configuration_qwen2_mm.py

# coding=utf-8
# Copyright 2024 Microsoft Research & University of Wisconsin-Madison and the HuggingFace Inc. team. All rights reserved.
# Licensed 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.
"""Qwen2Audio model configuration"""

from transformers.configuration_utils import PretrainedConfig
from transformers.utils import logging
from transformers import CONFIG_MAPPING

import os
from typing import Union

logger = logging.get_logger(__name__)


class Qwen2SeamlessEncoderConfig(PretrainedConfig):

    model_type = "qwen2_seamless_encoder"

    def __init__(
        self,
        speech_encoder_layers=24,
        speech_encoder_attention_heads=16,
        speech_encoder_intermediate_size=4096,
        speech_encoder_hidden_act="swish",
        speech_encoder_dropout=0.0,
        add_adapter=True,
        speech_encoder_layerdrop=0.1,
        feature_projection_input_dim=160,
        adaptor_kernel_size=8,
        adaptor_stride=8,
        adaptor_dropout=0.1,
        num_adapter_layers=1,
        position_embeddings_type="relative_key",
        conv_depthwise_kernel_size=31,
        left_max_position_embeddings=64,
        right_max_position_embeddings=8,
        speech_encoder_chunk_size=20000,
        speech_encoder_left_chunk_num=128,
        **kwargs,
    ):
        super().__init__(**kwargs)

        self.speech_encoder_layers = speech_encoder_layers
        self.speech_encoder_hidden_act = speech_encoder_hidden_act
        self.speech_encoder_dropout = speech_encoder_dropout
        self.speech_encoder_attention_heads = speech_encoder_attention_heads
        self.speech_encoder_layerdrop = speech_encoder_layerdrop
        self.speech_encoder_intermediate_size = speech_encoder_intermediate_size
        self.feature_projection_input_dim = feature_projection_input_dim
        self.adaptor_kernel_size = adaptor_kernel_size
        self.adaptor_stride = adaptor_stride
        self.adaptor_dropout = adaptor_dropout
        self.num_adapter_layers = num_adapter_layers
        self.position_embeddings_type = position_embeddings_type
        self.conv_depthwise_kernel_size = conv_depthwise_kernel_size
        self.add_adapter = add_adapter
        self.left_max_position_embeddings = left_max_position_embeddings
        self.right_max_position_embeddings = right_max_position_embeddings
        self.speech_encoder_chunk_size = speech_encoder_chunk_size
        self.speech_encoder_left_chunk_num = speech_encoder_left_chunk_num
        self.audio_path = "/mnt/diskhd/Backup/DownloadModel/seamless-m4t-v2-large/"



class Qwen2VLVisionConfig(PretrainedConfig):
    model_type = "qwen2_vl"

    def __init__(
        self,
        depth=32,
        embed_dim=1280,
        hidden_size=3584,
        hidden_act="quick_gelu",
        mlp_ratio=4,
        num_heads=16,
        in_channels=3,
        patch_size=14,
        spatial_merge_size=2,
        temporal_patch_size=2,
        **kwargs,
    ):
        super().__init__(**kwargs)

        self.depth = depth
        self.embed_dim = embed_dim
        self.hidden_size = hidden_size
        self.hidden_act = hidden_act
        self.mlp_ratio = mlp_ratio
        self.num_heads = num_heads
        self.in_channels = in_channels
        self.patch_size = patch_size
        self.spatial_merge_size = spatial_merge_size
        self.temporal_patch_size = temporal_patch_size

    @classmethod
    def from_pretrained(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs) -> "PretrainedConfig":
        cls._set_token_in_kwargs(kwargs)

        config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)

        if 1:#config_dict.get("model_type") == "qwen2_vl":
            config_dict = config_dict["vision_config"]

        if "model_type" in config_dict and hasattr(cls, "model_type") and config_dict["model_type"] != cls.model_type:
            logger.warning(
                f"You are using a model of type {config_dict['model_type']} to instantiate a model of type "
                f"{cls.model_type}. This is not supported for all configurations of models and can yield errors."
            )

        return cls.from_dict(config_dict, **kwargs)
        
        

class Qwen2MMConfig(PretrainedConfig):

    model_type = "qwen2_mm"
    is_composition = False

    def __init__(
        self,
        vocab_size=152064,
        hidden_size=8192,
        intermediate_size=29568,
        num_hidden_layers=80,
        num_attention_heads=64,
        num_key_value_heads=8,
        hidden_act="silu",
        max_position_embeddings=32768,
        initializer_range=0.02,
        rms_norm_eps=1e-05,
        use_cache=True,
        tie_word_embeddings=False,
        rope_theta=1000000.0,
        use_sliding_window=False,
        sliding_window=4096,
        max_window_layers=80,
        attention_dropout=0.0,
        audio_config=None,
        vision_config=None,
        rope_scaling=None,
        **kwargs,
    ):
        if isinstance(vision_config, dict):
            self.vision_config = Qwen2VLVisionConfig(**vision_config)
        elif vision_config is None:
            self.vision_config = Qwen2VLVisionConfig()
            
        if isinstance(audio_config, dict):
            self.audio_config = Qwen2SeamlessEncoderConfig(**audio_config)
        elif audio_config is None:
            self.audio_config = Qwen2SeamlessEncoderConfig()

        self.vocab_size = vocab_size
        self.max_position_embeddings = max_position_embeddings
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.use_sliding_window = use_sliding_window
        self.sliding_window = sliding_window
        self.max_window_layers = max_window_layers

        # for backward compatibility
        if num_key_value_heads is None:
            num_key_value_heads = num_attention_heads

        self.num_key_value_heads = num_key_value_heads
        self.hidden_act = hidden_act
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.rope_theta = rope_theta
        self.attention_dropout = attention_dropout
        self.llm_path = "/mnt/diskhd/Backup/DownloadModel/Qwen2.5-3B-Instruct/"
        self.auto_map = {
            "AutoConfig": "configuration_qwen2_seamless.Qwen2MMConfig",
            "AutoModel": "modeling_qwen2_seamless.Qwen2SeamlessForConditionalGeneration"
        }
        self.rope_scaling = rope_scaling

        super().__init__(**kwargs)