modeling_gemma3mm.py

import copy
from collections.abc import Callable
from dataclasses import dataclass
from typing import List, Optional, Tuple, Union

import torch
import torch.nn as nn

from transformers.activations import ACT2FN
from transformers.cache_utils import Cache, HybridCache, StaticCache
from transformers.generation import GenerationMixin
from transformers.modeling_flash_attention_utils import FlashAttentionKwargs
from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, ModelOutput
from transformers.modeling_rope_utils import ROPE_INIT_FUNCTIONS
from transformers.modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel
from transformers.processing_utils import Unpack
from transformers.utils import (
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    is_torchdynamo_compiling,
    logging,
    replace_return_docstrings,
)
from transformers.utils.deprecation import deprecate_kwarg
from transformers import AutoModel, AutoModelForCausalLM

from transformers.models.gemma3.modeling_gemma3 import Gemma3CausalLMOutputWithPast, Gemma3PreTrainedModel, Gemma3MultiModalProjector

from transformers import AutoConfig, AutoModelForCausalLM

from .configuration_gemma3mm import Gemma3MMConfig
from .processing_gemma3mm import InputMode
from .speech_conformer_encoder import ConformerEncoder

logger = logging.get_logger(__name__)
_CONFIG_FOR_DOC = "Gemma3MMConfig"

@dataclass
class Gemma3MMCausalLMOutputWithPast(Gemma3CausalLMOutputWithPast):  # ← 부모 클래스 변경
    """
    Multimodal version of `Gemma3CausalLMOutputWithPast`.
    Adds audio-specific hidden states.

    Args:
        audio_hidden_states (`torch.FloatTensor`, *optional*):
            A `torch.FloatTensor` of size `(batch_size, sequence_length, hidden_size)`.
            Audio hidden states produced by the audio encoder.
    """
    audio_hidden_states: Optional[torch.FloatTensor] = None


GEMMA3_START_DOCSTRING = r"""
    This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
    library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
    etc.)

    This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
    and behavior.

    Parameters:
        config ([`Gemma3Config`]):
            Model configuration class with all the parameters of the model. Initializing with a config file does not
            load the weights associated with the model, only the configuration. Check out the
            [`~PreTrainedModel.from_pretrained`] method to load the model weights.
"""



GEMMA3_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
            Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
            it.

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            [What are input IDs?](../glossary#input-ids)
        attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:

            - 1 for tokens that are **not masked**,
            - 0 for tokens that are **masked**.

            [What are attention masks?](../glossary#attention-mask)

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            If `past_key_values` is used, optionally only the last `input_ids` have to be input (see
            `past_key_values`).

            If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
            and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
            information on the default strategy.

            - 1 indicates the head is **not masked**,
            - 0 indicates the head is **masked**.
        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
            config.n_positions - 1]`.

            [What are position IDs?](../glossary#position-ids)
        past_key_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*):
            Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
            blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values`
            returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`.

            Two formats are allowed:
            - a [`~cache_utils.Cache`] instance, see our
            [kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache);
            - Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
            shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`). This is also known as the legacy
            cache format.

            The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the
            legacy cache format will be returned.

            If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't
            have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids`
            of shape `(batch_size, sequence_length)`.
        inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
            is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
            model's internal embedding lookup matrix.
        use_cache (`bool`, *optional*):
            If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
            `past_key_values`).
        output_attentions (`bool`, *optional*):
            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
            tensors for more detail.
        output_hidden_states (`bool`, *optional*):
            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
            more detail.
        return_dict (`bool`, *optional*):
            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
        cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
            Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`,
            this tensor is not affected by padding. It is used to update the cache in the correct position and to infer
            the complete sequence length.
"""

@add_start_docstrings(
    "The bare Gemma3 Model outputting raw hidden-states without any specific head on top.",
    GEMMA3_START_DOCSTRING,
)
class Gemma3MMPreTrainedModel(Gemma3PreTrainedModel):
    config_class = Gemma3MMConfig

@add_start_docstrings(
    """The GEMMA3 model which consists of a vision backbone and a language model.""",
    GEMMA3_START_DOCSTRING,
)
class Gemma3MMForConditionalGeneration(Gemma3MMPreTrainedModel, GenerationMixin):
    def __init__(self, config: Gemma3MMConfig):
        super().__init__(config)
        self.vision_tower = AutoModel.from_config(config=config.vision_config)
        audio_config = config.audio_config.to_diff_dict()
        for item in ['transformers_version', 'model_type', 'torch_dtype']:
            if item in audio_config:
                audio_config.pop(item)
        self.audio_tower = ConformerEncoder(**audio_config)
        self.audio_tower.post_init({})
        self.audio_projector = nn.Sequential(
            nn.Linear(in_features=config.audio_config.attention_dim, out_features=config.text_config.hidden_size, bias=True),
            nn.GELU(approximate='none'),
            nn.Linear(in_features=config.text_config.hidden_size, out_features=config.text_config.hidden_size, bias=True)
        ).to(dtype=self.dtype)

        self.multi_modal_projector = Gemma3MultiModalProjector(config)
        self.vocab_size = config.text_config.vocab_size

        language_model = AutoModelForCausalLM.from_config(config=config.text_config)

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

        self.pad_token_id = self.config.pad_token_id if self.config.pad_token_id is not None else -1
        
        # LoRA 어댑터 설정 추가
        if hasattr(config, "speech_lora") and config.speech_lora is not None:
            from peft import LoraConfig, get_peft_model
            import warnings
            
            speech_lora_config = LoraConfig(
                r=config.speech_lora['r'],
                lora_alpha=config.speech_lora['lora_alpha'],
                target_modules=config.speech_lora['layer'],
                use_rslora=config.speech_lora['use_rslora'],
                lora_dropout=config.speech_lora['dp'],
                task_type="CAUSAL_LM",
            )
            self.language_model.model = get_peft_model(self.language_model.model, speech_lora_config, adapter_name="speech")
        
        self.post_init()
        
    def set_lora_adapter(self, adapter_name) -> None:
        from peft.tuners.lora.layer import LoraLayer
        for module in self.modules():
            if isinstance(module, LoraLayer):
                if module.merged:
                    warnings.warn("Adapter cannot be set when the model is merged. Unmerging the model first.")
                    module.unmerge()
                module.set_adapter(adapter_name)
                module._disable_adapters = False

    def unset_lora_adapter(self) -> None:
        # Ref: peft/tuners/tuners_utils.py - enable_adapters()
        # Ref: peft/tuners/lora/layer.py
        from peft.tuners.lora.layer import LoraLayer
        for module in self.modules():
            if isinstance(module, LoraLayer):
                # disable grads on all adapter layers
                # TODO weijian: may use enable_adapters() instead
                for layer_name in module.adapter_layer_names:
                    layer = getattr(module, layer_name)
                    layer.requires_grad_(False)
                module._disable_adapters = True

    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 _update_causal_mask(
        self,
        attention_mask,
        token_type_ids,
        past_key_values,
        cache_position,
        input_tensor,
        is_training: bool = False,
    ):
        if self.config.text_config._attn_implementation == "flash_attention_2":
            return attention_mask

        if attention_mask is not None and attention_mask.dim() == 4:
            # In this case we assume that the mask comes already in inverted
            # form and requires no inversion or slicing.
            return attention_mask

        using_static_cache = isinstance(past_key_values, StaticCache)
        min_dtype = torch.finfo(self.dtype).min
        inputs_lead_dim, sequence_length = input_tensor.shape[:2]
        if using_static_cache:
            target_length = past_key_values.get_max_cache_shape()
        elif isinstance(past_key_values, HybridCache):
            target_length = past_key_values.get_max_cache_shape()
        else:
            target_length = (
                attention_mask.shape[-1]
                if isinstance(attention_mask, torch.Tensor)
                else cache_position[0] + sequence_length + 1
            )

        if attention_mask is not None and attention_mask.dim() == 4:
            # In this case we assume that the mask comes already in inverted form and requires no inversion or slicing.
            return attention_mask

        causal_mask = torch.full(
            (sequence_length, target_length), fill_value=min_dtype, dtype=self.dtype, device=cache_position.device
        )

        # Causal diagonal mask only if training, otherwise attend to the whole prefix. Training-specific attn for prefix is handled below
        if sequence_length != 1:
            causal_mask = torch.triu(causal_mask, diagonal=1)

        causal_mask *= torch.arange(target_length, device=cache_position.device) > cache_position.reshape(-1, 1)
        causal_mask = causal_mask[None, None, :, :].expand(inputs_lead_dim, 1, -1, -1)

        # Apply bidirectional mask on images if token type ids are provided
        if token_type_ids is not None and sequence_length != 1:
            token_type_mask = token_type_ids.unsqueeze(1) == token_type_ids.unsqueeze(2)
            token_type_mask[token_type_ids == 0] = False  # if text token do not change anything
            token_type_mask = token_type_mask.unsqueeze(1).to(causal_mask.device, dtype=torch.bool)
            causal_mask = causal_mask.clone()
            causal_mask[:, :, :, :sequence_length] = causal_mask[:, :, :, :sequence_length].masked_fill(
                token_type_mask, 0.0
            )

        if attention_mask is not None:
            causal_mask = causal_mask.clone()  # copy to contiguous memory for in-place edit
            mask_length = attention_mask.shape[-1]

            # Then apply padding mask (will mask pad tokens)
            padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(causal_mask.device)
            padding_mask = padding_mask == 0
            causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
                padding_mask, min_dtype
            )

        return causal_mask

    def get_image_features(self, pixel_values: torch.Tensor):
        """
        Projects the last hidden state from the vision model into language model space.

        Args:
            pixel_values (`torch.FloatTensor]` of shape `(batch_size, channels, height, width)`)
               The tensors corresponding to the input images.
        Returns:
            image_features (`torch.Tensor`): Image feature tensor of shape `(num_images, image_length, embed_dim)`).
        """
        vision_outputs = self.vision_tower(pixel_values=pixel_values).last_hidden_state
        image_features = self.multi_modal_projector(vision_outputs)
        return image_features
        
    def get_audio_features(self, input_audio_embeds: torch.FloatTensor, audio_attention_mask: torch.FloatTensor, audio_embed_sizes: torch.FloatTensor):
        """
        Projects the last hidden state from the audio model into language model space.
        
        Args:
            audio_inputs (`torch.FloatTensor]` of shape `(batch_size, sequence_length, feature_dim)`)
                The tensors corresponding to the input audio features.
                
        Returns:
            audio_features (`torch.Tensor`): Audio feature tensor of shape `(batch_size, audio_length, embed_dim)`).
        """
        audio_features, masks = self.audio_tower(input_audio_embeds, audio_attention_mask)
        audio_outputs = self.audio_projector(audio_features)
        return audio_outputs

    @deprecate_kwarg("num_logits_to_keep", version="4.50", new_name="logits_to_keep")
    @add_start_docstrings_to_model_forward(GEMMA3_INPUTS_DOCSTRING)
    @replace_return_docstrings(output_type=Gemma3MMCausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
    def forward(
        self,
        input_ids: torch.LongTensor = None,
        pixel_values: torch.FloatTensor = None,
        input_audio_embeds: torch.FloatTensor = None,
        audio_embed_sizes: torch.FloatTensor = None,
        audio_attention_mask: torch.FloatTensor = None,
        attention_mask: Optional[torch.Tensor] = None,
        input_modes: torch.LongTensor = None,
        position_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[Union[List[torch.FloatTensor], Cache]] = None,
        token_type_ids: Optional[torch.LongTensor] = None,
        cache_position: Optional[torch.LongTensor] = 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,
        logits_to_keep: Union[int, torch.Tensor] = 0,
        **lm_kwargs,
    ) -> Union[Tuple, Gemma3MMCausalLMOutputWithPast]:
        r"""
            labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
                Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
                config.text_config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
                (masked), the loss is only computed for the tokens with labels in `[0, ..., config.text_config.vocab_size]`.

            logits_to_keep (`int` or `torch.Tensor`, *optional*):
                If an `int`, compute logits for the last `logits_to_keep` tokens. If `0`, calculate logits for all
                `input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
                token can save memory, which becomes pretty significant for long sequences or large vocabulary size.
                If a `torch.Tensor`, must be 1D corresponding to the indices to keep in the sequence length dimension.
                This is useful when using packed tensor format (single dimension for batch and sequence length).

        Returns:

        Example:

        ```python
        >>> from PIL import Image
        >>> import requests
        >>> from transformers import AutoProcessor, Gemma3MMForConditionalGeneration

        >>> model = Gemma3MMForConditionalGeneration.from_pretrained("google/Gemma3-test-224px-hf")
        >>> processor = AutoProcessor.from_pretrained("google/Gemma3-test-224px-hf")

        >>> prompt = "answer en Where is the cow standing?"
        >>> url = "https://huggingface.co/gv-hf/Gemma3-test-224px-hf/resolve/main/cow_beach_1.png"
        >>> image = Image.open(requests.get(url, stream=True).raw)

        >>> inputs = processor(images=image, text=prompt,  return_tensors="pt")

        >>> # Generate
        >>> generate_ids = model.generate(**inputs, max_length=30)
        >>> processor.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
        "answer en Where is the cow standing?\nbeach"
        ```"""

        if (input_ids is None) ^ (inputs_embeds is not None):
            raise ValueError("You must specify exactly one of input_ids or inputs_embeds")

        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

        if isinstance(input_modes, torch.Tensor):
            # len(input_mode) == num_beams in beam search, and all elements of input_mode should have the same value
            input_modes = input_modes.unique()
            if len(input_modes) != 1:
                raise ValueError("Elements of input_modes should have the same value")

        input_mode = InputMode(input_modes.item())

        if input_mode in [InputMode.VISION_SPEECH, InputMode.VISION]:
            self.unset_lora_adapter()
            #self.set_lora_adapter('vision')
            #audio_projection_mode = 'vision'
        elif input_mode == InputMode.SPEECH:
            self.set_lora_adapter('speech')
            #audio_projection_mode = 'speech'
        elif input_mode == InputMode.LANGUAGE:
            self.unset_lora_adapter()
            #audio_projection_mode = 'speech'
        else:
            raise ValueError(f"Invalid input_mode: {input_mode}")

        is_training = token_type_ids is not None and labels is not None

        # Replace image id woth PAD if the image token if OOV, to avoid index-errors
        if input_ids is not None and self.config.image_token_index >= self.vocab_size or self.config.audio_token_index >= self.vocab_size:
            special_image_mask = input_ids == self.config.image_token_index
            special_audio_mask = input_ids == self.config.audio_token_index
            llm_input_ids = input_ids.clone()
            llm_input_ids[special_image_mask] = 0
            llm_input_ids[special_audio_mask] = 0
        else:
            llm_input_ids = input_ids

        if inputs_embeds is None:
            inputs_embeds = self.get_input_embeddings()(llm_input_ids)

        if cache_position is None:
            past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
            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) + 1  # Gemma3 positions are 1-indexed

        # Merge text and images
        if pixel_values is not None:
            image_features = self.get_image_features(pixel_values)

            if input_ids is None:
                special_image_mask = inputs_embeds == self.get_input_embeddings()(
                    torch.tensor(self.config.image_token_index, dtype=torch.long, device=inputs_embeds.device)
                )
            else:
                special_image_mask = (input_ids == self.config.image_token_index).unsqueeze(-1)
                special_image_mask = special_image_mask.expand_as(inputs_embeds).to(inputs_embeds.device)

            if not is_torchdynamo_compiling() and inputs_embeds[special_image_mask].numel() != image_features.numel():
                image_tokens_in_text = (special_image_mask).sum(dim=1).sum(dim=0)[0]
                raise ValueError(
                    f"Number of images does not match number of special image tokens in the input text. "
                    f"Got {image_tokens_in_text} image tokens in the text but {image_features.shape[0] * image_features.shape[1]} "
                    "tokens from image embeddings."
                )
            image_features = image_features.to(inputs_embeds.device, inputs_embeds.dtype)
            inputs_embeds = inputs_embeds.masked_scatter(special_image_mask, image_features)

        # Merge text and audios
        if input_audio_embeds is not None:
            audio_features = self.get_audio_features(input_audio_embeds, audio_attention_mask, audio_embed_sizes)
            if input_ids is None:
                special_audio_mask = inputs_embeds == self.get_input_embeddings()(
                    torch.tensor(self.config.audio_token_index, dtype=torch.long, device=inputs_embeds.device)
                )
            else:
                special_audio_mask = (input_ids == self.config.audio_token_index).unsqueeze(-1)
                special_audio_mask = special_audio_mask.expand_as(inputs_embeds).to(inputs_embeds.device)
                
            masked_audio_features = []
            for i, size in enumerate(audio_embed_sizes):
                masked_audio_features.append(audio_features[i, :size, :])
            masked_audio_features = torch.cat(masked_audio_features, dim=0)

            if not is_torchdynamo_compiling() and inputs_embeds[special_audio_mask].numel() != masked_audio_features.numel():
                audio_tokens_in_text = (special_audio_mask).sum(dim=1).sum(dim=0)[0]
                masked_audio_size = audio_embed_sizes.sum()[0]
                raise ValueError(
                    f"Number of images does not match number of special image tokens in the input text. "
                    f"Got {audio_tokens_in_text} image tokens in the text but {masked_audio_size} "
                    "tokens from image embeddings."
                )
            masked_audio_features = masked_audio_features.to(inputs_embeds.device, inputs_embeds.dtype)
            inputs_embeds = inputs_embeds.masked_scatter(special_audio_mask, masked_audio_features)

        # mask out pad-token-ids in labels for BC
        if labels is not None and self.pad_token_id in labels:
            logger.warning_once(
                "`labels` contains `pad_token_id` which will be masked with `config.ignore_index`. "
                "You have to mask out `pad_token_id` when preparing `labels`, this behavior will be removed in v.4.46.",
            )
            labels = torch.where(input_ids == self.pad_token_id, self.config.ignore_index, labels)

        causal_mask = self._update_causal_mask(
            attention_mask, token_type_ids, past_key_values, cache_position, inputs_embeds, is_training
        )
        outputs = self.language_model(
            attention_mask=causal_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,
            cache_position=cache_position,
            logits_to_keep=logits_to_keep,
            **lm_kwargs,
        )

        logits = outputs.logits
        loss = None
        if labels is not None:
            # Upcast to float if we need to compute the loss to avoid potential precision issues
            logits = logits.float()
            shift_logits = logits[..., :-1, :]
            shift_labels = labels[..., 1:]
            if attention_mask is not None:
                # we use the input attention mask to shift the logits and labels, because it is 2D.
                # we also crop attn mask in case it is longer, which happens in PrefixTuning with peft
                shift_attention_mask = attention_mask[:, -shift_logits.shape[1] :].to(logits.device)
                shift_logits = shift_logits[shift_attention_mask.to(logits.device) != 0].contiguous()
                shift_labels = shift_labels[shift_attention_mask.to(shift_labels.device) != 0].contiguous()
            else:
                shift_logits = shift_logits.contiguous()
                shift_labels = shift_labels.contiguous()
            # Flatten the tokens
            loss_fct = nn.CrossEntropyLoss()

            flat_logits = shift_logits.view(-1, self.config.text_config.vocab_size)
            flat_labels = shift_labels.view(-1).to(shift_logits.device)
            loss = loss_fct(flat_logits, flat_labels)
        if not return_dict:
            output = (logits,) + outputs[1:]
            return (loss,) + output if loss is not None else output

        return Gemma3MMCausalLMOutputWithPast(
            loss=loss,
            logits=logits,
            past_key_values=outputs.past_key_values,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
            image_hidden_states=image_features if pixel_values is not None else None,
            audio_hidden_states=audio_features if input_audio_embeds is not None else None,
        )

    def prepare_inputs_for_generation(
        self,
        input_ids,
        past_key_values=None,
        input_modes=None,
        inputs_embeds=None,
        cache_position=None,
        position_ids=None,
        pixel_values=None,
        input_audio_embeds=None,
        audio_embed_sizes=None,
        audio_attention_mask=None,
        attention_mask=None,
        token_type_ids=None,
        use_cache=True,
        logits_to_keep=None,
        labels=None,
        **kwargs,
    ):
        # Overwritten -- custom `position_ids` and `pixel_values` handling
        model_inputs = self.language_model.prepare_inputs_for_generation(
            input_ids,
            past_key_values=past_key_values,
            input_modes=input_modes,
            inputs_embeds=inputs_embeds,
            attention_mask=attention_mask,
            position_ids=position_ids,
            cache_position=cache_position,
            use_cache=use_cache,
            logits_to_keep=logits_to_keep,
            token_type_ids=token_type_ids,
            **kwargs,
        )

        # position_ids in Gemma3 are 1-indexed
        if model_inputs.get("position_ids") is not None:
            model_inputs["position_ids"] += 1
        # If we're in cached decoding stage, pixel values should be None because input ids do not contain special image token anymore
        # Otherwise we need pixel values to be passed to model. NOTE: use_cache=False needs pixel_values always
        if cache_position[0] == 0:
            model_inputs["pixel_values"] = pixel_values
            model_inputs["input_audio_embeds"] = input_audio_embeds
            model_inputs["audio_embed_sizes"] = audio_embed_sizes
            model_inputs["audio_attention_mask"] = audio_attention_mask
        model_inputs["input_modes"] = input_modes
        is_training = token_type_ids is not None and labels is not None
        if cache_position[0] == 0 and isinstance(past_key_values, HybridCache):
            input_tensor = inputs_embeds if inputs_embeds is not None else input_ids
            causal_mask = self._update_causal_mask(
                attention_mask, token_type_ids, past_key_values, cache_position, input_tensor, is_training
            )
            model_inputs["attention_mask"] = causal_mask

        return model_inputs

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