Skip to content

SUPRA-Tex: 3D Generation with Super-Aligned Texture

5.0 KB
Raw Blame
import torch, warnings
from transformers import Siglip2VisionModel
try:
    from transformers.models.siglip.modeling_siglip import SiglipVisionModel
except:
    warnings.warn(f"Cannot import `SiglipVisionModel`. `Siglip2ImageEncoder` is not available. Please update `transformers` by `pip install -U transformers`.")
    SiglipVisionModel = torch.nn.Module
from diffsynth.core.device.npu_compatible_device import get_device_type


class Siglip2ImageEncoder(SiglipVisionModel):
    def __init__(self):
        from transformers.models.siglip.modeling_siglip import SiglipVisionConfig
        from transformers import SiglipImageProcessor
        config = SiglipVisionConfig(
            attention_dropout = 0.0,
            dtype = "float32",
            hidden_act = "gelu_pytorch_tanh",
            hidden_size = 1536,
            image_size = 384,
            intermediate_size = 6144,
            layer_norm_eps = 1e-06,
            model_type = "siglip_vision_model",
            num_attention_heads = 16,
            num_channels = 3,
            num_hidden_layers = 40,
            patch_size = 16,
            transformers_version = "4.56.1",
            _attn_implementation = "sdpa"
        )
        super().__init__(config)
        self.processor = SiglipImageProcessor(
            do_convert_rgb = None,
            do_normalize = True,
            do_rescale = True,
            do_resize = True,
            image_mean = [
                0.5,
                0.5,
                0.5
            ],
            image_processor_type = "SiglipImageProcessor",
            image_std = [
                0.5,
                0.5,
                0.5
            ],
            processor_class = "SiglipProcessor",
            resample = 2,
            rescale_factor = 0.00392156862745098,
            size = {
                "height": 384,
                "width": 384
            }
        )
        
    def forward(self, image, torch_dtype=torch.bfloat16, device=get_device_type()):
        pixel_values = self.processor(images=[image], return_tensors="pt")["pixel_values"]
        pixel_values = pixel_values.to(device=device, dtype=torch_dtype)
        output_attentions = False
        output_hidden_states = False
        interpolate_pos_encoding = False

        hidden_states = self.embeddings(pixel_values, interpolate_pos_encoding=interpolate_pos_encoding)

        encoder_outputs = self.encoder(
            inputs_embeds=hidden_states,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
        )

        last_hidden_state = encoder_outputs.last_hidden_state
        last_hidden_state = self.post_layernorm(last_hidden_state)

        pooler_output = self.head(last_hidden_state) if self.use_head else None

        return pooler_output


class Siglip2ImageEncoder428M(Siglip2VisionModel):
    def __init__(self):
        from transformers import Siglip2VisionConfig, Siglip2ImageProcessor
        config = Siglip2VisionConfig(
            attention_dropout = 0.0,
            dtype = "bfloat16",
            hidden_act = "gelu_pytorch_tanh",
            hidden_size = 1152,
            intermediate_size = 4304,
            layer_norm_eps = 1e-06,
            model_type = "siglip2_vision_model",
            num_attention_heads = 16,
            num_channels = 3,
            num_hidden_layers = 27,
            num_patches = 256,
            patch_size = 16,
            transformers_version = "4.57.1"
        )
        super().__init__(config)
        self.processor = Siglip2ImageProcessor(
            **{
                "data_format": "channels_first",
                "default_to_square": True,
                "device": None,
                "disable_grouping": None,
                "do_convert_rgb": None,
                "do_normalize": True,
                "do_pad": None,
                "do_rescale": True,
                "do_resize": True,
                "image_mean": [
                    0.5,
                    0.5,
                    0.5
                ],
                "image_processor_type": "Siglip2ImageProcessor",
                "image_std": [
                    0.5,
                    0.5,
                    0.5
                ],
                "input_data_format": None,
                "max_num_patches": 256,
                "pad_size": None,
                "patch_size": 16,
                "processor_class": "Siglip2Processor",
                "resample": 2,
                "rescale_factor": 0.00392156862745098,
                "return_tensors": None,
            }
        )
        
    def forward(self, image, torch_dtype=torch.bfloat16, device="cuda"):
        siglip_inputs = self.processor(images=[image], return_tensors="pt").to(device)
        shape = siglip_inputs.spatial_shapes[0]
        hidden_state = super().forward(**siglip_inputs).last_hidden_state
        B, N, C = hidden_state.shape
        hidden_state = hidden_state[:, : shape[0] * shape[1]]
        hidden_state = hidden_state.view(shape[0], shape[1], C)
        hidden_state = hidden_state.to(torch_dtype)
        return hidden_state