Upload folder using huggingface_hub

configuration_aria.py

@@ -0,0 +1,114 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+import logging
+
+from transformers.configuration_utils import PretrainedConfig
+
+from .moe_lm import AriaMoELMConfig
+from .vision_encoder import AriaVisionConfig
+
+logger = logging.getLogger(__name__)
+
+
+# adapted from transformers.models.llava.configuration_llava.LlavaConfig
+class AriaConfig(PretrainedConfig):
+    """
+    Configuration class for Aria model.
+
+    This class handles the configuration for both vision and text components of the Aria model,
+    as well as additional parameters for image token handling and projector mapping.
+
+    Args:
+        vision_config (AriaVisionConfig or dict): Configuration for the vision component.
+        text_config (AriaMoELMConfig or dict): Configuration for the text component.
+        projector_patch_to_query_dict (dict): Mapping of patch sizes to query dimensions.
+        ignore_index (int): Index to ignore in loss calculation.
+        image_token_index (int): Index used to represent image tokens.
+        **kwargs: Additional keyword arguments passed to the parent class.
+
+    Attributes:
+        model_type (str): Type of the model, set to "aria".
+        is_composition (bool): Whether the model is a composition of multiple components.
+        ignore_index (int): Index to ignore in loss calculation.
+        image_token_index (int): Index used to represent image tokens.
+        projector_patch_to_query_dict (dict): Mapping of patch sizes to query dimensions.
+        vision_config (AriaVisionConfig): Configuration for the vision component.
+        text_config (AriaMoELMConfig): Configuration for the text component.
+    """
+
+    model_type = "aria"
+    is_composition = False
+
+    def __init__(
+        self,
+        vision_config=AriaVisionConfig(),
+        text_config=AriaMoELMConfig(),
+        projector_patch_to_query_dict={
+            1225: 128,
+            4900: 256,
+        },
+        ignore_index=-100,
+        image_token_index=32000,
+        tie_word_embeddings=False,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.ignore_index = ignore_index
+        self.image_token_index = image_token_index
+        self.tie_word_embeddings = tie_word_embeddings
+        attn_implementation = kwargs.pop("attn_implementation", None)
+
+        # Set the default attention implementation to flash_attention_2 if not specified
+        self._attn_implementation = (
+            "flash_attention_2" if attn_implementation is None else attn_implementation
+        )
+
+        # Convert the keys and values of projector_patch_to_query_dict to integers
+        # This ensures consistency even if they were provided as strings
+        self.projector_patch_to_query_dict = {
+            int(k): int(v) for k, v in projector_patch_to_query_dict.items()
+        }
+
+        if isinstance(vision_config, dict) and "model_type" in vision_config:
+            vision_config = AriaVisionConfig(**vision_config)
+            if attn_implementation is None:
+                vision_attn_implementation = "flash_attention_2"
+            elif attn_implementation == "sdpa":
+                logger.warning(
+                    "SDPA is not supported for vit, using flash_attention_2 instead"
+                )
+                vision_attn_implementation = "flash_attention_2"
+            else:
+                vision_attn_implementation = attn_implementation
+            vision_config._attn_implementation = vision_attn_implementation
+
+        self.vision_config = vision_config
+
+        if isinstance(text_config, dict) and "model_type" in text_config:
+            text_attn_implementation = (
+                "sdpa" if attn_implementation is None else attn_implementation
+            )
+            text_config = AriaMoELMConfig(**text_config)
+            text_config._attn_implementation = text_attn_implementation
+
+        self.text_config = text_config
+
+        # This is needed for the static kv cache
+        self.num_hidden_layers = self.text_config.num_hidden_layers

modeling_aria.py

@@ -0,0 +1,365 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+from dataclasses import dataclass
+from typing import List, Optional, Tuple, Union
+
+import torch
+import torch.nn as nn
+from torch import nn
+from transformers import GenerationMixin, PreTrainedModel
+from transformers.modeling_outputs import ModelOutput
+from transformers.utils import logging
+
+from .configuration_aria import AriaConfig
+from .moe_lm import AriaMoELMForCausalLM
+from .projector import AriaProjector
+from .vision_encoder import AriaVisionModel
+
+logger = logging.get_logger(__name__)
+
+
+class AriaPretrainedModel(PreTrainedModel):
+    """
+    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained models.
+    """
+
+    config_class = AriaConfig
+    base_model_prefix = "model"
+    _no_split_modules = []
+    supports_gradient_checkpointing = True
+    _skip_keys_device_placement = "past_key_values"
+    _supports_flash_attn_2 = True
+    _supports_cache_class = True
+    _supports_static_cache = True
+
+    @property
+    def _supports_sdpa(self):
+        """
+        Retrieve language_model's attribute to check whether the model supports
+        SDPA (Scaled Dot Product Attention) or not.
+        """
+        return self.language_model._supports_sdpa
+
+
+@dataclass
+# Copied from transformers.models.llava.modeling_llava.LlavaCausalLMOutputWithPast with Llava->Aria
+class AriaCausalLMOutputWithPast(ModelOutput):
+    """
+    Base class for Aria causal language model (or autoregressive) outputs.
+
+    Args:
+        loss (`torch.FloatTensor` of shape `(1,)`, *optional*, returned when `labels` is provided):
+            Language modeling loss (for next-token prediction).
+        logits (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`):
+            Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
+        past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
+            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)`)
+
+            Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see
+            `past_key_values` input) to speed up sequential decoding.
+        hidden_states (`tuple(torch.FloatTensor)`, *optional*, returned when `output_hidden_states=True` is passed or when `config.output_hidden_states=True`):
+            Tuple of `torch.FloatTensor` (one for the output of the embeddings, if the model has an embedding layer, +
+            one for the output of each layer) of shape `(batch_size, sequence_length, hidden_size)`.
+
+            Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
+        attentions (`tuple(torch.FloatTensor)`, *optional*, returned when `output_attentions=True` is passed or when `config.output_attentions=True`):
+            Tuple of `torch.FloatTensor` (one for each layer) of shape `(batch_size, num_heads, sequence_length,
+            sequence_length)`.
+
+            Attentions weights after the attention softmax, used to compute the weighted average in the self-attention
+            heads.
+        image_hidden_states (`tuple(torch.FloatTensor)`, *optional*):
+            Tuple of `torch.FloatTensor` (one for the output of the image embeddings, `(batch_size, num_images,
+            sequence_length, hidden_size)`.
+
+            image_hidden_states of the model produced by the vision encoder, and optionally by the perceiver
+    """
+
+    loss: Optional[torch.FloatTensor] = None
+    logits: torch.FloatTensor = None
+    past_key_values: Optional[List[torch.FloatTensor]] = None
+    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+    attentions: Optional[Tuple[torch.FloatTensor]] = None
+    image_hidden_states: Optional[Tuple[torch.FloatTensor]] = None
+
+
+def build_mm_projector(config: AriaConfig):
+    """
+    Builds and returns an AriaProjector instance based on the provided configuration.
+
+    Args:
+        config (AriaConfig): The configuration object containing necessary parameters.
+
+    Returns:
+        AriaProjector: An instance of the AriaProjector class.
+    """
+    return AriaProjector(
+        patch_to_query_dict=config.projector_patch_to_query_dict,
+        embed_dim=config.vision_config.hidden_size,
+        num_heads=config.vision_config.num_attention_heads,
+        kv_dim=config.vision_config.hidden_size,
+        ff_dim=config.text_config.hidden_size,
+        output_dim=config.text_config.hidden_size,
+    )
+
+
+# adapted from transformers.models.llava.modeling_llava.LlavaForConditionalGeneration
+class AriaForConditionalGeneration(AriaPretrainedModel, GenerationMixin):
+    """
+    Aria model for conditional generation tasks.
+
+    This model combines a vision tower, a multi-modal projector, and a language model
+    to perform tasks that involve both image and text inputs.
+    """
+
+    def __init__(self, config: AriaConfig):
+        super().__init__(config)
+
+        self.vision_tower = AriaVisionModel(config.vision_config)
+        self.multi_modal_projector = build_mm_projector(config)
+        self.vocab_size = config.text_config.vocab_size
+        self.language_model = AriaMoELMForCausalLM(config.text_config)
+        self.pad_token_id = (
+            self.config.pad_token_id if self.config.pad_token_id is not None else -1
+        )
+        self.post_init()
+
+    def freeze_vit(self):
+        """Freeze the parameters of the vision tower."""
+        for param in self.vision_tower.parameters():
+            param.requires_grad = False
+
+    def freeze_projector(self):
+        """Freeze the parameters of the multi-modal projector."""
+        for param in self.multi_modal_projector.parameters():
+            param.requires_grad = False
+
+    def freeze_llm(self):
+        """Freeze the parameters of the language model."""
+        for param in self.language_model.parameters():
+            param.requires_grad = False
+
+    def get_input_embeddings(self) -> nn.Module:
+        """Retrieve the input embeddings from the language model."""
+        return self.language_model.get_input_embeddings()
+
+    def set_input_embeddings(self, value):
+        """Set the input embeddings for the language model."""
+        self.language_model.set_input_embeddings(value)
+
+    def get_output_embeddings(self):
+        """Retrieve the output embeddings from the language model."""
+        return self.language_model.get_output_embeddings()
+
+    def set_output_embeddings(self, value):
+        """Set the output embeddings for the language model."""
+        self.language_model.set_output_embeddings(value)
+
+    def set_moe_z_loss_coeff(self, value):
+        """
+        Set the z-loss coefficient for Mixture of Experts (MoE) models.
+
+        Args:
+            value: The z-loss coefficient value to set.
+        """
+        self.language_model.set_z_loss_coeff(value)
+
+    def set_moe_aux_loss_coeff(self, value):
+        """
+        Set the auxiliary loss coefficient for Mixture of Experts (MoE) models.
+
+        Args:
+            value: The auxiliary loss coefficient value to set.
+        """
+        self.language_model.set_aux_loss_coeff(value)
+
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        pixel_values: torch.FloatTensor = None,
+        pixel_mask: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        num_logits_to_keep: int = 0,
+    ) -> Union[Tuple, AriaCausalLMOutputWithPast]:
+        """
+        Forward pass of the AriaForConditionalGeneration model.
+
+        This method processes both text and image inputs, merges them if necessary,
+        and generates output using the language model.
+
+        Args:
+            input_ids (torch.LongTensor, optional): Input token ids.
+            pixel_values (torch.FloatTensor, optional): Pixel values of the images.
+            pixel_mask (torch.LongTensor, optional): Mask for the pixel values.
+            attention_mask (torch.Tensor, optional): Attention mask.
+            position_ids (torch.LongTensor, optional): Position ids.
+            past_key_values (List[torch.FloatTensor], optional): Past key values for efficient processing.
+            inputs_embeds (torch.FloatTensor, optional): Input embeddings.
+            labels (torch.LongTensor, optional): Labels for computing the language modeling loss.
+            use_cache (bool, optional): Whether to use the model's cache mechanism.
+            output_attentions (bool, optional): Whether to output attention weights.
+            output_hidden_states (bool, optional): Whether to output hidden states.
+            return_dict (bool, optional): Whether to return a ModelOutput object.
+
+        Returns:
+            Union[Tuple, AriaCausalLMOutputWithPast]: Model outputs.
+        """
+        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 inputs_embeds is None:
+            # 1. Extra the input embeddings
+            inputs_embeds = self.get_input_embeddings()(input_ids)
+
+        image_features = None
+        if pixel_values is not None:
+            image_outputs, image_attn_mask = self.vision_tower(
+                pixel_values,
+                pixel_mask=pixel_mask,
+            )
+
+            selected_image_feature = image_outputs.last_hidden_state
+            image_features = self.multi_modal_projector(
+                selected_image_feature, attn_mask=image_attn_mask
+            )
+
+        if image_features is not None:
+            n_image_tokens = (input_ids == self.config.image_token_index).sum().item()
+            n_image_features = image_features.shape[0] * image_features.shape[1]
+
+            if n_image_tokens != n_image_features:
+                raise ValueError(
+                    f"Image features and image tokens do not match: tokens: {n_image_tokens}, features {n_image_features}"
+                )
+            special_image_mask = (
+                (input_ids == self.config.image_token_index)
+                .unsqueeze(-1)
+                .expand_as(inputs_embeds)
+                .to(inputs_embeds.device)
+            )
+            image_features = image_features.to(
+                inputs_embeds.device, inputs_embeds.dtype
+            )
+            inputs_embeds = inputs_embeds.masked_scatter(
+                special_image_mask, image_features
+            )
+
+        outputs = self.language_model(
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+            cache_position=cache_position,
+            num_logits_to_keep=num_logits_to_keep,
+        )
+
+        logits = outputs[0]
+
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            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[:, -(logits.shape[1] - 1) :].to(
+                    logits.device
+                )
+                shift_logits = logits[..., :-1, :][
+                    shift_attention_mask.to(logits.device) != 0
+                ].contiguous()
+                shift_labels = labels[..., 1:][
+                    shift_attention_mask.to(labels.device) != 0
+                ].contiguous()
+            else:
+                shift_logits = logits[..., :-1, :].contiguous()
+                shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = nn.CrossEntropyLoss()
+            loss = loss_fct(
+                shift_logits.view(-1, shift_logits.size(-1)),
+                shift_labels.view(-1).to(shift_logits.device),
+            )
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return (loss,) + output if loss is not None else output
+
+        return AriaCausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def prepare_inputs_for_generation(
+        self,
+        input_ids,
+        past_key_values=None,
+        inputs_embeds=None,
+        pixel_values=None,
+        pixel_mask=None,
+        attention_mask=None,
+        cache_position=None,
+        num_logits_to_keep=None,
+        **kwargs,
+    ):
+        model_inputs = self.language_model.prepare_inputs_for_generation(
+            input_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            attention_mask=attention_mask,
+            cache_position=cache_position,
+            num_logits_to_keep=num_logits_to_keep,
+            **kwargs,
+        )
+
+        if cache_position[0] == 0:
+            # 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
+            model_inputs["pixel_values"] = pixel_values
+            model_inputs["pixel_mask"] = pixel_mask
+
+        return model_inputs

moe_lm.py

@@ -0,0 +1,679 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+import logging
+import os
+from typing import Tuple
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch import nn
+from transformers import GenerationMixin, LlamaConfig
+from transformers.models.llama.modeling_llama import (
+    ACT2FN,
+    LLAMA_ATTENTION_CLASSES,
+    LlamaDecoderLayer,
+    LlamaForCausalLM,
+    LlamaMLP,
+    LlamaModel,
+    LlamaRMSNorm,
+    LlamaRotaryEmbedding,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class AriaMoELMConfig(LlamaConfig):
+    """
+    Configuration class for AriaMoE language model.
+
+    This class extends the LlamaConfig to include additional parameters specific to the Mixture of Experts (MoE) architecture.
+    """
+
+    model_type = "aria_moe_lm"
+
+    def __init__(
+        self,
+        moe_intermediate_size: int = 4096,
+        moe_num_experts: int = 8,
+        moe_topk: int = 2,
+        moe_z_loss_coeff: float = 1e-5,
+        moe_aux_loss_coeff: float = 1e-3,
+        moe_num_shared_experts: int = 2,
+        **kwargs,
+    ):
+        """
+        Initialize the AriaMoELMConfig.
+
+        Args:
+            moe_intermediate_size (int): The intermediate size for MoE layers. Default is 4096.
+            moe_num_experts (int): The number of experts in the MoE layer. Default is 8.
+            moe_topk (int): The number of top experts to route to for each token. Default is 2.
+            moe_z_loss_coeff (float): The coefficient for the auxiliary z-loss. Default is 1e-5.
+            moe_aux_loss_coeff (float): The coefficient for the auxiliary load balancing loss. Default is 1e-3.
+            moe_num_shared_experts (int): The number of shared experts. Default is 2.
+            **kwargs: Additional keyword arguments to be passed to the parent LlamaConfig.
+        """
+        super().__init__(**kwargs)
+        self.moe_intermediate_size = moe_intermediate_size
+        self.moe_num_experts = moe_num_experts
+        self.moe_topk = moe_topk
+        self.moe_z_loss_coeff = moe_z_loss_coeff
+        self.moe_aux_loss_coeff = moe_aux_loss_coeff
+        self.moe_num_shared_experts = moe_num_shared_experts
+
+
+# copied from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/moe_utils.py#L101-L142
+class MoEAuxLossAutoScaler(torch.autograd.Function):
+    """An AutoScaler that compute and scales the grad for auxiliary loss."""
+
+    main_loss_backward_scale: torch.Tensor = torch.tensor(1.0)
+
+    @staticmethod
+    def forward(ctx, output: torch.Tensor, aux_loss: torch.Tensor):
+        """Preserve the aux_loss by storing it in the context to avoid garbage collection.
+
+        Args:
+            output (torch.Tensor): The output tensor.
+            aux_loss (torch.Tensor): The auxiliary loss tensor.
+
+        Returns:
+            torch.Tensor: The output tensor.
+        """
+        ctx.save_for_backward(aux_loss)
+        return output
+
+    @staticmethod
+    def backward(ctx, grad_output: torch.Tensor):
+        """Compute and scale the gradient for auxiliary loss..
+
+        Args:
+            grad_output (torch.Tensor): The gradient of the output.
+
+        Returns:
+            Tuple[torch.Tensor, torch.Tensor]: The gradient of the output, scaled auxiliary loss gradient.
+        """
+        (aux_loss,) = ctx.saved_tensors
+        aux_loss_backward_scale = MoEAuxLossAutoScaler.main_loss_backward_scale
+        scaled_aux_loss_grad = torch.ones_like(aux_loss) * aux_loss_backward_scale
+        return grad_output, scaled_aux_loss_grad
+
+    @staticmethod
+    def set_loss_scale(scale: torch.Tensor):
+        """set the scale of the aux loss.
+
+        Args:
+            scale (torch.Tensor): The scale value to set. Please ensure that the scale passed in matches the scale of the main_loss.
+        """
+        MoEAuxLossAutoScaler.main_loss_backward_scale = scale
+
+
+def z_loss_func(logits, z_loss_coeff):
+    """Encourages the router's logits to remain small to enhance stability.
+    Please refer to the ST-MoE paper (https://arxiv.org/pdf/2202.08906.pdf) for details.
+
+    Args:
+        logits (torch.Tensor): The logits of the router.
+
+    Returns:
+        torch.Tensor: The logits after applying the z-loss.
+    """
+
+    z_loss = torch.mean(torch.square(torch.logsumexp(logits, dim=-1))) * z_loss_coeff
+    return z_loss
+
+
+def switch_load_balancing_loss_func(
+    probs: torch.Tensor,
+    tokens_per_expert: torch.Tensor,
+    topk: int,
+    moe_aux_loss_coeff: float,
+):
+    """Calculate the auxiliary loss for better load balancing.
+    Please refer to the Switch Transformer paper (https://arxiv.org/abs/2101.03961) for details.
+
+    Args:
+        probs (torch.Tensor): The softmax probs output by the router for each token. [num_tokens, num_experts]
+        tokens_per_expert (torch.Tensor): The number of assigned tokens for each expert. [num_experts]
+
+    Returns:
+        torch.Tensor: The auxiliary loss for load balancing.
+    """
+    num_tokens = probs.shape[0] * topk
+    num_experts = probs.shape[1]
+
+    probs_mean_per_expert = probs.mean(dim=0)
+    aux_loss = torch.sum(probs_mean_per_expert * tokens_per_expert) * (
+        num_experts / num_tokens * moe_aux_loss_coeff
+    )
+    return aux_loss
+
+
+# adapted from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/router.py#L96-L304
+class TopKRouter(nn.Module):
+    """
+    Top-K Router for Mixture of Experts (MoE) models.
+
+    This router determines which experts should process each token based on the top-k scoring experts.
+    It also applies auxiliary losses to encourage load balancing among experts.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object containing MoE-related parameters.
+    """
+
+    def __init__(self, config: AriaMoELMConfig):
+        super().__init__()
+        self.config = config
+
+        self.weight = nn.Parameter(
+            torch.empty((self.config.moe_num_experts, self.config.hidden_size))
+        )
+        # FIXME: initialize the weight
+
+    def gating(self, input: torch.Tensor) -> torch.Tensor:
+        """
+        Compute the gating logits for each token-expert pair.
+
+        Args:
+            input (torch.Tensor): Input tensor of shape [batch_size * seq_len, hidden_size].
+
+        Returns:
+            torch.Tensor: Logits tensor of shape [batch_size * seq_len, num_experts].
+        """
+        logits = torch.nn.functional.linear(input, self.weight)
+        return logits
+
+    def apply_z_loss(self, logits: torch.Tensor) -> torch.Tensor:
+        """
+        Apply z-loss to encourage router logits to remain small for enhanced stability.
+
+        Args:
+            logits (torch.Tensor): Router logits.
+
+        Returns:
+            torch.Tensor: Logits with z-loss applied.
+        """
+        z_loss = z_loss_func(logits, self.config.moe_z_loss_coeff)
+        logits = MoEAuxLossAutoScaler.apply(logits, z_loss)
+        return logits
+
+    def apply_aux_loss(
+        self,
+        logits: torch.Tensor,
+        tokens_per_expert: torch.Tensor,
+        activation: torch.Tensor,
+    ) -> torch.Tensor:
+        """
+        Apply auxiliary loss for load balancing among experts.
+
+        Args:
+            logits (torch.Tensor): Router logits.
+            tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
+            activation (torch.Tensor): Activation values.
+
+        Returns:
+            torch.Tensor: Activation with auxiliary loss applied.
+        """
+        probs = torch.softmax(logits, dim=-1, dtype=torch.float32)
+        aux_loss = switch_load_balancing_loss_func(
+            probs,
+            tokens_per_expert,
+            self.config.moe_topk,
+            self.config.moe_aux_loss_coeff,
+        )
+        return MoEAuxLossAutoScaler.apply(activation, aux_loss)
+
+    def routing(
+        self, logits: torch.Tensor
+    ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+        """
+        Perform the routing operation to determine expert assignments.
+
+        Args:
+            logits (torch.Tensor): Router logits.
+
+        Returns:
+            Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+                - scores: Softmax probabilities for top-k experts.
+                - top_indices: Indices of top-k experts for each token.
+                - tokens_per_expert: Number of tokens assigned to each expert.
+        """
+        if self.training:
+            logits = self.apply_z_loss(logits)
+
+        top_logits, top_indices = torch.topk(logits, k=self.config.moe_topk, dim=1)
+        scores = torch.softmax(top_logits, dim=-1, dtype=torch.float32).type_as(logits)
+
+        tokens_per_expert = torch.histc(
+            top_indices.flatten(),
+            bins=self.config.moe_num_experts,
+            min=0,
+            max=self.config.moe_num_experts - 1,
+        )
+
+        if self.training:
+            scores = self.apply_aux_loss(logits, tokens_per_expert, scores)
+        return scores, top_indices, tokens_per_expert
+
+    def forward(
+        self, input: torch.Tensor
+    ) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+        """
+        Forward pass of the TopKRouter.
+
+        Args:
+            input (torch.Tensor): Input tensor of shape [batch_size * seq_len, hidden_size].
+
+        Returns:
+            Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+                - scores: Softmax probabilities for top-k experts.
+                - top_indices: Indices of top-k experts for each token.
+                - tokens_per_expert: Number of tokens assigned to each expert.
+        """
+        logits = self.gating(input)
+        logits = logits.view(-1, self.config.moe_num_experts)
+        scores, top_indices, tokens_per_expert = self.routing(logits)
+        return scores, top_indices, tokens_per_expert
+
+
+# adapted from https://github.com/NVIDIA/Megatron-LM/blob/54f1f78529cbc2b9cddad313e7f9d96ac0420a27/megatron/core/transformer/moe/token_dispatcher.py#L291-L587
+class TokenDispatcher:
+    """
+    Handles the dispatching and gathering of tokens to and from experts.
+
+    This class is responsible for permuting tokens based on expert assignments and
+    unpermuting them after expert processing.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object containing MoE-related parameters.
+    """
+
+    def __init__(self, config: AriaMoELMConfig):
+        self.config = config
+        self.hidden_states_shape = None
+        self.reversed_input_permutation_mapping = None
+
+    def token_permutation(
+        self, hidden_states: torch.Tensor, indices: torch.Tensor
+    ) -> torch.Tensor:
+        """
+        Permute tokens based on expert assignments.
+
+        Args:
+            hidden_states (torch.Tensor): Input hidden states.
+            indices (torch.Tensor): Expert assignment indices.
+
+        Returns:
+            torch.Tensor: Permuted tokens.
+        """
+        self.hidden_states_shape = hidden_states.shape
+        hidden_states = hidden_states.view(-1, hidden_states.size(-1))
+        flatten_indices = indices.flatten()
+        sorted_indices = torch.argsort(flatten_indices, stable=True)
+        permuted_tokens = hidden_states.index_select(
+            0, sorted_indices // self.config.moe_topk
+        )
+        self.reversed_input_permutation_mapping = sorted_indices
+        return permuted_tokens
+
+    def token_unpermutation(
+        self, permuted_tokens: torch.Tensor, scores: torch.Tensor
+    ) -> torch.Tensor:
+        """
+        Unpermute tokens and combine expert outputs.
+
+        Args:
+            permuted_tokens (torch.Tensor): Tokens after expert processing.
+            scores (torch.Tensor): Expert assignment scores.
+
+        Returns:
+            torch.Tensor: Unpermuted and combined output.
+        """
+        num_unpermuted_tokens = scores.numel()
+        unpermuted_tokens = torch.zeros(
+            (num_unpermuted_tokens, permuted_tokens.size(1)),
+            dtype=permuted_tokens.dtype,
+            device=permuted_tokens.device,
+        )
+        unpermuted_tokens.index_copy_(
+            0, self.reversed_input_permutation_mapping, permuted_tokens
+        )
+        unpermuted_tokens = unpermuted_tokens.reshape(
+            -1, self.config.moe_topk, permuted_tokens.size(1)
+        )
+
+        unpermuted_tokens = unpermuted_tokens * scores.unsqueeze(-1)
+        unpermuted_tokens = unpermuted_tokens.sum(dim=1).type_as(permuted_tokens)
+        output = unpermuted_tokens.view(self.hidden_states_shape)
+        return output
+
+
+class SharedExpertMLP(LlamaMLP):
+    """
+    Shared Expert MLP for shared experts.
+
+    Unlike routed experts, shared experts process all tokens without routing.
+    This class reconfigures the intermediate size in comparison to the LlamaMLP.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object for the AriaMoE language model.
+    """
+
+    def __init__(self, config: AriaMoELMConfig):
+        nn.Module.__init__(self)
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.intermediate_size = (
+            config.moe_intermediate_size * config.moe_num_shared_experts
+        )
+        self.gate_proj = nn.Linear(
+            self.hidden_size, self.intermediate_size, bias=config.mlp_bias
+        )
+        self.up_proj = nn.Linear(
+            self.hidden_size, self.intermediate_size, bias=config.mlp_bias
+        )
+        self.down_proj = nn.Linear(
+            self.intermediate_size, self.hidden_size, bias=config.mlp_bias
+        )
+        self.act_fn = ACT2FN[config.hidden_act]
+
+
+def sequential_gemm(input, weight, tokens_per_expert):
+    """
+    Compute the matrix multiplication (GEMM) for each expert sequentially. This approach is computationally inefficient, especially when dealing with a large number of experts.
+
+    Args:
+        input (torch.Tensor): Input tensor of shape (num_tokens, in_features).
+        weight (torch.Tensor): Weight tensor of shape (num_experts, in_features, out_features).
+        tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
+
+    Returns:
+        torch.Tensor: Output tensor of shape (num_tokens, out_features).
+    """
+    num_tokens = input.shape[0]
+    out_features = weight.shape[-1]
+    output = torch.zeros(
+        num_tokens, out_features, dtype=input.dtype, device=input.device
+    )
+
+    cumsum_num_tokens = torch.cumsum(tokens_per_expert, dim=0)
+    # Insert zero at the begining for offset index's convenience
+    zero_tensor = torch.zeros(1, dtype=torch.long, device=cumsum_num_tokens.device)
+    cumsum_num_tokens = torch.cat((zero_tensor, cumsum_num_tokens))
+
+    for expert_num in range(weight.shape[0]):
+        start = cumsum_num_tokens[expert_num]
+        end = cumsum_num_tokens[expert_num + 1]
+        tokens = input[start:end]
+
+        out = torch.matmul(tokens, weight[expert_num])
+        output[start:end] = out
+    return output
+
+
+try:
+    from grouped_gemm.ops import gmm as experts_gemm
+
+    if os.environ.get("USE_GROUPED_GEMM", "1") == "0":
+        logger.warning(
+            "environment variable USE_GROUPED_GEMM is set to 0, using sequential GEMM instead."
+        )
+        experts_gemm = sequential_gemm
+except ImportError:
+    logger.warning(
+        "`grouped_gemm` is not installed, using sequential GEMM, which is slower."
+    )
+    experts_gemm = sequential_gemm
+
+
+class GroupedGEMM(nn.Module):
+    """
+    Grouped GEMM (General Matrix Multiplication) module for efficient expert computation.
+    This module utilizes the grouped_gemm library (https://github.com/fanshiqing/grouped_gemm)
+    for optimized performance. If the grouped_gemm library is not installed, it gracefully
+    falls back to a sequential GEMM implementation, which may be slower but ensures
+    functionality.
+
+    Args:
+        in_features (int): Number of input features.
+        out_features (int): Number of output features.
+        groups (int): Number of expert groups.
+    """
+
+    def __init__(self, in_features, out_features, groups):
+        super().__init__()
+        self.in_features = in_features
+        self.out_features = out_features
+        self.groups = groups
+        self.weight = nn.Parameter(torch.empty(groups, in_features, out_features))
+
+    def forward(self, input, tokens_per_expert):
+        """
+        Perform grouped matrix multiplication.
+
+        Args:
+            input (torch.Tensor): Input tensor of shape (num_tokens, in_features).
+            tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
+
+        Returns:
+            torch.Tensor: Output tensor of shape (num_tokens, out_features).
+        """
+        tokens_per_expert = tokens_per_expert.cpu()
+
+        # Ensure the CUDA device matches the input tensor's device.
+        # This mismatch can occur when using `transformers.AutoModel.from_pretrained`
+        # with `device_map="auto"` on a multi-GPU setup.
+        torch.cuda.set_device(input.device)
+        return experts_gemm(input, self.weight, tokens_per_expert)
+
+
+class GroupedMLP(nn.Module):
+    """
+    Grouped MLP module for Mixture of Experts.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object for the model.
+    """
+
+    def __init__(self, config: AriaMoELMConfig) -> None:
+        super().__init__()
+        self.config = config
+        self.fc1 = GroupedGEMM(
+            config.hidden_size, config.moe_intermediate_size * 2, config.moe_num_experts
+        )
+        self.fc2 = GroupedGEMM(
+            config.moe_intermediate_size, config.hidden_size, config.moe_num_experts
+        )
+
+        def glu(x):
+            x = torch.chunk(x, 2, dim=-1)
+            return F.silu(x[0]) * x[1]
+
+        self.activation_func = glu
+
+    def forward(self, permuted_tokens, tokens_per_expert):
+        """
+        Forward pass of the Grouped MLP.
+
+        Args:
+            permuted_tokens (torch.Tensor): Permuted input tokens.
+            tokens_per_expert (torch.Tensor): Number of tokens assigned to each expert.
+
+        Returns:
+            torch.Tensor: Output tensor after passing through the MLP.
+        """
+        fc1_output = self.fc1(permuted_tokens, tokens_per_expert)
+        fc1_output = self.activation_func(fc1_output)
+        fc2_output = self.fc2(fc1_output, tokens_per_expert)
+        return fc2_output
+
+
+class MoELayer(nn.Module):
+    """
+    Mixture of Experts (MoE) Layer for the AriaMoE model.
+
+    This layer implements the MoE mechanism, which routes input tokens to different experts
+    based on a routing algorithm, processes them through the experts, and then combines
+    the outputs.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object for the MoE layer.
+    """
+
+    def __init__(self, config: AriaMoELMConfig):
+        super().__init__()
+
+        self.router = TopKRouter(config)
+        self.token_dispatcher = TokenDispatcher(config)
+        self.experts = GroupedMLP(config)
+        self.shared_experts = SharedExpertMLP(config)
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        """
+        Forward pass of the MoE Layer.
+
+        Args:
+            hidden_states (torch.Tensor): Input tensor of shape (batch_size, sequence_length, hidden_size).
+
+        Returns:
+            torch.Tensor: Output tensor after passing through the MoE layer.
+
+        Process:
+        1. Route tokens to experts using the router.
+        2. Permute tokens based on routing decisions.
+        3. Process tokens through experts.
+        4. Unpermute and combine expert outputs.
+        5. Add shared expert output to the final result.
+        """
+        scores, indices, tokens_per_expert = self.router(hidden_states)
+
+        permuted_tokens = self.token_dispatcher.token_permutation(
+            hidden_states, indices
+        )
+
+        expert_output = self.experts(permuted_tokens, tokens_per_expert)
+
+        output = self.token_dispatcher.token_unpermutation(expert_output, scores)
+
+        shared_expert_output = self.shared_experts(hidden_states)
+        output += shared_expert_output
+        return output
+
+
+class MoEDecoderLayer(LlamaDecoderLayer):
+    """
+    Custom Decoder Layer for the AriaMoE model which modifies the standard `LlamaDecoderLayer` by
+    replacing the traditional MLP with a Mixture of Experts (MoE) Layer.
+
+    Args:
+        config (LlamaConfig): Configuration object for the layer.
+        layer_idx (int): Index of the current layer in the model.
+    """
+
+    def __init__(self, config: LlamaConfig, layer_idx: int):
+        nn.Module.__init__(self)
+        self.hidden_size = config.hidden_size
+
+        self.self_attn = LLAMA_ATTENTION_CLASSES[config._attn_implementation](
+            config=config, layer_idx=layer_idx
+        )
+
+        self.mlp = MoELayer(config)
+        self.input_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = LlamaRMSNorm(
+            config.hidden_size, eps=config.rms_norm_eps
+        )
+
+
+class AriaMoELMModel(LlamaModel):
+    """
+    Custom LlamaModel for the AriaMoE model which modifies the standard LlamaModel by
+    replacing the `LlamaDecoderLayer` with `MoEDecoderLayer`.
+
+    This model implements a Mixture of Experts (MoE) approach, where each layer contains
+    multiple expert networks that specialize in different aspects of the input.
+
+    Args:
+        config (LlamaConfig): Configuration object for the model.
+    """
+
+    def __init__(self, config: LlamaConfig):
+        super().__init__(config)
+        self.padding_idx = config.pad_token_id
+        self.vocab_size = config.vocab_size
+
+        self.embed_tokens = nn.Embedding(
+            config.vocab_size, config.hidden_size, self.padding_idx
+        )
+        self.layers = nn.ModuleList(
+            [
+                MoEDecoderLayer(config, layer_idx)
+                for layer_idx in range(config.num_hidden_layers)
+            ]
+        )
+        self.norm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.rotary_emb = LlamaRotaryEmbedding(config=config)
+        self.gradient_checkpointing = False
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+
+class AriaMoELMForCausalLM(LlamaForCausalLM, GenerationMixin):
+    """
+    AriaMoE model for causal language modeling tasks.
+
+    This class extends LlamaForCausalLM to incorporate the Mixture of Experts (MoE) approach,
+    allowing for more efficient and scalable language modeling.
+
+    Args:
+        config (AriaMoELMConfig): Configuration object for the model.
+    """
+
+    _tied_weights_keys = ["lm_head.weight"]
+    config_class = AriaMoELMConfig
+    _no_split_modules = ["MoEDecoderLayer"]
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.model = AriaMoELMModel(config)
+        self.vocab_size = config.vocab_size
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def set_z_loss_coeff(self, z_loss_coeff: float):
+        """
+        Set the coefficient for the z-loss in the MoE routing.
+
+        Args:
+            z_loss_coeff (float): The coefficient for the z-loss.
+        """
+        self.config.moe_z_loss_coeff = z_loss_coeff
+
+    def set_aux_loss_coeff(self, aux_loss_coeff: float):
+        """
+        Set the coefficient for the auxiliary loss in the MoE routing.
+
+        Args:
+            aux_loss_coeff (float): The coefficient for the auxiliary loss.
+        """
+        self.config.moe_aux_loss_coeff = aux_loss_coeff

preprocessor_config.json

@@ -0,0 +1,21 @@
+{
+  "_transform": null,
+  "auto_map": {
+    "AutoImageProcessor": "vision_processor.AriaVisionProcessor",
+    "AutoProcessor": "processing_aria.AriaProcessor"
+  },
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "AriaVisionProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "max_image_size": 980,
+  "min_image_size": 336,
+  "processor_class": "AriaProcessor"
+}

processing_aria.py

@@ -0,0 +1,305 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+import inspect
+import logging
+import re
+from typing import List, Optional, Union
+
+from transformers import AutoTokenizer, BatchFeature
+from transformers.image_utils import ImageInput
+from transformers.processing_utils import ProcessorMixin
+from transformers.tokenization_utils import (
+    PaddingStrategy,
+    PreTokenizedInput,
+    TensorType,
+    TextInput,
+    TruncationStrategy,
+)
+
+from .vision_processor import AriaVisionProcessor
+
+logger = logging.getLogger(__name__)
+
+
+class AriaProcessor(ProcessorMixin):
+    """
+    AriaProcessor is a processor for the Aria model which wraps the Aria image preprocessor and the LLama slow tokenizer.
+    Args:
+        image_processor(AriaVisionProcessor): The AriaVisionProcessor to use for image preprocessing.
+        tokenizer(AutoTokenizer): The AutoTokenizer to use for tokenizing the text.
+        patch_size(int): The patch size to use for the image processor.
+        chat_template(str): The chat template to use for the tokenizer.
+        image_token(str): The image token to use for the tokenizer.
+    """
+
+    attributes = []
+    valid_kwargs = ["chat_template", "patch_size", "image_token"]
+    image_processor_class = None
+    tokenizer_class = "AutoTokenizer"
+
+    def __init__(
+        self,
+        image_processor: AriaVisionProcessor = None,
+        tokenizer: Union[AutoTokenizer, str] = None,
+        patch_size: int = 490,
+        chat_template: str = None,
+        image_token: str = "<|img|>",
+    ):
+        super().__init__(chat_template=chat_template)
+
+        if image_processor is None:
+            self.image_processor = AriaVisionProcessor(max_image_size=patch_size)
+        else:
+            self.image_processor = image_processor
+
+        if isinstance(tokenizer, str):
+            self.tokenizer = AutoTokenizer.from_pretrained(
+                tokenizer, trust_remote_code=True, use_fast=False
+            )
+        else:
+            self.tokenizer = tokenizer
+
+        if self.tokenizer is not None and self.tokenizer.pad_token is None:
+            self.tokenizer.pad_token = self.tokenizer.unk_token
+
+        self.image_token = image_token
+
+    # Copied from transformers.models.llava_next.processing_llave_next.LlavaNextProcessor.__call__
+    def __call__(
+        self,
+        text: Union[
+            TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]
+        ],
+        images: ImageInput = None,
+        padding: Union[bool, str, PaddingStrategy] = False,
+        truncation: Union[bool, str, TruncationStrategy] = None,
+        max_length: Optional[int] = None,
+        max_image_size: Optional[int] = 980,
+        split_image: Optional[bool] = False,
+        return_tensors: Optional[Union[str, TensorType]] = TensorType.PYTORCH,
+        return_final_prompts: Optional[bool] = False,
+    ) -> BatchFeature:
+        """
+        Main method to prepare for the model one or several sequences(s) and image(s). Please refer to the doctsring
+        of the above two methods for more information.
+
+        Args:
+            text (`str`, `List[str]`, `List[List[str]]`):
+                The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings
+                (pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set
+                `is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
+            images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
+                The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
+                tensor. Both channels-first and channels-last formats are supported.
+            padding (`bool`, `str` or [`~utils.PaddingStrategy`], *optional*, defaults to `False`):
+                Select a strategy to pad the returned sequences (according to the model's padding side and padding
+                index) among:
+                - `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
+                  sequence if provided).
+                - `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
+                  acceptable input length for the model if that argument is not provided.
+                - `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
+                  lengths).
+            max_length (`int`, *optional*):
+                Maximum length of the returned list and optionally padding length (see above).
+            max_image_size (`int`, *optional*):
+                Maximum size of the image to be processed.
+            split_image (`bool`, *optional*):
+                Whether to split the image into patches before processing.
+            truncation (`bool`, *optional*):
+                Activates truncation to cut input sequences longer than `max_length` to `max_length`.
+            return_tensors (`str` or [`~utils.TensorType`], *optional*):
+                If set, will return tensors of a particular framework. Acceptable values are:
+
+                - `'tf'`: Return TensorFlow `tf.constant` objects.
+                - `'pt'`: Return PyTorch `torch.Tensor` objects.
+                - `'np'`: Return NumPy `np.ndarray` objects.
+                - `'jax'`: Return JAX `jnp.ndarray` objects.
+
+        Returns:
+            [`BatchFeature`]: A [`BatchFeature`] with the following fields:
+
+            - **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.
+            - **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
+              `return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not
+              `None`).
+            - **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.
+            - **pixel_mask** -- Pixel mask to be fed to a model. Returned when `images` is not `None`.
+        """
+        if isinstance(text, str):
+            text = [text]
+        elif not isinstance(text, list) and not isinstance(text[0], str):
+            raise ValueError(
+                "Invalid input text. Please provide a string, or a list of strings"
+            )
+
+        if images is not None:
+            image_inputs = self.image_processor(
+                images,
+                return_tensors=return_tensors,
+                max_image_size=max_image_size,
+                split_image=split_image,
+            )
+            # expand the image_token according to the num_crops of image
+            prompt_strings = []
+            crop_iter = iter(image_inputs.pop("num_crops"))
+            for prompt in text:
+                prompt_strings.append(
+                    re.sub(
+                        re.escape(self.image_token),
+                        lambda _: next(crop_iter) * self.image_token,
+                        prompt,
+                    )
+                )
+
+            max_image_size = (
+                max_image_size
+                if max_image_size is not None
+                else self.image_processor.max_image_size
+            )
+            if max_image_size == 490:
+                num_image_tokens = 128
+            elif max_image_size == 980:
+                num_image_tokens = 256
+            else:
+                raise ValueError(
+                    f"max_image_size must be either 490 or 980, got {max_image_size}"
+                )
+            prompt_strings = [
+                sample.replace(self.image_token, self.image_token * num_image_tokens)
+                for sample in prompt_strings
+            ]
+
+        else:
+            image_inputs = {}
+            prompt_strings = text
+
+        text_inputs = self.tokenizer(
+            prompt_strings,
+            return_tensors=return_tensors,
+            padding=padding,
+            truncation=truncation,
+            max_length=max_length,
+        )
+
+        if return_final_prompts:
+            return BatchFeature(data={**text_inputs, **image_inputs}), prompt_strings
+        else:
+            return BatchFeature(data={**text_inputs, **image_inputs})
+
+    @staticmethod
+    def _extract_kwargs(func: callable, **kwargs) -> dict:
+        """
+        Extract the kwargs that are valid for the given function.
+        """
+        return {
+            k: v for k, v in kwargs.items() if k in inspect.signature(func).parameters
+        }
+
+    def save_pretrained(self, save_directory, **kwargs):
+        """
+        Save both the image processor and tokenizer.
+        """
+        if self.image_processor is not None:
+            self.image_processor.save_pretrained(
+                save_directory,
+                **self._extract_kwargs(self.image_processor.save_pretrained, **kwargs),
+            )
+        if self.tokenizer is not None:
+            self.tokenizer.save_pretrained(
+                save_directory,
+                **self._extract_kwargs(self.tokenizer.save_pretrained, **kwargs),
+            )
+
+    @classmethod
+    def from_pretrained(
+        cls,
+        pretrained_model_name_or_path,
+        tokenizer_path=None,
+        image_processor_path=None,
+        **kwargs,
+    ):
+        """
+        Load both the image processor and tokenizer from a pretrained model path.
+        """
+        tokenizer_path = (
+            tokenizer_path
+            if tokenizer_path is not None
+            else pretrained_model_name_or_path
+        )
+        image_processor_path = (
+            image_processor_path
+            if image_processor_path is not None
+            else pretrained_model_name_or_path
+        )
+        image_processor = AriaVisionProcessor.from_pretrained(
+            image_processor_path,
+            **cls._extract_kwargs(AriaVisionProcessor.from_pretrained, **kwargs),
+        )
+        if "use_fast" in kwargs:
+            logger.warning("use_fast is not supported for AriaProcessor. Ignoring...")
+            kwargs.pop("use_fast")
+        try:
+            tokenizer = AutoTokenizer.from_pretrained(
+                tokenizer_path,
+                use_fast=False,
+                **cls._extract_kwargs(AutoTokenizer.from_pretrained, **kwargs),
+            )
+            chat_template = tokenizer.chat_template
+        except Exception as e:
+            logger.warning(f"Failed to load tokenizer from {tokenizer_path}: {e}")
+            tokenizer = None
+            chat_template = None
+        return cls(
+            image_processor=image_processor,
+            tokenizer=tokenizer,
+            chat_template=chat_template,
+        )
+
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.batch_decode with CLIP->Llama
+    def batch_decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
+        refer to the docstring of this method for more information.
+        """
+        if self.tokenizer is None:
+            raise ValueError(
+                "Tokenizer is not initialized. Please provide a valid tokenizer."
+            )
+        return self.tokenizer.batch_decode(*args, **kwargs)
+
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.decode with CLIP->Llama
+    def decode(self, *args, **kwargs):
+        """
+        This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
+        the docstring of this method for more information.
+        """
+        if self.tokenizer is None:
+            raise ValueError(
+                "Tokenizer is not initialized. Please provide a valid tokenizer."
+            )
+        return self.tokenizer.decode(*args, **kwargs)
+
+    @property
+    # Copied from transformers.models.clip.processing_clip.CLIPProcessor.model_input_names
+    def model_input_names(self):
+        tokenizer_input_names = self.tokenizer.model_input_names
+        image_processor_input_names = self.image_processor.model_input_names
+        return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))

projector.py

@@ -0,0 +1,189 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+import torch
+import torch.nn as nn
+from torch.nn.init import trunc_normal_
+from transformers.activations import ACT2FN
+
+
+class FFN(nn.Module):
+    """
+    Feed-Forward Network module.
+
+    Args:
+        embed_dim (int): Input embedding dimension.
+        ff_dim (int): Hidden dimension of the feed-forward network.
+        output_dim (int): Output dimension.
+    """
+
+    def __init__(self, embed_dim, ff_dim, output_dim):
+        super().__init__()
+        self.linear_in = nn.Linear(embed_dim, ff_dim, bias=False)
+        self.linear_out = nn.Linear(ff_dim, output_dim, bias=False)
+        self.act = ACT2FN["gelu_new"]
+
+    def forward(self, hidden_states):
+        hidden_states = self.act(self.linear_in(hidden_states))
+        hidden_states = self.linear_out(hidden_states)
+        return hidden_states
+
+
+class CrossAttention(nn.Module):
+    """
+    Cross-Attention module.
+
+    Args:
+        kv_dim (int): Dimension of key and value.
+        embed_dim (int): Embedding dimension.
+        num_heads (int): Number of attention heads.
+        drop_out_rate (float): Dropout rate. Default is 0.
+    """
+
+    def __init__(self, kv_dim, embed_dim, num_heads, drop_out_rate=0):
+        super().__init__()
+        self.num_heads = num_heads
+        self.q_proj = nn.Linear(embed_dim, embed_dim, bias=False)
+        self.k_proj = nn.Linear(kv_dim, embed_dim, bias=False)
+        self.v_proj = nn.Linear(kv_dim, embed_dim, bias=False)
+
+        self.multihead_attn = nn.MultiheadAttention(embed_dim, num_heads)
+        self.linear = nn.Linear(embed_dim, embed_dim)
+        self.dropout = nn.Dropout(drop_out_rate)
+
+        self.layer_norm = nn.LayerNorm(embed_dim)
+        self.ln_kv = nn.LayerNorm(kv_dim)
+
+    def forward(self, x, hidden_states, attn_mask=None, add_residual=False):
+        """
+        Forward pass of the CrossAttention module.
+
+        Args:
+            x (torch.Tensor): Input tensor for key and value.
+            hidden_states (torch.Tensor): Input tensor for query.
+            attn_mask (torch.Tensor, optional): Attention mask. Default is None.
+            add_residual (bool): Whether to add residual connection. Default is False.
+
+        Returns:
+            torch.Tensor: Output tensor after cross-attention.
+        """
+        normed_hidden_states = self.layer_norm(hidden_states)
+        query = self.q_proj(normed_hidden_states).permute(1, 0, 2)
+
+        x = self.ln_kv(x)
+        key = self.k_proj(x).permute(1, 0, 2)
+        value = self.v_proj(x).permute(1, 0, 2)
+
+        attn_output, _ = self.multihead_attn(query, key, value, attn_mask=attn_mask)
+
+        attn_output = attn_output.permute(1, 0, 2)
+
+        if add_residual:
+            attn_output = hidden_states + self.dropout(self.linear(attn_output))
+        else:
+            attn_output = self.dropout(self.linear(attn_output))
+
+        return attn_output
+
+
+class AriaProjector(nn.Module):
+    """
+    A projection module with one cross attention layer and one FFN layer, which projects ViT's outputs into MoE's inputs.
+
+    Args:
+        patch_to_query_dict (dict): Maps patch numbers to their corresponding query numbers,
+            e.g., {1225: 128, 4900: 256}. This allows for different query sizes based on image resolution.
+        embed_dim (int): Embedding dimension.
+        num_heads (int): Number of attention heads.
+        kv_dim (int): Dimension of key and value.
+        ff_dim (int): Hidden dimension of the feed-forward network.
+        output_dim (int): Output dimension.
+        norm_layer (nn.Module): Normalization layer. Default is nn.LayerNorm.
+
+    Outputs:
+        A tensor with the shape of (batch_size, query_number, output_dim)
+    """
+
+    def __init__(
+        self,
+        patch_to_query_dict,
+        embed_dim,
+        num_heads,
+        kv_dim,
+        ff_dim,
+        output_dim,
+        norm_layer=nn.LayerNorm,
+    ):
+        super().__init__()
+        self.patch_to_query_dict = patch_to_query_dict
+        self.embed_dim = embed_dim
+        self.num_heads = num_heads
+
+        self.query = nn.Parameter(
+            torch.zeros(max(patch_to_query_dict.values()), self.embed_dim)
+        )
+
+        trunc_normal_(self.query, std=0.02)
+
+        self.cross_attn = CrossAttention(kv_dim, embed_dim, num_heads)
+
+        self.ln_ffn = norm_layer(embed_dim)
+        self.ffn = FFN(embed_dim, ff_dim, output_dim)
+
+        self.apply(self._init_weights)
+
+    def _init_weights(self, m):
+        if isinstance(m, nn.Linear):
+            trunc_normal_(m.weight, std=0.02)
+            if isinstance(m, nn.Linear) and m.bias is not None:
+                nn.init.constant_(m.bias, 0)
+        elif isinstance(m, nn.LayerNorm):
+            nn.init.constant_(m.bias, 0)
+            nn.init.constant_(m.weight, 1.0)
+
+    def forward(self, x, attn_mask=None):
+        """
+        Forward pass of the Projector module.
+
+        Args:
+            x (torch.Tensor): Input tensor of shape (batch_size, num_patches, kv_dim).
+            attn_mask (torch.Tensor, optional): Attention mask. Default is None.
+
+        Returns:
+            torch.Tensor: Output tensor of shape (batch_size, query_number, output_dim).
+        """
+        bs = x.shape[0]
+        queries = self.query.unsqueeze(0).repeat(bs, 1, 1)
+
+        query_num = self.patch_to_query_dict.get(x.shape[1], None)
+        assert (
+            query_num is not None
+        ), f"Query number for {x.shape[1]} patches is not provided"
+
+        queries = queries[:, :query_num, :]
+
+        if attn_mask is not None:
+            attn_mask = attn_mask.repeat_interleave(self.num_heads, 0)
+            attn_mask = attn_mask.unsqueeze(1).expand(-1, queries.size(1), -1)
+
+        attention_out = self.cross_attn(x, queries, attn_mask=attn_mask)
+
+        out = self.ffn(self.ln_ffn(attention_out))
+
+        return out

vision_encoder.py

@@ -0,0 +1,152 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+"""PyTorch Aria vision transformer."""
+
+from typing import Optional, Tuple, Union
+
+import torch
+import torch.utils.checkpoint
+from transformers import SiglipVisionConfig, SiglipVisionModel
+from transformers.modeling_outputs import BaseModelOutputWithPooling
+from transformers.models.idefics2.modeling_idefics2 import Idefics2VisionTransformer
+
+
+class AriaVisionConfig(SiglipVisionConfig):
+    """Configuration class for AriaVisionModel."""
+
+    model_type = "aria_vision_model"
+
+    def __init__(
+        self,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+
+
+class IdentityOp(torch.nn.Module):
+    """
+    An identity operation that returns the input unchanged.
+
+    This can be used as a placeholder or to maintain architectural consistency
+    when a specific operation is not needed.
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__()
+
+    def forward(self, x, *args, **kwargs):
+        return x
+
+
+class AriaVisionTransformer(Idefics2VisionTransformer):
+    """
+    Aria Vision Transformer model based on Idefics2VisionTransformer.
+
+    This class extends the original Idefics2VisionTransformer by removing the post-layernorm operation.
+    """
+
+    def __init__(self, config: AriaVisionConfig):
+        super().__init__(config)
+        self.post_layernorm = IdentityOp()
+
+
+class AriaVisionModel(SiglipVisionModel):
+    """
+    Aria Vision Model extends SiglipVisionModel to support pixel_mask.
+
+    The pixel_mask is a 2D boolean tensor that indicates which pixels in the input
+    image are actual content and which are padding. It has the same height and width
+    as the input image, where:
+    - True (1) values represent pixels from the original image
+    - False (0) values represent padding pixels
+
+    This mask helps the model focus on the relevant parts of the image during processing.
+    """
+
+    config_class = AriaVisionConfig
+    main_input_name = "pixel_values"
+    _supports_sdpa = False
+
+    def __init__(self, config: AriaVisionConfig):
+        super().__init__(config)
+        self.vision_model = AriaVisionTransformer(config)
+
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def forward(
+        self,
+        pixel_values: torch.Tensor,
+        pixel_mask: Optional[torch.BoolTensor] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPooling]:
+        """
+        Forward pass of the AriaVisionModel.
+
+        Args:
+            pixel_values (torch.Tensor): The pixel values of the input images.
+            pixel_mask (Optional[torch.BoolTensor]): Mask for the pixel values.
+            output_attentions (Optional[bool]): Whether to output attentions.
+            output_hidden_states (Optional[bool]): Whether to output hidden states.
+            return_dict (Optional[bool]): Whether to return a ModelOutput object.
+
+        Returns:
+            Union[Tuple, BaseModelOutputWithPooling]: The model's output.
+        """
+        return_dict = (
+            return_dict if return_dict is not None else self.config.use_return_dict
+        )
+        patch_attention_mask = self._create_patch_attention_mask(pixel_mask)
+
+        vit_oup = self.vision_model(
+            pixel_values=pixel_values,
+            patch_attention_mask=patch_attention_mask,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            return_dict=return_dict,
+        )
+
+        image_atts = self._create_image_attention_mask(patch_attention_mask)
+
+        return vit_oup, image_atts
+
+    def _create_patch_attention_mask(self, pixel_mask):
+        if pixel_mask is None:
+            return None
+
+        patches_subgrid = pixel_mask.unfold(
+            dimension=1,
+            size=self.vision_model.config.patch_size,
+            step=self.vision_model.config.patch_size,
+        ).unfold(
+            dimension=2,
+            size=self.vision_model.config.patch_size,
+            step=self.vision_model.config.patch_size,
+        )
+        return (patches_subgrid.sum(dim=(-1, -2)) > 0).bool()
+
+    def _create_image_attention_mask(self, patch_attention_mask):
+        if patch_attention_mask is None:
+            return None
+
+        flattened_mask = patch_attention_mask.flatten(1)
+        return torch.logical_not(flattened_mask)

vision_processor.py

@@ -0,0 +1,321 @@
+# Copyright 2024 Rhymes AI. All rights reserved.
+#
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+
+from typing import List, Optional, Union
+
+import numpy as np
+import torch
+from PIL import Image, ImageOps
+from torchvision import transforms
+from transformers import BaseImageProcessor, BatchFeature, TensorType
+
+
+def _select_best_resolution(
+    img_width: int, img_height: int, target_ratios: List[List[int]], patch_size: int
+):
+    """
+    Selects the best resolution from a list of possible resolutions based on the original size.
+
+    Args:
+        img_width: the original widths of images.
+        img_height: the original heights of images.
+        target_ratios (2d numpy array): dimension size (M,2)
+        patch_size (int): image patch size
+
+    Returns:
+        tuple: The best fit resolution in the format (width, height).
+    """
+
+    aspect_ratio = img_width / img_height
+    best_ratio_diff = float("inf")
+    best_ratio_w, best_ratio_h = 1, 1
+    area = np.int32(img_width) * np.int32(img_height)
+    for ratio in target_ratios:
+        target_aspect_ratio = ratio[0] / ratio[1]
+        ratio_diff = abs(aspect_ratio - target_aspect_ratio)
+        if ratio_diff < best_ratio_diff:
+            best_ratio_diff = ratio_diff
+            best_ratio_w, best_ratio_h = ratio[0], ratio[1]
+        elif (
+            ratio_diff == best_ratio_diff
+            and area > 0.5 * patch_size * patch_size * ratio[0] * ratio[1]
+        ):
+            best_ratio_w, best_ratio_h = ratio[0], ratio[1]
+
+    return best_ratio_w, best_ratio_h
+
+
+def _split_image(
+    image: Image.Image,
+    split_image: bool,
+    split_ratio: List[List[int]],
+    patch_size: int,
+) -> List[Image.Image]:
+    """
+    Split image into multiple patches
+
+    Args:
+        image (PIL.Image): Input image.
+        split_image (bool): Whether to split the image into patches.
+        split_ratio (2d numpy array): dimension size (M,2)
+        patch_size (int): image patch size
+
+    Returns:
+        List[PIL.Image]: List of splitted images.
+    """
+    if split_image:
+        ratio_width, ratio_height = _select_best_resolution(
+            image.width, image.height, split_ratio, patch_size
+        )
+        resize_width = patch_size * ratio_width
+        resize_height = patch_size * ratio_height
+        blocks = ratio_width * ratio_height
+        resized_img = image.resize((resize_width, resize_height))
+        processed_images = []
+        for i in range(blocks):
+            box = (
+                (i % (resize_width // patch_size)) * patch_size,
+                (i // (resize_width // patch_size)) * patch_size,
+                ((i % (resize_width // patch_size)) + 1) * patch_size,
+                ((i // (resize_width // patch_size)) + 1) * patch_size,
+            )
+            # split the image
+            split_img = resized_img.crop(box)
+            processed_images.append(split_img)
+        assert len(processed_images) == blocks
+        if len(processed_images) != 1:
+            processed_images.insert(0, image)
+        return processed_images
+    else:
+        return [image]
+
+
+def keep_ratio_resize_and_pixel_mask(
+    img: Image.Image, max_size, min_size=336, padding_value=0
+):
+    """
+    Resize an image while maintaining aspect ratio and create a pixel mask.
+
+    Args:
+        img (PIL.Image): Input image.
+        max_size (int): Maximum size for the larger dimension of the image.
+        min_size (int, optional): Minimum size for the smaller dimension. Defaults to 336.
+        padding_value (int, optional): Value used for padding. Defaults to 0.
+
+    Returns:
+        tuple: A tuple containing:
+            - PIL.Image: Resized and padded image.
+            - torch.Tensor: Boolean pixel mask. This mask is a 2D tensor of shape (max_size, max_size) where:
+                - True (1) values indicate pixels that belong to the original resized image.
+                - False (0) values indicate pixels that are part of the padding.
+              The mask helps distinguish between actual image content and padded areas in subsequent processing steps.
+    """
+    img = img.convert("RGB")
+    # rescale the given image, keep the aspect ratio
+    scale = max_size / max(img.size)
+
+    w, h = img.size
+    if w >= h:
+        new_size = (max_size, max(int(h * scale), min_size))  # w, h
+    else:
+        new_size = (max(int(w * scale), min_size), max_size)  # w, h
+
+    img_resized = img.resize(new_size, resample=Image.Resampling.BICUBIC)
+
+    # padding the right/bottom
+    padding_right, padding_bottom = max_size - new_size[0], max_size - new_size[1]
+    img_padded = ImageOps.expand(
+        img_resized, (0, 0, padding_right, padding_bottom), fill=padding_value
+    )
+
+    # Create a pixel mask
+    pixel_mask = torch.zeros(max_size, max_size)
+    pixel_mask[: new_size[1], : new_size[0]] = 1
+    pixel_mask = pixel_mask.bool()
+    return img_padded, pixel_mask
+
+
+class AriaVisionProcessor(BaseImageProcessor):
+    """
+    A vision processor for the Aria model that handles image preprocessing.
+    """
+
+    def __init__(
+        self,
+        max_image_size=980,
+        min_image_size=336,
+        image_mean=[0.5, 0.5, 0.5],
+        image_std=[0.5, 0.5, 0.5],
+        **kwargs,
+    ):
+        """
+        Initialize the AriaVisionProcessor.
+
+        Args:
+            max_image_size (int, optional): Maximum image size. Defaults to 980.
+            min_image_size (int, optional): Minimum image size. Defaults to 336.
+            mean (list, optional): Mean values for normalization. Defaults to [0.5, 0.5, 0.5].
+            std (list, optional): Standard deviation values for normalization. Defaults to [0.5, 0.5, 0.5].
+        """
+        super().__init__(**kwargs)
+
+        self.max_image_size = max_image_size
+        self.min_image_size = min_image_size
+        self.image_mean = image_mean
+        self.image_std = image_std
+        self.auto_map = {
+            "AutoProcessor": "processing_aria.AriaProcessor",
+            "AutoImageProcessor": "vision_processor.AriaVisionProcessor",
+        }
+
+        # we make the transform a property so that it is lazily initialized,
+        # this could avoid the error "TypeError: Object of type Normalize is not JSON serializable"
+        # when we used save_pretrained or from_pretrained.
+        self._transform = None
+        self._set_processor_class("AriaProcessor")
+
+    @property
+    def transform(self):
+        if self._transform is None:
+            # Recreate the transform when accessed
+            self._transform = transforms.Compose(
+                [
+                    transforms.ToTensor(),
+                    transforms.Normalize(self.image_mean, self.image_std),
+                ]
+            )
+        return self._transform
+
+    def __call__(
+        self,
+        images: Union[Image.Image, List[Image.Image]],
+        max_image_size: Optional[int] = 980,
+        min_image_size: Optional[int] = 336,
+        return_tensors: Optional[Union[str, TensorType]] = "pt",
+        split_image: Optional[bool] = False,
+        split_ratio: Optional[List[List[int]]] = [
+            [1, 2],
+            [1, 3],
+            [1, 4],
+            [1, 5],
+            [1, 6],
+            [1, 7],
+            [1, 8],
+            [2, 4],
+            [2, 3],
+            [2, 2],
+            [2, 1],
+            [3, 1],
+            [3, 2],
+            [4, 1],
+            [4, 2],
+            [5, 1],
+            [6, 1],
+            [7, 1],
+            [8, 1],
+        ],
+    ):
+        """
+        Process a list of images.
+
+        Args:
+            images (list): List of PIL.Image objects.
+            max_image_size (int, optional): Override the default max image size. Defaults to None.
+            return_tensors (str or TensorType, optional): The type of tensor to return. Defaults to "pt".
+            split_image (bool, optional): Whether to split the image. Defaults to False.
+            split_ratio (list, optional): The ratio for splitting the image. Defaults to a list of common split ratios.
+        Returns:
+            BatchFeature: A BatchFeature object containing:
+                - 'pixel_values': Tensor of processed image pixel values.
+                - 'pixel_mask': Boolean pixel mask. This mask is a 2D tensor of shape (max_size, max_size) where:
+                    - True (1) values indicate pixels that belong to the original resized image.
+                    - False (0) values indicate pixels that are part of the padding.
+                  The mask helps distinguish between actual image content and padded areas in subsequent processing steps.
+                - 'num_crops': Tensor of the number of crops for each image.
+        """
+        max_size = self.max_image_size if max_image_size is None else max_image_size
+        min_size = self.min_image_size if min_image_size is None else min_image_size
+
+        if max_size not in [490, 980]:
+            raise ValueError("max_image_size must be either 490 or 980")
+
+        if isinstance(images, Image.Image):
+            images = [images]
+
+        pixel_values = []
+        pixel_masks = []
+        num_crops = []
+
+        for image in images:
+            crop_images = _split_image(image, split_image, split_ratio, max_size)
+            num_crops.append(torch.tensor(len(crop_images)))
+            for crop_image in crop_images:
+                img_padded, pixel_mask = keep_ratio_resize_and_pixel_mask(
+                    crop_image, max_size, min_size
+                )
+                img_padded = self.transform(img_padded)
+                pixel_values.append(img_padded)
+                pixel_masks.append(pixel_mask)
+
+        return BatchFeature(
+            data={
+                "pixel_values": torch.stack(pixel_values),
+                "pixel_mask": torch.stack(pixel_masks),
+                "num_crops": torch.stack(num_crops),
+            },
+            tensor_type=return_tensors,
+        )
+
+    def preprocess(
+        self,
+        images,
+        max_image_size=None,
+        min_image_size=None,
+        return_tensors: Optional[Union[str, TensorType]] = None,
+        split_image: Optional[bool] = False,
+        split_ratio: Optional[List[List[int]]] = [
+            [1, 2],
+            [1, 3],
+            [1, 4],
+            [1, 5],
+            [1, 6],
+            [1, 7],
+            [1, 8],
+            [2, 4],
+            [2, 3],
+            [2, 2],
+            [2, 1],
+            [3, 1],
+            [3, 2],
+            [4, 1],
+            [4, 2],
+            [5, 1],
+            [6, 1],
+            [7, 1],
+            [8, 1],
+        ],
+    ):
+        return self.__call__(
+            images,
+            max_image_size=max_image_size,
+            min_image_size=min_image_size,
+            return_tensors=return_tensors,
+            split_image=split_image,
+            split_ratio=split_ratio,
+        )