dc/model/multimodal_projector/builder.py

import torch
import torch.nn as nn
import re


class IdentityMap(nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, x, *args, **kwargs):
        return x

    @property
    def config(self):
        return {"mm_projector_type": 'identity'}


class SimpleResBlock(nn.Module):
    def __init__(self, channels):
        super().__init__()
        self.pre_norm = nn.LayerNorm(channels)

        self.proj = nn.Sequential(
            nn.Linear(channels, channels),
            nn.GELU(),
            nn.Linear(channels, channels)
        )
    def forward(self, x):
        x = self.pre_norm(x)
        return x + self.proj(x)

def build_vision_projector(config, delay_load=False, **kwargs):
    projector_type = getattr(config, 'mm_projector_type', 'linear')

    if projector_type == 'linear':
        return nn.Linear(config.mm_hidden_size, config.hidden_size)

    mlp_gelu_match = re.match(r'^mlp(\d+)x_gelu$', projector_type)
    if mlp_gelu_match:
        mlp_depth = int(mlp_gelu_match.group(1))
        modules = [nn.Linear(config.mm_hidden_size, config.hidden_size)]
        # modules = []
        for _ in range(1, mlp_depth):
            modules.append(nn.GELU())
            modules.append(nn.Linear(config.hidden_size, config.hidden_size))
        return nn.Sequential(*modules)

    if projector_type == 'identity':
        return IdentityMap()

    raise ValueError(f'Unknown projector type: {projector_type}')

class DenseConnector(nn.Module):
    def __init__(self, config, mlp_gelu_match):
        super().__init__()    

        self.dense_connector_type = dense_connector_type
        if self.dense_connector_type == 'token_cat':   # else channel cat
            self.avg_pooling_k8 = nn.AvgPool1d(kernel_size=8, stride=8)
        elif self.dense_connector_type == 'channel_cat':
            pass
        elif self.dense_connector_type == 'channel_sum_cat':
            self.dc_ln = nn.LayerNorm(mm_hidden_size)
            self.dc_linear = nn.Linear(mm_hidden_size, mm_hidden_size),
        else:
            raise ValueError(f'Unknown dense connector type: {dense_connector_type}')
            

    def forward(self, image_forward_outs, selected_features):
        return x

def build_dense_connector(config, delay_load=False, **kwargs):
    projector_type = getattr(config, 'mm_projector_type', 'linear')

    if projector_type == 'linear':
        return nn.Linear(config.mm_hidden_size, config.hidden_size)

    mlp_gelu_match = re.match(r'^mlp(\d+)x_gelu$', projector_type)
    if mlp_gelu_match:
        DenseConnector(config, mlp_gelu_match)

        mlp_depth = int(mlp_gelu_match.group(1))
        modules = [nn.Linear(config.mm_hidden_size, config.hidden_size)]
        # modules = []
        for _ in range(1, mlp_depth):
            modules.append(nn.GELU())
            modules.append(nn.Linear(config.hidden_size, config.hidden_size))
        return nn.Sequential(*modules)

    if projector_type == 'identity':
        return IdentityMap()

    raise ValueError(f'Unknown projector type: {projector_type}')