_model.PeftModel.get_nb_trainable_parameters

        trainable_params, all_param = count_params(self)

        logging.info(
            f"trainable params: {trainable_params:,d} || all params: {all_param:,d}"
            f" || trainable%: {100 * trainable_params / all_param:.1f}%"
        )

        lm_trainable_params, lm_all_params = count_params(self.language_model)
        if hasattr(self, "audio_tower") and self.audio_tower is not None:
            audio_trainable_params, audio_all_params = count_params(self.audio_tower)
        else:
            audio_trainable_params, audio_all_params = 0, 0

        projector_trainable_params = (
            trainable_params - lm_trainable_params - audio_trainable_params
        )
        projector_all_params = all_param - lm_all_params - audio_all_params

        # Calculate percentages only if the total parameters are non-zero
        audio_percent = (
            0.0
            if audio_all_params == 0
            else 100 * audio_trainable_params / audio_all_params
        )
        projector_percent = (
            0.0
            if projector_all_params == 0
            else 100 * projector_trainable_params / projector_all_params
        )

        logging.info(
            f"Trainable%:   "
            f" LLM: {100 * lm_trainable_params / lm_all_params:.1f}%"
            f" || Audio Encoder: {audio_percent:.1f}%"
            f" || Projector: {projector_percent:.1f}%"
        )


def get_checkpoint_files(
    model_id: str,
) -> tuple[list[str], dict | None, list[str]]:
    resolved_archive_file = transformers.utils.cached_file(
        model_id,
        transformers.utils.SAFE_WEIGHTS_NAME,
        _raise_exceptions_for_missing_entries=False,
    )

    if resolved_archive_file is not None:
        # not sharded
        sharded_metadata = None
        state_dict = transformers.modeling_utils.load_state_dict(resolved_archive_file)
        loaded_state_dict_keys = list(state_dict.keys())
    else:
        # sharded
        resolved_archive_file = transformers.utils.cached_file(
            model_id, transformers.utils.SAFE_WEIGHTS_INDEX_NAME
        )
        resolved_archive_file, sharded_metadata = (
            transformers.modeling_utils.get_checkpoint_shard_files(
                model_id,
                resolved_archive_file,
            )
        )
        loaded_state_dict_keys = sharded_metadata["all_checkpoint_keys"]

    if isinstance(resolved_archive_file, str):
        resolved_archive_file = [resolved_archive_file]

    return resolved_archive_file, sharded_metadata, loaded_state_dict_keys


# TODO: refactor common parts to a shared module
def is_cache_empty(
    past_key_values: Optional[Union[Tuple, transformers.cache_utils.Cache]],
) -> bool:
    """
    Check if the cache is empty.
    """
    if past_key_values is None:
        return True
    if isinstance(past_key_values, tuple):
        return all(len(c) == 0 for c in past_key_values)
    return past_key_values.get_seq_length() == 0


T = TypeVar("T", bound=torch.nn.Module)


def apply_lora(model: T, lora_config: dict) -> T:
    """
    Applies LoRA finetuning to the model. If the `r` parameter is set to 0, the model is frozen instead.
    """
    unfreeze_layers = lora_config.pop("unfreeze_layers", None)
    lora_config = peft.LoraConfig(**lora_config or {})

    if lora_config.r == 0:
        # freeze the model entirely, except for the specified layers
        for name, param in model.named_parameters():
            if not unfreeze_layers or not any(
                re.match(layer, name) for layer in unfreeze_layers
            ):
                param.requires_grad = False
            else:
                logging.info(f"Unfreezing layer: {name} with #{param.numel()} params")
    else:
        model = peft.get_peft_model(model, lora_config)

    return model


class StackAudioFrames(nn.Module):
    """
    Stack the audio embedding frames to reduce the sequence length by a factor
    of `stack_factor`.
    """

    def __init__(self, stack_factor: int = 8):
        super().__init__()
        self.stack_factor = stack_factor

    def forward(self, audio_embeds: torch.Tensor) -> torch.Tensor:
        B, T, C = audio_embeds.shape
        T_pad = (T + self.stack_factor - 1) // self.stack_factor * self.stack_factor
        audio_embeds = F.pad(audio_embeds, (0, 0, 0, T_pad - T))
        B, T, C = audio_embeds.shape
        audio_embeds = audio_embeds.view(
            B, T // self.stack_factor, C * self.stack_factor
        )
        return audio_embeds


class RMSNorm(transformers.models.llama.modeling_llama.LlamaRMSNorm):
    def __init__(self, hidden_size: int, init: float = 1, eps: float = 1e-6):
        super().__init__(hidden_size=hidden_size, eps=eps)
        self.weight.data.fill_(init)


class SwiGLU(nn.Module):
    def forward(self, x):
        x, gate = x.chunk(2, dim=-1)
        return F.silu(gate) * x


class UltravoxProjector(nn.Module):
    def __init__(self, config: UltravoxConfig):
        super().__init__()
        self.hidden_dim = config.hidden_size
        self._pad_and_stack = StackAudioFrames(config.stack_factor)
        dim_in = config.audio_config.hidden_size * config.stack_factor
        self.ln_pre = RMSNorm(dim_in, init=config.norm_init)
        self.linear_1 = nn.Linear(dim_in, self.hidden_dim, bias=False)
        dim_mid = self.hidden_dim
        self.act = transformers.activations.get_activation(config.projector_act)
        dim_mid = dim_mid // 2 if config.projector_act == "swiglu" else dim_mid
        dim_out = config.text_config.hidden_size
        self.linear_2 = nn.Linear(dim_mid, dim_out, bias=False)

        # Ultravox v0.4.1 and below uses layer_norm after the second linear layer,
        # while v0.5.0 and above uses layer_norm after the first linear layer.
        if config.projector_ln_mid:
            self.ln_mid: nn.Module = RMSNorm(dim_mid, init=config.norm_init)
            self.ln_post: nn.Module = nn.Identity()
        else:
            self.ln_mid = nn.Identity()
            self.ln_post = RMSNorm(dim_out, init=config.norm_init)

    def forward(self, audio_features: torch.Tensor) -> torch.Tensor:
        """
        Takes in audio features from the audio tower and projects them to the text model's embedding space.
        It reduces the number of frames by a factor of `stack_factor` and increases the number of channels by the same factor.
        If the number of audio frames are not a multiple of the stack factor, the last few frames will be padded with zeros.

        Input shape:
            audio_features: B, T*S, C
        Output shape:
            hidden_states: B, T, D
        Where:
            B: batch size
            F: number of frames in the audio tower
            T: number of output embeddings
                T = ceil(F / S)
            S: stack factor
            C: number of channels out of the encoder (aka audio tower)
            H: hidden size of the projector (config.hidden_size)
            D: dimension of the text model (config.text_config.hidden_size)

        """
        # B, F, C -> B, T, C*S
        audio_features = self._pad_and_stack(audio_features)
        audio_features = self.ln_pre(audio_features)
        # B, T, C*S -> B, T, H
        hidden_states = self.linear_1(audio_features)
        # B, T, H -> B, T, H/2 (assuming swiglu)
        hidden_states = self.act(hidden_states)
        hidden_states = self.ln_mid(hidden_states)
        # B, T, H/2 -> B, T, D
        hidden_states = self.linear_2(hidden_states)
        hidden_states = self.ln_post(hidden_states)
        return hidden_states


class ModifiedWhisperEncoder(
    whisper.WhisperEncoder, transformers.modeling_utils.ModuleUtilsMixin
):
    """
    Encoder portion of OpenAI's Whisper model.

    This implementation is a slightly modified version of HF Transformers' Whisper Encoder, with only a few fixes:
    1. base_model_prefix updated to allow for doing `.from_pretrained` directly on the encoder
    2. allow less than 30 second of audio padding to be passed in:
        - relaxed ValueError check for `input_features` length to be less than or equal to `expected_seq_length` instead of strictly equal
        - embed_pos is now sliced to match the length of `inputs_embeds`

    Original: https://github.com/huggingface/transformers/blob/main/src/transformers/models/whisper/modeling_whisper.py
    """

    base_model_prefix = "model.encoder"
    _no_split_modules = ["WhisperEncoderLayer"]
    _keys_to_ignore_on_load_unexpected = ["model.decoder.*"]

    def __init__(self, config: transformers.WhisperConfig):
        super().__init__(config)
        self.config.is_decoder = False

    @property
    def max_context_length(self):
        return (
            self.config.max_source_positions
            * self.conv1.stride[0]
            * self.conv2.stride[0]
        )

    def init_latency_mask(
        self, audio_latency_block_size: int | None, dtype: torch.dtype
    ):
        if audio_latency_block_size is None:
            self.audio_streaming_mask = None
            return

        # Use max_context_length directly in the calculation
        max_seqlen = self.max_context_length
        assert (
            max_seqlen > 0
        ), f"maximum sequence length must be positive, got {max_seqlen}"
        assert (
            max_seqlen % audio_latency_block_size == 0
        ), f"audio_latency_block_size {audio_latency_block_size} must divide {max_seqlen} evenly."
        # Given the block size, we calculate number of blocks.
        audio_latency_nblocks = max_seqlen // audio_latency_block_size
        audio_streaming_mask = (
            torch.tril(
                torch.ones(audio_latency_nblocks, audio_latency_nblocks),
                diagonal=0,
            )
            .repeat_interleave(audio_latency_block_size, dim=0)
            .repeat_interleave(audio_latency_block_size, dim=1)
        )
        audio_streaming_mask = (1.0 - audio_streaming_mask) * torch.finfo(dtype).min
        audio_streaming_mask = audio_streaming_mask[None, None, :, :]
        self.register_buffer(
            "audio_streaming_mask", audio_streaming_mask, persistent=False
        )

    def forward(
        self,
        input_features,
        audio_len=None,
        head_mask=None,
        output_attentions=None,
        output_hidden_states=None,
        return_dict=None,
    ):
        expected_seq_length = self.max_context_length
        if input_features.shape[-1] > expected_seq_length:
            raise ValueError(
                f"Whisper expects the mel input features to be of length {expected_seq_length} or less, but found {input_features.shape[-1]}. Make sure to pad the input mel features to {expected_seq_length}."
            )

        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
        )
        inputs_embeds = nn.functional.gelu(self.conv1(input_features))
        inputs_embeds = nn.functional.gelu(self.conv2(inputs_embeds))

        inputs_embeds = inputs_embeds.permute(0, 2, 1)
        embed_pos = self.embed_positions.weight[: inputs_embeds.size(-2)]

        hidden_states = inputs_embeds + embed_pos
        hidden_states = nn.functional.dropout(
            hidden_states, p=self.dropout, training=self.training
        )

        encoder_states = () if output_hidden_states else None
        all_attentions = () if output_attentions else None

        # Create attention mask based on audio lengths to mask out padding tokens
        # For each sample in batch:
        # - Convert raw audio length to feature length after convolutions
        # - Create boolean mask that is True for valid positions and False for padding
        # - Convert to extended attention mask format expected by transformer layers
        #   (1.0 for positions to attend to, large negative for positions to ignore)
        # This masking ensures consistent behavior between training and inference
        # by preventing the model from attending to padding tokens in both cases
        attention_mask = None
        if audio_len is not None:
            audio_feature_len = self._get_feat_extract_output_lengths(audio_len)
            max_seq_len = hidden_states.shape[1]
            attention_mask = torch.arange(max_seq_len, device=hidden_states.device)[
                None, :
            ].lt(audio_feature_len.view(-1, 1))
            attention_mask = self.get_extended_attention_mask(
                attention_mask,
                None,
                dtype=hidden_states.dtype,
            )

        if self.audio_streaming_mask is not None:
            seqlen = hidden_states.size(-2)
            if attention_mask is not None:
                attention_mask = torch.minimum(
                    self.audio_streaming_mask[:, :, :seqlen, :seqlen], attention_mask
                )  # merge
            else:
                attention_mask = self.audio_streaming_mask[:, :, :seqlen, :seqlen]
            attention_mask = attention_mask.to(hidden_states.dtype)

        # check if head_mask has a correct number of layers specified if desired
        if head_mask is not None:
            assert head_mask.size()[0] == (
                len(self.layers)
            ), f"The head_mask should be specified for {len(self.layers)} layers, but it is for {head_mask.size()[0]}."

        for idx, encoder_layer in enumerate(self.layers):
            if output_hidden_states:
                encoder_states = encoder_states + (hidden_states,)
            # add LayerDrop (see https://arxiv.org/abs/1909.11556 for description)
            to_drop = False
            if self.training:
                dropout_probability = torch.rand([])
                if dropout_probability < self.layerdrop:  # skip the layer
                    to_drop = True

            if to_drop:
                layer_outputs = (None, None)
            else:
                if self.gradient_checkpointing and self.training:
                    layer_outputs = self._gradient_checkpointing_func(
                        encoder_layer.__call__,
                        hidden_states,
                        attention_mask,
                        (head_mask[idx] if head_mask is not None else None),
                        output_attentions,
                    )
                else:
                    layer_outputs = encoder_layer(
                        hidden_states,
                        attention_mask,
                        layer_head_mask=(
                            head_mask[idx] if head_mask is not None else None
                        ),
                        output_attentions=output_attentions,
                    )

                hidden_states = layer_outputs[0]

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

        hidden_states = self.layer_norm(hidden_states)
        if output_hidden_states:
            encoder_states = encoder_states + (hidden_states,)

        if not return_dict:
            return tuple(
                v
                for v in [hidden_states, encoder_states, all_attentions]
                if v is not None
            )
        return transformers.modeling_outputs.BaseModelOutput(
            last_hidden_state=hidden_states,
            hidden_states=encoder_states,
            attentions=all_attentions,
        )


UltravoxConfig.register_for_auto_class()
UltravoxModel.register_for_auto_class()

transformers.AutoConfig.register("ultravox", UltravoxConfig)
transformers.AutoModel.register(UltravoxConfig, UltravoxModel)

transformers.activations.ACT2FN["swiglu"] = SwiGLU