Upload folder using huggingface_hub

.gitattributes

@@ -36,3 +36,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 DeCo-XL-16-512/decoder/diffusion_pytorch_model.safetensors.bak filter=lfs diff=lfs merge=lfs -text
 DeCo-XL-16-512/demo.png filter=lfs diff=lfs merge=lfs -text
 DeCo-XL-16-512/transformer/diffusion_pytorch_model.safetensors.bak filter=lfs diff=lfs merge=lfs -text
+DeCo-XXL-16-512-t2i/demo.png filter=lfs diff=lfs merge=lfs -text
+DeCo-XXL-16-512-t2i/text_encoder/tokenizer.json filter=lfs diff=lfs merge=lfs -text

DeCo-XL-16-256/pipeline.py

@@ -1,63 +1,34 @@
-"""Hub custom pipeline: DeCoPipeline (class-conditioned c2i).
+"""Hub custom pipeline: DeCoPipeline.
 Load with native Hugging Face diffusers and trust_remote_code=True.
 """
 
 from __future__ import annotations
 
-import json
-from pathlib import Path
-from typing import Dict, List, Optional, Tuple, Union
+import inspect
 
-import torch
 from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
 from diffusers.utils.torch_utils import randn_tensor
+import json
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple, Union, Any
 
-EXAMPLE_DOC_STRING = """
-    Examples:
-        ```py
-        >>> from pathlib import Path
-        >>> from diffusers import DiffusionPipeline
-        >>> import torch
-
-        >>> model_dir = Path("./DeCo-XL-16-512").resolve()
-        >>> pipe = DiffusionPipeline.from_pretrained(
-        ...     str(model_dir),
-        ...     local_files_only=True,
-        ...     custom_pipeline=str(model_dir / "pipeline.py"),
-        ...     trust_remote_code=True,
-        ...     torch_dtype=torch.bfloat16,
-        ... )
-        >>> pipe.to("cuda")
-
-        >>> print(pipe.id2label[207])
-        >>> print(pipe.get_label_ids("golden retriever"))
-
-        >>> generator = torch.Generator(device="cuda").manual_seed(42)
-        >>> image = pipe(
-        ...     class_labels="golden retriever",
-        ...     num_inference_steps=100,
-        ...     guidance_scale=5.0,
-        ...     generator=generator,
-        ... ).images[0]
-        ```
-"""
-
+import torch
 
 class DeCoPipeline(DiffusionPipeline):
-    r"""
-    Pipeline for class-conditional image generation with DeCo.
-
-    Parameters:
-        transformer ([`DeCoTransformer2DModel`]):
-            Class-conditional DeCo transformer.
-        scheduler ([`DeCoFlowMatchEulerDiscreteScheduler`]):
-            Flow-matching Euler scheduler for DeCo.
-        decoder ([`DeCoPatchDecoderModel`]):
-            Per-patch RGB decoder (NerfEmbedder + AdaLN MLP).
-        id2label (`dict[int, str]`, *optional*):
-            ImageNet class id to English label mapping. Values may contain comma-separated synonyms.
-    """
 
+    @staticmethod
+    def prepare_extra_step_kwargs(
+        scheduler,
+        generator=None,
+        eta: float | None = None,
+    ):
+        kwargs = {}
+        step_params = set(inspect.signature(scheduler.step).parameters.keys())
+        if "generator" in step_params:
+            kwargs["generator"] = generator
+        if eta is not None and "eta" in step_params:
+            kwargs["eta"] = eta
+        return kwargs
     model_cpu_offload_seq = "transformer->decoder"
 
     def __init__(
@@ -112,20 +83,27 @@ class DeCoPipeline(DiffusionPipeline):
                     label2id[synonym] = int(class_id)
         return dict(sorted(label2id.items()))
 
+    @staticmethod
+    def _effective_guidance_scale(
+        timestep: Union[torch.Tensor, float],
+        guidance_scale: float,
+        do_cfg: bool,
+        guidance_interval_min: float,
+        guidance_interval_max: float,
+    ) -> float:
+        if not do_cfg:
+            return 1.0
+        t = float(timestep)
+        if t > guidance_interval_min and t <= guidance_interval_max:
+            return float(guidance_scale)
+        return 1.0
+
     @property
     def id2label(self) -> Dict[int, str]:
-        r"""ImageNet class id to English label string (comma-separated synonyms)."""
         self._ensure_labels_loaded()
         return self._id2label
 
     def get_label_ids(self, label: Union[str, List[str]]) -> List[int]:
-        r"""
-        Map ImageNet label strings to class ids.
-
-        Args:
-            label (`str` or `list[str]`):
-                One or more English label strings. Each string must match a synonym in `id2label`.
-        """
         self._ensure_labels_loaded()
         label2id = self.labels
         if not label2id:
@@ -170,34 +148,12 @@ class DeCoPipeline(DiffusionPipeline):
         width: Optional[int] = None,
         num_inference_steps: int = 50,
         guidance_scale: float = 1.0,
+        guidance_interval_min: float = 0.1,
+        guidance_interval_max: float = 1.0,
         generator: Optional[Union[torch.Generator, list[torch.Generator]]] = None,
         output_type: str = "pil",
         return_dict: bool = True,
     ) -> Union[ImagePipelineOutput, Tuple]:
-        r"""
-        Generate class-conditional images with DeCo.
-
-        Args:
-            class_labels (`int`, `str`, `list[int]`, `list[str]`, or `torch.LongTensor`):
-                ImageNet class indices or human-readable English label strings.
-            batch_size (`int`, *optional*):
-                Number of images to generate. Defaults to the number of class labels. When a single
-                class label is provided, repeats it to match `batch_size`.
-            height (`int`, *optional*):
-                Output image height in pixels. Defaults to `transformer.config.sample_size`.
-            width (`int`, *optional*):
-                Output image width in pixels. Defaults to `transformer.config.sample_size`.
-            num_inference_steps (`int`, defaults to `50`):
-                Number of denoising steps.
-            guidance_scale (`float`, defaults to `1.0`):
-                Classifier-free guidance scale. CFG is active when `guidance_scale > 1.0`.
-            generator (`torch.Generator`, *optional*):
-                RNG for reproducibility.
-            output_type (`str`, defaults to `"pil"`):
-                `"pil"`, `"np"`, or `"latent"`.
-            return_dict (`bool`, defaults to `True`):
-                Return [`ImagePipelineOutput`] if True.
-        """
         device = self._execution_device
         dtype = next(self.transformer.parameters()).dtype
         do_cfg = guidance_scale is not None and float(guidance_scale) > 1.0
@@ -230,6 +186,13 @@ class DeCoPipeline(DiffusionPipeline):
 
         for timestep in self.progress_bar(timesteps):
             latent_model_input = self.scheduler.scale_model_input(latents, timestep)
+            effective_guidance = self._effective_guidance_scale(
+                timestep,
+                guidance_scale,
+                do_cfg,
+                guidance_interval_min,
+                guidance_interval_max,
+            )
 
             if do_cfg:
                 latent_model_input = torch.cat([latent_model_input, latent_model_input], dim=0)
@@ -240,13 +203,15 @@ class DeCoPipeline(DiffusionPipeline):
                     decoder=self.decoder,
                 ).sample
                 model_output_uncond, model_output_cond = model_output.chunk(2)
-                model_output = model_output_uncond + float(guidance_scale) * (model_output_cond - model_output_uncond)
+                model_output = model_output_uncond + effective_guidance * (
+                    model_output_cond - model_output_uncond
+                )
             else:
                 model_output = self.transformer(
                     latent_model_input, timestep, class_labels=class_labels, decoder=self.decoder
                 ).sample
 
-            latents = self.scheduler.step(model_output, timestep, latents).prev_sample
+            latents = self.scheduler.step(model_output, timestep, latents, **extra_step_kwargs).prev_sample
 
         image = latents
 
@@ -265,4 +230,4 @@ class DeCoPipeline(DiffusionPipeline):
 
         if not return_dict:
             return (image,)
-        return ImagePipelineOutput(images=image)
+        return ImagePipelineOutput(images=image)

DeCo-XL-16-512/pipeline.py

@@ -1,63 +1,34 @@
-"""Hub custom pipeline: DeCoPipeline (class-conditioned c2i).
+"""Hub custom pipeline: DeCoPipeline.
 Load with native Hugging Face diffusers and trust_remote_code=True.
 """
 
 from __future__ import annotations
 
-import json
-from pathlib import Path
-from typing import Dict, List, Optional, Tuple, Union
+import inspect
 
-import torch
 from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
 from diffusers.utils.torch_utils import randn_tensor
+import json
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple, Union, Any
 
-EXAMPLE_DOC_STRING = """
-    Examples:
-        ```py
-        >>> from pathlib import Path
-        >>> from diffusers import DiffusionPipeline
-        >>> import torch
-
-        >>> model_dir = Path("./DeCo-XL-16-512").resolve()
-        >>> pipe = DiffusionPipeline.from_pretrained(
-        ...     str(model_dir),
-        ...     local_files_only=True,
-        ...     custom_pipeline=str(model_dir / "pipeline.py"),
-        ...     trust_remote_code=True,
-        ...     torch_dtype=torch.bfloat16,
-        ... )
-        >>> pipe.to("cuda")
-
-        >>> print(pipe.id2label[207])
-        >>> print(pipe.get_label_ids("golden retriever"))
-
-        >>> generator = torch.Generator(device="cuda").manual_seed(42)
-        >>> image = pipe(
-        ...     class_labels="golden retriever",
-        ...     num_inference_steps=100,
-        ...     guidance_scale=5.0,
-        ...     generator=generator,
-        ... ).images[0]
-        ```
-"""
-
+import torch
 
 class DeCoPipeline(DiffusionPipeline):
-    r"""
-    Pipeline for class-conditional image generation with DeCo.
-
-    Parameters:
-        transformer ([`DeCoTransformer2DModel`]):
-            Class-conditional DeCo transformer.
-        scheduler ([`DeCoFlowMatchEulerDiscreteScheduler`]):
-            Flow-matching Euler scheduler for DeCo.
-        decoder ([`DeCoPatchDecoderModel`]):
-            Per-patch RGB decoder (NerfEmbedder + AdaLN MLP).
-        id2label (`dict[int, str]`, *optional*):
-            ImageNet class id to English label mapping. Values may contain comma-separated synonyms.
-    """
 
+    @staticmethod
+    def prepare_extra_step_kwargs(
+        scheduler,
+        generator=None,
+        eta: float | None = None,
+    ):
+        kwargs = {}
+        step_params = set(inspect.signature(scheduler.step).parameters.keys())
+        if "generator" in step_params:
+            kwargs["generator"] = generator
+        if eta is not None and "eta" in step_params:
+            kwargs["eta"] = eta
+        return kwargs
     model_cpu_offload_seq = "transformer->decoder"
 
     def __init__(
@@ -112,20 +83,27 @@ class DeCoPipeline(DiffusionPipeline):
                     label2id[synonym] = int(class_id)
         return dict(sorted(label2id.items()))
 
+    @staticmethod
+    def _effective_guidance_scale(
+        timestep: Union[torch.Tensor, float],
+        guidance_scale: float,
+        do_cfg: bool,
+        guidance_interval_min: float,
+        guidance_interval_max: float,
+    ) -> float:
+        if not do_cfg:
+            return 1.0
+        t = float(timestep)
+        if t > guidance_interval_min and t <= guidance_interval_max:
+            return float(guidance_scale)
+        return 1.0
+
     @property
     def id2label(self) -> Dict[int, str]:
-        r"""ImageNet class id to English label string (comma-separated synonyms)."""
         self._ensure_labels_loaded()
         return self._id2label
 
     def get_label_ids(self, label: Union[str, List[str]]) -> List[int]:
-        r"""
-        Map ImageNet label strings to class ids.
-
-        Args:
-            label (`str` or `list[str]`):
-                One or more English label strings. Each string must match a synonym in `id2label`.
-        """
         self._ensure_labels_loaded()
         label2id = self.labels
         if not label2id:
@@ -170,34 +148,12 @@ class DeCoPipeline(DiffusionPipeline):
         width: Optional[int] = None,
         num_inference_steps: int = 50,
         guidance_scale: float = 1.0,
+        guidance_interval_min: float = 0.1,
+        guidance_interval_max: float = 1.0,
         generator: Optional[Union[torch.Generator, list[torch.Generator]]] = None,
         output_type: str = "pil",
         return_dict: bool = True,
     ) -> Union[ImagePipelineOutput, Tuple]:
-        r"""
-        Generate class-conditional images with DeCo.
-
-        Args:
-            class_labels (`int`, `str`, `list[int]`, `list[str]`, or `torch.LongTensor`):
-                ImageNet class indices or human-readable English label strings.
-            batch_size (`int`, *optional*):
-                Number of images to generate. Defaults to the number of class labels. When a single
-                class label is provided, repeats it to match `batch_size`.
-            height (`int`, *optional*):
-                Output image height in pixels. Defaults to `transformer.config.sample_size`.
-            width (`int`, *optional*):
-                Output image width in pixels. Defaults to `transformer.config.sample_size`.
-            num_inference_steps (`int`, defaults to `50`):
-                Number of denoising steps.
-            guidance_scale (`float`, defaults to `1.0`):
-                Classifier-free guidance scale. CFG is active when `guidance_scale > 1.0`.
-            generator (`torch.Generator`, *optional*):
-                RNG for reproducibility.
-            output_type (`str`, defaults to `"pil"`):
-                `"pil"`, `"np"`, or `"latent"`.
-            return_dict (`bool`, defaults to `True`):
-                Return [`ImagePipelineOutput`] if True.
-        """
         device = self._execution_device
         dtype = next(self.transformer.parameters()).dtype
         do_cfg = guidance_scale is not None and float(guidance_scale) > 1.0
@@ -230,6 +186,13 @@ class DeCoPipeline(DiffusionPipeline):
 
         for timestep in self.progress_bar(timesteps):
             latent_model_input = self.scheduler.scale_model_input(latents, timestep)
+            effective_guidance = self._effective_guidance_scale(
+                timestep,
+                guidance_scale,
+                do_cfg,
+                guidance_interval_min,
+                guidance_interval_max,
+            )
 
             if do_cfg:
                 latent_model_input = torch.cat([latent_model_input, latent_model_input], dim=0)
@@ -240,13 +203,15 @@ class DeCoPipeline(DiffusionPipeline):
                     decoder=self.decoder,
                 ).sample
                 model_output_uncond, model_output_cond = model_output.chunk(2)
-                model_output = model_output_uncond + float(guidance_scale) * (model_output_cond - model_output_uncond)
+                model_output = model_output_uncond + effective_guidance * (
+                    model_output_cond - model_output_uncond
+                )
             else:
                 model_output = self.transformer(
                     latent_model_input, timestep, class_labels=class_labels, decoder=self.decoder
                 ).sample
 
-            latents = self.scheduler.step(model_output, timestep, latents).prev_sample
+            latents = self.scheduler.step(model_output, timestep, latents, **extra_step_kwargs).prev_sample
 
         image = latents
 
@@ -265,4 +230,4 @@ class DeCoPipeline(DiffusionPipeline):
 
         if not return_dict:
             return (image,)
-        return ImagePipelineOutput(images=image)
+        return ImagePipelineOutput(images=image)

DeCo-XXL-16-512-t2i/decoder/decoder_deco.py

@@ -53,8 +53,8 @@ class NerfEmbedder(nn.Module):
     @lru_cache
     def fetch_pos(self, patch_size: int, device: torch.device, dtype: torch.dtype) -> torch.Tensor:
         pos = _precompute_freqs_cis_ex2d(self.max_freqs**2 * 2, patch_size, patch_size)
-        # Official code casts complex cis to real when concatenating with patch pixels.
-        return pos[None, :, :].to(device=device, dtype=dtype)
+        # Official DeCo uses real part of complex cis (same as .to(dtype) on complex tensors).
+        return pos[None, :, :].real.to(device=device, dtype=dtype)
 
     def forward(self, inputs: torch.Tensor) -> torch.Tensor:
         batch_size, patch_tokens, _ = inputs.shape

DeCo-XXL-16-512-t2i/pipeline.py

@@ -1,90 +1,87 @@
-"""Hub custom pipeline: DeCoT2IPipeline (text-to-image, 512×512).
-
-Sampling matches official DeCo AdamLMSampler:
-https://github.com/Zehong-Ma/DeCo/blob/main/src/diffusion/flow_matching/adam_sampling.py
+"""Hub custom pipeline: DeCoPipeline.
+Load with native Hugging Face diffusers and trust_remote_code=True.
 """
 
 from __future__ import annotations
 
-from pathlib import Path
-from typing import List, Optional, Tuple, Union
+import inspect
 
-import torch
 from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
 from diffusers.utils.torch_utils import randn_tensor
+import json
+from pathlib import Path
+from typing import Dict, List, Optional, Tuple, Union, Any
 
-DEFAULT_TEXT_ENCODER_REPO = "Qwen/Qwen3-1.7B"
+import torch
 
+class DeCoPipeline(DiffusionPipeline):
 
-class DeCoT2IPipeline(DiffusionPipeline):
-    model_cpu_offload_seq = "text_encoder->transformer->decoder"
-    _optional_components = ["text_encoder", "tokenizer"]
+    @staticmethod
+    def prepare_extra_step_kwargs(
+        scheduler,
+        generator=None,
+        eta: float | None = None,
+    ):
+        kwargs = {}
+        step_params = set(inspect.signature(scheduler.step).parameters.keys())
+        if "generator" in step_params:
+            kwargs["generator"] = generator
+        if eta is not None and "eta" in step_params:
+            kwargs["eta"] = eta
+        return kwargs
+    model_cpu_offload_seq = "transformer->decoder"
 
     def __init__(
         self,
         transformer,
         scheduler,
         decoder,
-        text_encoder=None,
-        tokenizer=None,
+        id2label: Optional[Dict[Union[int, str], str]] = None,
     ):
         super().__init__()
-        self.register_modules(
-            transformer=transformer,
-            scheduler=scheduler,
-            decoder=decoder,
-            text_encoder=text_encoder,
-            tokenizer=tokenizer,
-        )
-
-    @classmethod
-    def from_pretrained(cls, pretrained_model_name_or_path, *args, **kwargs):
-        pipe = super().from_pretrained(pretrained_model_name_or_path, *args, **kwargs)
-        if pipe.text_encoder is None or pipe.tokenizer is None:
-            model_dir = Path(getattr(pipe.config, "_name_or_path", pretrained_model_name_or_path)).resolve()
-            pipe._load_text_encoder(model_dir, **kwargs)
-        return pipe
+        self.register_modules(transformer=transformer, scheduler=scheduler, decoder=decoder)
+        self._id2label = self._normalize_id2label(id2label)
+        self.labels = self._build_label2id(self._id2label)
+        self._labels_loaded_from_model_index = bool(self._id2label)
+
+    def _ensure_labels_loaded(self) -> None:
+        if self._labels_loaded_from_model_index:
+            return
+        loaded = self._read_id2label_from_model_index(getattr(self.config, "_name_or_path", None))
+        if loaded:
+            self._id2label = loaded
+            self.labels = self._build_label2id(self._id2label)
+        self._labels_loaded_from_model_index = True
 
     @staticmethod
-    def _resolve_text_encoder_path(model_dir: Path) -> Path:
-        hint = model_dir / "text_encoder_pretrained_model_name_or_path.txt"
-        if hint.exists():
-            raw = hint.read_text(encoding="utf-8").strip().splitlines()[0].strip()
-            path = Path(raw)
-            if not path.is_absolute():
-                path = (model_dir / path).resolve()
-            if path.exists():
-                return path
-        local = model_dir / "text_encoder"
-        if local.exists():
-            return local.resolve()
-        return Path(DEFAULT_TEXT_ENCODER_REPO)
-
-    def _load_text_encoder(self, model_dir: Path, **kwargs) -> None:
-        from transformers import Qwen2Tokenizer, Qwen3Model
-
-        text_path = self._resolve_text_encoder_path(model_dir)
-        load_kwargs = {
-            k: kwargs[k]
-            for k in ("torch_dtype", "device_map", "local_files_only", "revision", "cache_dir")
-            if k in kwargs
-        }
-        text_encoder = Qwen3Model.from_pretrained(str(text_path), **load_kwargs)
-        tokenizer = Qwen2Tokenizer.from_pretrained(
-            str(text_path),
-            max_length=self.txt_max_length,
-            padding_side="right",
-            **{k: v for k, v in load_kwargs.items() if k in ("local_files_only", "revision", "cache_dir")},
-        )
-        self.register_modules(text_encoder=text_encoder, tokenizer=tokenizer)
+    def _normalize_id2label(id2label: Optional[Dict[Union[int, str], str]]) -> Dict[int, str]:
+        if not id2label:
+            return {}
+        return {int(key): value for key, value in id2label.items()}
 
-    @property
-    def txt_embed_dim(self) -> int:
-        return int(getattr(self.transformer.config, "txt_embed_dim", 2048))
+    @staticmethod
+    def _read_id2label_from_model_index(variant_path: Optional[str]) -> Dict[int, str]:
+        if not variant_path:
+            return {}
+        variant_dir = Path(variant_path).resolve()
+        model_index_path = variant_dir / "model_index.json"
+        if not model_index_path.exists():
+            return {}
+        raw = json.loads(model_index_path.read_text(encoding="utf-8"))
+        id2label = raw.get("id2label")
+        if not isinstance(id2label, dict):
+            return {}
+        return {int(key): value for key, value in id2label.items()}
 
-    @property
-    def txt_max_length(self) -> int:
-        return int(getattr(self.transformer.config, "txt_max_length", 128))
+    @staticmethod
+    def _build_label2id(id2label: Dict[int, str]) -> Dict[str, int]:
+        label2id: Dict[str, int] = {}
+        for class_id, value in id2label.items():
+            for synonym in value.split(","):
+                synonym = synonym.strip()
+                if synonym:
+                    label2id[synonym] = int(class_id)
+        return dict(sorted(label2id.items()))
 
     @staticmethod
     def _effective_guidance_scale(
@@ -94,87 +91,65 @@ class DeCoT2IPipeline(DiffusionPipeline):
         guidance_interval_min: float,
         guidance_interval_max: float,
     ) -> float:
-        """Match official AdamLMSampler: CFG when t > min and t < max."""
         if not do_cfg:
             return 1.0
         t = float(timestep)
-        if t > guidance_interval_min and t < guidance_interval_max:
+        if t > guidance_interval_min and t <= guidance_interval_max:
             return float(guidance_scale)
         return 1.0
 
-    @staticmethod
-    def _fp_to_uint8(image: torch.Tensor) -> torch.Tensor:
-        return torch.clip_((image + 1) * 127.5 + 0.5, 0, 255).to(torch.uint8)
-
-    def encode_prompt(
+    @property
+    def id2label(self) -> Dict[int, str]:
+        self._ensure_labels_loaded()
+        return self._id2label
+
+    def get_label_ids(self, label: Union[str, List[str]]) -> List[int]:
+        self._ensure_labels_loaded()
+        label2id = self.labels
+        if not label2id:
+            raise ValueError("No English labels loaded. Ensure `id2label` exists in model_index.json.")
+
+        if isinstance(label, str):
+            label = [label]
+
+        missing = [item for item in label if item not in label2id]
+        if missing:
+            preview = ", ".join(list(label2id.keys())[:8])
+            raise ValueError(f"Unknown English label(s): {missing}. Example valid labels: {preview}, ...")
+        return [label2id[item] for item in label]
+
+    def _normalize_class_labels(
         self,
-        prompt: Union[str, List[str]],
-        negative_prompt: Optional[Union[str, List[str]]] = None,
-        device: Optional[torch.device] = None,
-        dtype: Optional[torch.dtype] = None,
-    ) -> Tuple[torch.Tensor, torch.Tensor]:
-        if self.text_encoder is None or self.tokenizer is None:
-            raise ValueError("text_encoder and tokenizer must be loaded for t2i inference.")
-
-        device = device or self._execution_device
-        dtype = dtype or next(self.text_encoder.parameters()).dtype
-
-        if isinstance(prompt, str):
-            prompt = [prompt]
-        batch_size = len(prompt)
-
-        if negative_prompt is None:
-            negative_prompt = [""] * batch_size
-        elif isinstance(negative_prompt, str):
-            negative_prompt = [negative_prompt] * batch_size
-
-        def _encode(texts: List[str]) -> torch.Tensor:
-            tokenized = self.tokenizer(
-                texts,
-                truncation=True,
-                max_length=self.txt_max_length,
-                padding="max_length",
-                return_tensors="pt",
-            )
-            input_ids = tokenized.input_ids.to(device)
-            attention_mask = tokenized.attention_mask.to(device)
-            outputs = self.text_encoder(input_ids=input_ids, attention_mask=attention_mask)
-            hidden = outputs[0]
-            embed_dim = self.txt_embed_dim
-            if hidden.shape[-1] < embed_dim:
-                pad = torch.zeros(
-                    hidden.shape[0],
-                    hidden.shape[1],
-                    embed_dim - hidden.shape[-1],
-                    device=hidden.device,
-                    dtype=hidden.dtype,
-                )
-                hidden = torch.cat([hidden, pad], dim=-1)
-            elif hidden.shape[-1] > embed_dim:
-                hidden = hidden[:, :, :embed_dim]
-            return hidden.to(dtype=dtype)
+        class_labels: Union[int, str, List[Union[int, str]], torch.LongTensor],
+    ) -> torch.LongTensor:
+        if torch.is_tensor(class_labels):
+            return class_labels.to(device=self._execution_device, dtype=torch.long).reshape(-1)
+
+        if isinstance(class_labels, int):
+            class_label_ids = [class_labels]
+        elif isinstance(class_labels, str):
+            class_label_ids = self.get_label_ids(class_labels)
+        elif class_labels and isinstance(class_labels[0], str):
+            class_label_ids = self.get_label_ids(class_labels)
+        else:
+            class_label_ids = list(class_labels)
 
-        return _encode(prompt), _encode(negative_prompt)
+        return torch.tensor(class_label_ids, device=self._execution_device, dtype=torch.long).reshape(-1)
 
     def _default_sample_size(self) -> int:
-        return int(getattr(self.transformer.config, "sample_size", 512))
+        return int(getattr(self.transformer.config, "sample_size", 256))
 
     @torch.no_grad()
-    @torch.autocast(device_type="cuda", dtype=torch.bfloat16, enabled=torch.cuda.is_available())
     def __call__(
         self,
-        prompt: Optional[Union[str, List[str]]] = None,
-        negative_prompt: Optional[Union[str, List[str]]] = None,
-        prompt_embeds: Optional[torch.Tensor] = None,
-        negative_prompt_embeds: Optional[torch.Tensor] = None,
+        class_labels: Union[int, str, List[Union[int, str]], torch.LongTensor],
+        batch_size: Optional[int] = None,
         height: Optional[int] = None,
         width: Optional[int] = None,
-        num_inference_steps: int = 25,
-        guidance_scale: float = 4.0,
-        timeshift: Optional[float] = None,
-        order: Optional[int] = None,
-        guidance_interval_min: Optional[float] = None,
-        guidance_interval_max: Optional[float] = None,
+        num_inference_steps: int = 50,
+        guidance_scale: float = 1.0,
+        guidance_interval_min: float = 0.1,
+        guidance_interval_max: float = 1.0,
         generator: Optional[Union[torch.Generator, list[torch.Generator]]] = None,
         output_type: str = "pil",
         return_dict: bool = True,
@@ -183,109 +158,76 @@ class DeCoT2IPipeline(DiffusionPipeline):
         dtype = next(self.transformer.parameters()).dtype
         do_cfg = guidance_scale is not None and float(guidance_scale) > 1.0
 
-        if prompt_embeds is not None:
-            batch_size = int(prompt_embeds.shape[0])
-        elif prompt is None:
-            raise ValueError("Either `prompt` or `prompt_embeds` must be provided.")
-        elif isinstance(prompt, str):
-            batch_size = 1
-        else:
-            batch_size = len(prompt)
-
         sample_size = self._default_sample_size()
         height = int(height if height is not None else sample_size)
         width = int(width if width is not None else sample_size)
-        height = height // 16 * 16
-        width = width // 16 * 16
 
-        interval_min = (
-            float(guidance_interval_min)
-            if guidance_interval_min is not None
-            else float(getattr(self.scheduler.config, "guidance_interval_min", 0.0))
-        )
-        interval_max = (
-            float(guidance_interval_max)
-            if guidance_interval_max is not None
-            else float(getattr(self.scheduler.config, "guidance_interval_max", 1.0))
+        class_labels = self._normalize_class_labels(class_labels)
+        if batch_size is None:
+            batch_size = int(class_labels.numel())
+        elif class_labels.numel() == 1 and batch_size > 1:
+            class_labels = class_labels.repeat(batch_size)
+        elif class_labels.numel() != batch_size:
+            raise ValueError("class_labels batch size must match batch_size")
+
+        if do_cfg:
+            null_label = int(self.transformer.config.num_classes)
+            uncond_labels = torch.full((batch_size,), null_label, device=device, dtype=torch.long)
+
+        latents = randn_tensor(
+            (batch_size, int(self.transformer.config.in_channels), height, width),
+            generator=generator,
+            device=device,
+            dtype=dtype,
         )
 
-        if prompt_embeds is None:
-            prompt_embeds, negative_embeds = self.encode_prompt(
-                prompt=prompt,
-                negative_prompt=negative_prompt,
-                device=device,
-                dtype=dtype,
+        self.scheduler.set_timesteps(num_inference_steps, device=device)
+        timesteps = self.scheduler.timesteps[:-1]
+
+        for timestep in self.progress_bar(timesteps):
+            latent_model_input = self.scheduler.scale_model_input(latents, timestep)
+            effective_guidance = self._effective_guidance_scale(
+                timestep,
+                guidance_scale,
+                do_cfg,
+                guidance_interval_min,
+                guidance_interval_max,
             )
-        else:
-            negative_embeds = negative_prompt_embeds
-            if negative_embeds is None:
-                negative_embeds = torch.zeros_like(prompt_embeds)
-
-        # Official DeCo t2i: float32 noise on CPU, then move to device (see app.py / GenEval).
-        noise_shape = (batch_size, int(self.transformer.config.in_channels), height, width)
-        if generator is not None:
-            gen_device = getattr(generator, "device", None)
-            if gen_device is not None and str(gen_device).startswith("cuda"):
-                latents = randn_tensor(
-                    noise_shape, generator=generator, device=device, dtype=torch.float32
+
+            if do_cfg:
+                latent_model_input = torch.cat([latent_model_input, latent_model_input], dim=0)
+                model_output = self.transformer(
+                    latent_model_input,
+                    timestep,
+                    class_labels=torch.cat([uncond_labels, class_labels], dim=0),
+                    decoder=self.decoder,
+                ).sample
+                model_output_uncond, model_output_cond = model_output.chunk(2)
+                model_output = model_output_uncond + effective_guidance * (
+                    model_output_cond - model_output_uncond
                 )
             else:
-                latents = randn_tensor(
-                    noise_shape, generator=generator, device="cpu", dtype=torch.float32
-                ).to(device)
-        else:
-            latents = randn_tensor(noise_shape, device="cpu", dtype=torch.float32).to(device)
-
-        set_kwargs = {
-            "num_inference_steps": num_inference_steps,
-            "guidance_scale": guidance_scale,
-            "device": device,
-        }
-        if timeshift is not None:
-            set_kwargs["timeshift"] = timeshift
-        if order is not None:
-            set_kwargs["order"] = order
-        self.scheduler.set_timesteps(**set_kwargs)
-
-        cfg_condition = torch.cat([negative_embeds, prompt_embeds], dim=0)
-        pred_trajectory: list[torch.Tensor] = []
-        t_cur = torch.zeros(batch_size, device=device, dtype=torch.float32)
-        timedeltas = self.scheduler._timedeltas
-        solver_coeffs = self.scheduler._solver_coeffs
-
-        for i in self.progress_bar(range(len(timedeltas))):
-            cfg_x = torch.cat([latents, latents], dim=0)
-            cfg_t = t_cur.repeat(2)
-            out = self.transformer(cfg_x, cfg_t, encoder_hidden_states=cfg_condition, decoder=self.decoder).sample
-
-            if do_cfg and t_cur[0] > interval_min and t_cur[0] < interval_max:
-                cfg_scale = float(guidance_scale)
-            else:
-                cfg_scale = 1.0
-            uncond, cond = out.chunk(2, dim=0)
-            out = uncond + cfg_scale * (cond - uncond)
-
-            pred_trajectory.append(out)
-            combined = torch.zeros_like(out)
-            order = len(solver_coeffs[i])
-            for j in range(order):
-                combined = combined + solver_coeffs[i][j] * pred_trajectory[-order:][j]
-            latents = latents + combined * timedeltas[i]
-            t_cur = t_cur + timedeltas[i]
+                model_output = self.transformer(
+                    latent_model_input, timestep, class_labels=class_labels, decoder=self.decoder
+                ).sample
+
+            latents = self.scheduler.step(model_output, timestep, latents, **extra_step_kwargs).prev_sample
+
+        image = latents
 
         if output_type == "latent":
             if not return_dict:
-                return (latents,)
-            return ImagePipelineOutput(images=latents)
+                return (image,)
+            return ImagePipelineOutput(images=image)
+
+        image = (image / 2 + 0.5).clamp(0, 1)
+        image = image.cpu().permute(0, 2, 3, 1).float().numpy()
 
-        images_uint8 = self._fp_to_uint8(latents.float()).permute(0, 2, 3, 1).cpu().numpy()
         if output_type == "pil":
-            image = self.numpy_to_pil(images_uint8)
-        elif output_type == "np":
-            image = images_uint8
-        else:
+            image = self.numpy_to_pil(image)
+        elif output_type != "np":
             raise ValueError("output_type must be one of {'pil', 'np', 'latent'}")
 
         if not return_dict:
             return (image,)
-        return ImagePipelineOutput(images=image)
+        return ImagePipelineOutput(images=image)

DeCo-XXL-16-512-t2i/scheduler/scheduling_deco_flow_match_adam_discrete.py

@@ -24,9 +24,9 @@ class DeCoFlowMatchAdamDiscreteScheduler(SchedulerMixin, ConfigMixin):
 
     @staticmethod
     def _lagrange_coeffs(order: int, pre_ts: torch.Tensor, t_start: torch.Tensor, t_end: torch.Tensor) -> List[float]:
-        ts = [float(v) for v in pre_ts[-order:].tolist()]
-        a = float(t_start)
-        b = float(t_end)
+        ts = [float(v) for v in pre_ts[-order:].to(dtype=torch.float64).tolist()]
+        a = float(t_start.to(dtype=torch.float64).item() if isinstance(t_start, torch.Tensor) else t_start)
+        b = float(t_end.to(dtype=torch.float64).item() if isinstance(t_end, torch.Tensor) else t_end)
 
         if order == 1:
             return [1.0]
@@ -55,6 +55,62 @@ class DeCoFlowMatchAdamDiscreteScheduler(SchedulerMixin, ConfigMixin):
             int3 = int3 / int3_denom
             total = int1 + int2 + int3
             return [int1 / total, int2 / total, int3 / total]
+        if order == 4:
+            t1, t2, t3, t4 = ts
+            int1_denom = (t1 - t2) * (t1 - t3) * (t1 - t4)
+            int1 = (
+                (1 / 4) * b**4
+                - (1 / 3) * (t2 + t3 + t4) * b**3
+                + 0.5 * (t3 * t4 + t2 * t3 + t2 * t4) * b**2
+                - (t2 * t3 * t4) * b
+            ) - (
+                (1 / 4) * a**4
+                - (1 / 3) * (t2 + t3 + t4) * a**3
+                + 0.5 * (t3 * t4 + t2 * t3 + t2 * t4) * a**2
+                - (t2 * t3 * t4) * a
+            )
+            int1 = int1 / int1_denom
+            int2_denom = (t2 - t1) * (t2 - t3) * (t2 - t4)
+            int2 = (
+                (1 / 4) * b**4
+                - (1 / 3) * (t1 + t3 + t4) * b**3
+                + 0.5 * (t3 * t4 + t1 * t3 + t1 * t4) * b**2
+                - (t1 * t3 * t4) * b
+            ) - (
+                (1 / 4) * a**4
+                - (1 / 3) * (t1 + t3 + t4) * a**3
+                + 0.5 * (t3 * t4 + t1 * t3 + t1 * t4) * a**2
+                - (t1 * t3 * t4) * a
+            )
+            int2 = int2 / int2_denom
+            int3_denom = (t3 - t1) * (t3 - t2) * (t3 - t4)
+            int3 = (
+                (1 / 4) * b**4
+                - (1 / 3) * (t1 + t2 + t4) * b**3
+                + 0.5 * (t4 * t2 + t1 * t2 + t1 * t4) * b**2
+                - (t1 * t2 * t4) * b
+            ) - (
+                (1 / 4) * a**4
+                - (1 / 3) * (t1 + t2 + t4) * a**3
+                + 0.5 * (t4 * t2 + t1 * t2 + t1 * t4) * a**2
+                - (t1 * t2 * t4) * a
+            )
+            int3 = int3 / int3_denom
+            int4_denom = (t4 - t1) * (t4 - t2) * (t4 - t3)
+            int4 = (
+                (1 / 4) * b**4
+                - (1 / 3) * (t1 + t2 + t3) * b**3
+                + 0.5 * (t3 * t2 + t1 * t2 + t1 * t3) * b**2
+                - (t1 * t2 * t3) * b
+            ) - (
+                (1 / 4) * a**4
+                - (1 / 3) * (t1 + t2 + t3) * a**3
+                + 0.5 * (t3 * t2 + t1 * t2 + t1 * t3) * a**2
+                - (t1 * t2 * t3) * a
+            )
+            int4 = int4 / int4_denom
+            total = int1 + int2 + int3 + int4
+            return [int1 / total, int2 / total, int3 / total, int4 / total]
         raise ValueError(f"Unsupported solver order: {order}.")
 
     @register_to_config
@@ -100,10 +156,10 @@ class DeCoFlowMatchAdamDiscreteScheduler(SchedulerMixin, ConfigMixin):
         if last_step is None:
             last_step = 1.0 / float(num_inference_steps)
 
-        endpoints = torch.linspace(0.0, 1.0 - float(last_step), int(num_inference_steps), dtype=torch.float32)
-        endpoints = torch.cat([endpoints, torch.tensor([1.0], dtype=torch.float32)], dim=0)
-        timesteps = self._shift_respace_fn(endpoints, timeshift).to(device=device)
-        timedeltas = (timesteps[1:] - timesteps[:-1]).to(device=device)
+        endpoints = torch.linspace(0.0, 1.0 - float(last_step), int(num_inference_steps), dtype=torch.float64)
+        endpoints = torch.cat([endpoints, torch.tensor([1.0], dtype=torch.float64)], dim=0)
+        timesteps = self._shift_respace_fn(endpoints, timeshift).to(device=device, dtype=torch.float64)
+        timedeltas = (timesteps[1:] - timesteps[:-1]).to(device=device, dtype=torch.float64)
 
         solver_coeffs: List[List[float]] = [[] for _ in range(int(num_inference_steps))]
         for i in range(int(num_inference_steps)):
@@ -149,14 +205,30 @@ class DeCoFlowMatchAdamDiscreteScheduler(SchedulerMixin, ConfigMixin):
     def scale_model_input(self, sample: torch.Tensor, timestep: Optional[torch.Tensor] = None) -> torch.Tensor:
         return sample
 
+    def effective_guidance_scale(self, timestep: Union[torch.Tensor, float]) -> float:
+        """Match AdamLMSampler: apply CFG only when t > min and t < max."""
+        t = float(timestep)
+        interval_min = float(getattr(self.config, "guidance_interval_min", self.guidance_interval_min))
+        interval_max = float(getattr(self.config, "guidance_interval_max", self.guidance_interval_max))
+        if t > interval_min and t < interval_max:
+            return float(self.guidance_scale)
+        return 1.0
+
     def classifier_free_guidance(
         self,
         model_output: torch.Tensor,
         guidance_scale: Optional[float] = None,
+        timestep: Optional[Union[torch.Tensor, float]] = None,
     ) -> torch.Tensor:
         if model_output.shape[0] % 2 != 0:
             raise ValueError("Classifier-free guidance expects concatenated unconditional/conditional batches.")
-        scale = self.guidance_scale if guidance_scale is None else float(guidance_scale)
+        if guidance_scale is None:
+            if timestep is None:
+                scale = float(self.guidance_scale)
+            else:
+                scale = self.effective_guidance_scale(timestep)
+        else:
+            scale = float(guidance_scale)
         uncond, cond = model_output.chunk(2, dim=0)
         return uncond + scale * (cond - uncond)
 
@@ -182,7 +254,9 @@ class DeCoFlowMatchAdamDiscreteScheduler(SchedulerMixin, ConfigMixin):
         for coeff, output in zip(coeffs, recent):
             pred = pred + coeff * output
 
-        prev_sample = sample + pred * self._timedeltas[self._step_index]
+        delta = self._timedeltas[self._step_index].to(device=sample.device, dtype=torch.float64)
+        prev_sample = sample.to(dtype=torch.float64) + pred.to(dtype=torch.float64) * delta
+        prev_sample = prev_sample.to(dtype=sample.dtype)
         self._step_index += 1
 
         if not return_dict:

DeCo-XXL-16-512-t2i/scripts/compare_official_ckpt.py

@@ -0,0 +1,123 @@
+#!/usr/bin/env python3
+"""Compare diffusers pipeline output vs official DeCo t2i_DeCo.ckpt sampling."""
+
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+import torch
+
+ROOT = Path(__file__).resolve().parents[1]
+MODELS_ROOT = ROOT.parents[2]
+OFFICIAL = Path("/tmp/DeCo-official")
+sys.path.insert(0, str(OFFICIAL))
+
+from diffusers import DiffusionPipeline  # noqa: E402
+from src.diffusion.base.guidance import simple_guidance_fn  # noqa: E402
+from src.diffusion.flow_matching.adam_sampling import AdamLMSampler  # noqa: E402
+from src.diffusion.flow_matching.scheduling import LinearScheduler  # noqa: E402
+from src.models.conditioner.qwen3_text_encoder import Qwen3TextEncoder  # noqa: E402
+from src.models.transformer.dit_t2i_DeCo import PixNerDiT  # noqa: E402
+
+PROMPT = "a golden retriever playing in the snow, high quality photograph"
+NEG = "Unrealistic, JPEG artifacts."
+SEED = 42
+STEPS = 25
+GUIDANCE = 4.0
+TIMESHIFT = 3.0
+ORDER = 2
+
+
+def official_sample(ckpt_path: Path, text_path: Path, device: str) -> torch.Tensor:
+    ckpt = torch.load(ckpt_path, map_location="cpu", weights_only=False)
+    sd = ckpt["state_dict"]
+    den = PixNerDiT(
+        in_channels=3,
+        patch_size=16,
+        num_groups=24,
+        hidden_size=1536,
+        txt_embed_dim=2048,
+        txt_max_length=128,
+        num_text_blocks=4,
+        decoder_hidden_size=32,
+        num_encoder_blocks=16,
+        num_decoder_blocks=3,
+    )
+    prefix = "ema_denoiser."
+    for k, v in den.state_dict().items():
+        v.copy_(sd[prefix + k])
+    den.eval().to(device=device, dtype=torch.bfloat16)
+
+    cond = Qwen3TextEncoder(str(text_path), embed_dim=2048, max_length=128)
+    cond.model.to(device=device, dtype=torch.bfloat16)
+
+    sampler = AdamLMSampler(
+        order=ORDER,
+        scheduler=LinearScheduler(),
+        guidance_fn=simple_guidance_fn,
+        num_steps=STEPS,
+        guidance=GUIDANCE,
+        timeshift=TIMESHIFT,
+    )
+    gen = torch.Generator(device="cpu").manual_seed(SEED)
+    noise = torch.randn((1, 3, 512, 512), device="cpu", generator=gen, dtype=torch.float32).to(device)
+
+    with torch.no_grad(), torch.autocast(device, dtype=torch.bfloat16):
+        condition, uncondition = cond([PROMPT], {"negative_prompt": NEG})
+        out = sampler(den, noise, condition, uncondition, return_x_trajs=True)
+        samples = out[1] if isinstance(out, tuple) else [out]
+    return samples[-1].float().cpu()
+
+
+def diffusers_sample(model_dir: Path, device: str) -> torch.Tensor:
+    pipe = DiffusionPipeline.from_pretrained(
+        str(model_dir),
+        local_files_only=True,
+        custom_pipeline=str(model_dir / "pipeline.py"),
+        trust_remote_code=True,
+        torch_dtype=torch.bfloat16,
+    ).to(device)
+    gen = torch.Generator(device="cpu").manual_seed(SEED)
+    with torch.inference_mode():
+        out = pipe(
+            prompt=PROMPT,
+            negative_prompt=NEG,
+            num_inference_steps=STEPS,
+            guidance_scale=GUIDANCE,
+            timeshift=TIMESHIFT,
+            order=ORDER,
+            generator=gen,
+            output_type="np",
+        )
+    # pipeline returns uint8 HWC; convert to float CHW [-1,1]
+    img = torch.from_numpy(out.images[0]).permute(2, 0, 1).float() / 127.5 - 1.0
+    return img
+
+
+def main() -> None:
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    ckpt = ROOT.parent / "t2i_DeCo.ckpt"
+    text_path = MODELS_ROOT / "Qwen" / "Qwen3-1.7B"
+    if not text_path.exists():
+        text_path = ROOT / "text_encoder"
+
+    print("device:", device)
+    print("ckpt:", ckpt)
+    print("text:", text_path)
+
+    off = official_sample(ckpt, text_path, device)
+    dif = diffusers_sample(ROOT, device)
+
+    diff = (off - dif).abs()
+    print("official:", tuple(off.shape), off.mean().item(), off.std().item())
+    print("diffusers:", tuple(dif.shape), dif.mean().item(), dif.std().item())
+    print("abs diff: max", diff.max().item(), "mean", diff.mean().item())
+    if diff.max().item() < 0.05:
+        print("MATCH (close)")
+    else:
+        print("MISMATCH")
+
+
+if __name__ == "__main__":
+    main()

DeCo-XXL-16-512-t2i/scripts/run_t2i_demo.py

@@ -12,17 +12,18 @@ MODEL_DIR = Path(__file__).resolve().parents[1]
 
 def main() -> None:
     device = "cuda" if torch.cuda.is_available() else "cpu"
-    dtype = torch.bfloat16 if device == "cuda" else torch.float32
+    load_dtype = torch.bfloat16 if device == "cuda" else torch.float32
 
     pipe = DiffusionPipeline.from_pretrained(
         str(MODEL_DIR),
         local_files_only=True,
         custom_pipeline=str(MODEL_DIR / "pipeline.py"),
         trust_remote_code=True,
-        torch_dtype=dtype,
+        torch_dtype=load_dtype,
     )
     print("text_encoder:", type(pipe.text_encoder).__name__)
     pipe.to(device)
+    # Denoiser runs float32 at inference (matches official GenEval / Lightning predict).
 
     prompt = "a golden retriever playing in the snow, high quality photograph"
     # Official DeCo uses CPU generator for reproducible noise (app.py / GenEval).

DeCo-XXL-16-512-t2i/text_encoder/.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

DeCo-XXL-16-512-t2i/text_encoder/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2024 Alibaba Cloud
+
+   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.

DeCo-XXL-16-512-t2i/text_encoder/README.md

@@ -0,0 +1,301 @@
+---
+library_name: transformers
+license: apache-2.0
+license_link: https://huggingface.co/Qwen/Qwen3-1.7B/blob/main/LICENSE
+pipeline_tag: text-generation
+base_model:
+- Qwen/Qwen3-1.7B-Base
+---
+
+# Qwen3-1.7B
+<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
+    <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
+</a>
+
+## Qwen3 Highlights
+
+Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features:
+
+- **Uniquely support of seamless switching between thinking mode** (for complex logical reasoning, math, and coding) and **non-thinking mode** (for efficient, general-purpose dialogue) **within single model**, ensuring optimal performance across various scenarios.
+- **Significantly enhancement in its reasoning capabilities**, surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking mode) on mathematics, code generation, and commonsense logical reasoning.
+- **Superior human preference alignment**, excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a more natural, engaging, and immersive conversational experience.
+- **Expertise in agent capabilities**, enabling precise integration with external tools in both thinking and unthinking modes and achieving leading performance among open-source models in complex agent-based tasks.
+- **Support of 100+ languages and dialects** with strong capabilities for **multilingual instruction following** and **translation**.
+
+## Model Overview
+
+**Qwen3-1.7B** has the following features:
+- Type: Causal Language Models
+- Training Stage: Pretraining & Post-training
+- Number of Parameters: 1.7B
+- Number of Paramaters (Non-Embedding): 1.4B
+- Number of Layers: 28
+- Number of Attention Heads (GQA): 16 for Q and 8 for KV
+- Context Length: 32,768
+
+For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).
+
+> [!TIP]
+> If you encounter significant endless repetitions, please refer to the [Best Practices](#best-practices) section for optimal sampling parameters, and set the ``presence_penalty`` to 1.5.
+
+## Quickstart
+
+The code of Qwen3 has been in the latest Hugging Face `transformers` and we advise you to use the latest version of `transformers`.
+
+With `transformers<4.51.0`, you will encounter the following error:
+```
+KeyError: 'qwen3'
+```
+
+The following contains a code snippet illustrating how to use the model generate content based on given inputs. 
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model_name = "Qwen/Qwen3-1.7B"
+
+# load the tokenizer and the model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype="auto",
+    device_map="auto"
+)
+
+# prepare the model input
+prompt = "Give me a short introduction to large language model."
+messages = [
+    {"role": "user", "content": prompt}
+]
+text = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True,
+    enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.
+)
+model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+
+# conduct text completion
+generated_ids = model.generate(
+    **model_inputs,
+    max_new_tokens=32768
+)
+output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 
+
+# parsing thinking content
+try:
+    # rindex finding 151668 (</think>)
+    index = len(output_ids) - output_ids[::-1].index(151668)
+except ValueError:
+    index = 0
+
+thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
+content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
+
+print("thinking content:", thinking_content)
+print("content:", content)
+```
+
+For deployment, you can use `sglang>=0.4.6.post1` or `vllm>=0.8.5` or to create an OpenAI-compatible API endpoint:
+- SGLang:
+    ```shell
+    python -m sglang.launch_server --model-path Qwen/Qwen3-1.7B --reasoning-parser qwen3
+    ```
+- vLLM:
+    ```shell
+    vllm serve Qwen/Qwen3-1.7B --enable-reasoning --reasoning-parser deepseek_r1
+    ```
+
+For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.
+
+## Switching Between Thinking and Non-Thinking Mode
+
+> [!TIP]
+> The `enable_thinking` switch is also available in APIs created by SGLang and vLLM. 
+> Please refer to our documentation for [SGLang](https://qwen.readthedocs.io/en/latest/deployment/sglang.html#thinking-non-thinking-modes) and [vLLM](https://qwen.readthedocs.io/en/latest/deployment/vllm.html#thinking-non-thinking-modes) users.
+
+### `enable_thinking=True`
+
+By default, Qwen3 has thinking capabilities enabled, similar to QwQ-32B. This means the model will use its reasoning abilities to enhance the quality of generated responses. For example, when explicitly setting `enable_thinking=True` or leaving it as the default value in `tokenizer.apply_chat_template`, the model will engage its thinking mode.
+
+```python
+text = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True,
+    enable_thinking=True  # True is the default value for enable_thinking
+)
+```
+
+In this mode, the model will generate think content wrapped in a `<think>...</think>` block, followed by the final response.
+
+> [!NOTE]
+> For thinking mode, use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0` (the default setting in `generation_config.json`). **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
+
+
+### `enable_thinking=False`
+
+We provide a hard switch to strictly disable the model's thinking behavior, aligning its functionality with the previous Qwen2.5-Instruct models. This mode is particularly useful in scenarios where disabling thinking is essential for enhancing efficiency.
+
+```python
+text = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True,
+    enable_thinking=False  # Setting enable_thinking=False disables thinking mode
+)
+```
+
+In this mode, the model will not generate any think content and will not include a `<think>...</think>` block.
+
+> [!NOTE]
+> For non-thinking mode, we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`. For more detailed guidance, please refer to the [Best Practices](#best-practices) section.
+
+### Advanced Usage: Switching Between Thinking and Non-Thinking Modes via User Input
+
+We provide a soft switch mechanism that allows users to dynamically control the model's behavior when `enable_thinking=True`. Specifically, you can add `/think` and `/no_think` to user prompts or system messages to switch the model's thinking mode from turn to turn. The model will follow the most recent instruction in multi-turn conversations.
+
+Here is an example of a multi-turn conversation:
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+class QwenChatbot:
+    def __init__(self, model_name="Qwen/Qwen3-1.7B"):
+        self.tokenizer = AutoTokenizer.from_pretrained(model_name)
+        self.model = AutoModelForCausalLM.from_pretrained(model_name)
+        self.history = []
+
+    def generate_response(self, user_input):
+        messages = self.history + [{"role": "user", "content": user_input}]
+
+        text = self.tokenizer.apply_chat_template(
+            messages,
+            tokenize=False,
+            add_generation_prompt=True
+        )
+
+        inputs = self.tokenizer(text, return_tensors="pt")
+        response_ids = self.model.generate(**inputs, max_new_tokens=32768)[0][len(inputs.input_ids[0]):].tolist()
+        response = self.tokenizer.decode(response_ids, skip_special_tokens=True)
+
+        # Update history
+        self.history.append({"role": "user", "content": user_input})
+        self.history.append({"role": "assistant", "content": response})
+
+        return response
+
+# Example Usage
+if __name__ == "__main__":
+    chatbot = QwenChatbot()
+
+    # First input (without /think or /no_think tags, thinking mode is enabled by default)
+    user_input_1 = "How many r's in strawberries?"
+    print(f"User: {user_input_1}")
+    response_1 = chatbot.generate_response(user_input_1)
+    print(f"Bot: {response_1}")
+    print("----------------------")
+
+    # Second input with /no_think
+    user_input_2 = "Then, how many r's in blueberries? /no_think"
+    print(f"User: {user_input_2}")
+    response_2 = chatbot.generate_response(user_input_2)
+    print(f"Bot: {response_2}") 
+    print("----------------------")
+
+    # Third input with /think
+    user_input_3 = "Really? /think"
+    print(f"User: {user_input_3}")
+    response_3 = chatbot.generate_response(user_input_3)
+    print(f"Bot: {response_3}")
+```
+
+> [!NOTE]
+> For API compatibility, when `enable_thinking=True`, regardless of whether the user uses `/think` or `/no_think`, the model will always output a block wrapped in `<think>...</think>`. However, the content inside this block may be empty if thinking is disabled.
+> When `enable_thinking=False`, the soft switches are not valid. Regardless of any `/think` or `/no_think` tags input by the user, the model will not generate think content and will not include a `<think>...</think>` block.
+
+## Agentic Use
+
+Qwen3 excels in tool calling capabilities. We recommend using [Qwen-Agent](https://github.com/QwenLM/Qwen-Agent) to make the best use of agentic ability of Qwen3. Qwen-Agent encapsulates tool-calling templates and tool-calling parsers internally, greatly reducing coding complexity.
+
+To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.
+```python
+from qwen_agent.agents import Assistant
+
+# Define LLM
+llm_cfg = {
+    'model': 'Qwen3-1.7B',
+
+    # Use the endpoint provided by Alibaba Model Studio:
+    # 'model_type': 'qwen_dashscope',
+    # 'api_key': os.getenv('DASHSCOPE_API_KEY'),
+
+    # Use a custom endpoint compatible with OpenAI API:
+    'model_server': 'http://localhost:8000/v1',  # api_base
+    'api_key': 'EMPTY',
+
+    # Other parameters:
+    # 'generate_cfg': {
+    #         # Add: When the response content is `<think>this is the thought</think>this is the answer;
+    #         # Do not add: When the response has been separated by reasoning_content and content.
+    #         'thought_in_content': True,
+    #     },
+}
+
+# Define Tools
+tools = [
+    {'mcpServers': {  # You can specify the MCP configuration file
+            'time': {
+                'command': 'uvx',
+                'args': ['mcp-server-time', '--local-timezone=Asia/Shanghai']
+            },
+            "fetch": {
+                "command": "uvx",
+                "args": ["mcp-server-fetch"]
+            }
+        }
+    },
+  'code_interpreter',  # Built-in tools
+]
+
+# Define Agent
+bot = Assistant(llm=llm_cfg, function_list=tools)
+
+# Streaming generation
+messages = [{'role': 'user', 'content': 'https://qwenlm.github.io/blog/ Introduce the latest developments of Qwen'}]
+for responses in bot.run(messages=messages):
+    pass
+print(responses)
+```
+
+## Best Practices
+
+To achieve optimal performance, we recommend the following settings:
+
+1. **Sampling Parameters**:
+   - For thinking mode (`enable_thinking=True`), use `Temperature=0.6`, `TopP=0.95`, `TopK=20`, and `MinP=0`. **DO NOT use greedy decoding**, as it can lead to performance degradation and endless repetitions.
+   - For non-thinking mode (`enable_thinking=False`), we suggest using `Temperature=0.7`, `TopP=0.8`, `TopK=20`, and `MinP=0`.
+   - For supported frameworks, you can adjust the `presence_penalty` parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
+
+2. **Adequate Output Length**: We recommend using an output length of 32,768 tokens for most queries. For benchmarking on highly complex problems, such as those found in math and programming competitions, we suggest setting the max output length to 38,912 tokens. This provides the model with sufficient space to generate detailed and comprehensive responses, thereby enhancing its overall performance.
+
+3. **Standardize Output Format**: We recommend using prompts to standardize model outputs when benchmarking.
+   - **Math Problems**: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
+   - **Multiple-Choice Questions**: Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the `answer` field with only the choice letter, e.g., `"answer": "C"`."
+
+4. **No Thinking Content in History**: In multi-turn conversations, the historical model output should only include the final output part and does not need to include the thinking content. It is implemented in the provided chat template in Jinja2. However, for frameworks that do not directly use the Jinja2 chat template, it is up to the developers to ensure that the best practice is followed.
+
+### Citation
+
+If you find our work helpful, feel free to give us a cite.
+
+```
+@misc{qwen3technicalreport,
+      title={Qwen3 Technical Report}, 
+      author={Qwen Team},
+      year={2025},
+      eprint={2505.09388},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL},
+      url={https://arxiv.org/abs/2505.09388}, 
+}
+```

DeCo-XXL-16-512-t2i/text_encoder/config.json

@@ -0,0 +1,30 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "max_position_embeddings": 40960,
+  "max_window_layers": 28,
+  "model_type": "qwen3",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.0",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

DeCo-XXL-16-512-t2i/text_encoder/generation_config.json

@@ -0,0 +1,13 @@
+{
+    "bos_token_id": 151643,
+    "do_sample": true,
+    "eos_token_id": [
+        151645,
+        151643
+    ],
+    "pad_token_id": 151643,
+    "temperature": 0.6,
+    "top_k": 20,
+    "top_p": 0.95,
+    "transformers_version": "4.51.0"
+}

DeCo-XXL-16-512-t2i/text_encoder/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:169ad53ec313c3a34b06c0809216e4fc072cce444a5d4ff2b59690d064130ed5
+size 3441185608

DeCo-XXL-16-512-t2i/text_encoder/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:912becff8d60672aa8628ef08c05898d9adf17c2ad4ae3caf99b065622fdeff9
+size 622329984

DeCo-XXL-16-512-t2i/text_encoder/model.safetensors.index.json

@@ -0,0 +1,318 @@
+{
+  "metadata": {
+    "total_size": 4063479808
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00001-of-00002.safetensors"
+  }
+}

DeCo-XXL-16-512-t2i/text_encoder/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

DeCo-XXL-16-512-t2i/text_encoder/tokenizer_config.json

@@ -0,0 +1,239 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {{- messages[0].content + '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n    {%- set index = (messages|length - 1) - loop.index0 %}\n    {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n        {%- set ns.multi_step_tool = false %}\n        {%- set ns.last_query_index = index %}\n    {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n    {%- if message.content is string %}\n        {%- set content = message.content %}\n    {%- else %}\n        {%- set content = '' %}\n    {%- endif %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set reasoning_content = '' %}\n        {%- if message.reasoning_content is string %}\n            {%- set reasoning_content = message.reasoning_content %}\n        {%- else %}\n            {%- if '</think>' in content %}\n                {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n                {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n            {%- endif %}\n        {%- endif %}\n        {%- if loop.index0 > ns.last_query_index %}\n            {%- if loop.last or (not loop.last and reasoning_content) %}\n                {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n            {%- else %}\n                {{- '<|im_start|>' + message.role + '\\n' + content }}\n            {%- endif %}\n        {%- else %}\n            {{- '<|im_start|>' + message.role + '\\n' + content }}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n    {%- if enable_thinking is defined and enable_thinking is false %}\n        {{- '<think>\\n\\n</think>\\n\\n' }}\n    {%- endif %}\n{%- endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

DeCo-XXL-16-512-t2i/transformer/transformer_deco_t2i.py

@@ -58,18 +58,21 @@ class TimestepEmbedder(nn.Module):
     @staticmethod
     def timestep_embedding(t: torch.Tensor, dim: int, max_period: int = 10) -> torch.Tensor:
         half = dim // 2
+        compute_dtype = torch.float64 if t.dtype == torch.float64 else torch.float32
         freqs = torch.exp(
-            -math.log(max_period) * torch.arange(start=0, end=half, dtype=torch.float32, device=t.device) / half
+            -math.log(max_period)
+            * torch.arange(start=0, end=half, dtype=compute_dtype, device=t.device)
+            / half
         )
-        args = t[..., None].float() * freqs[None, ...]
+        args = t[..., None].to(compute_dtype) * freqs[None, ...]
         embedding = torch.cat([torch.cos(args), torch.sin(args)], dim=-1)
         if dim % 2:
             embedding = torch.cat([embedding, torch.zeros_like(embedding[:, :1])], dim=-1)
-        return embedding.to(t.dtype)
+        return embedding
 
     def forward(self, t: torch.Tensor) -> torch.Tensor:
         t_freq = self.timestep_embedding(t, self.frequency_embedding_size)
-        return self.mlp(t_freq)
+        return self.mlp(t_freq.to(self.mlp[0].weight.dtype))
 
 
 class DeCoSwiGLU(nn.Module):
@@ -379,8 +382,8 @@ class DeCoT2ITransformer2DModel(ModelMixin, ConfigMixin):
         self, timestep: Union[torch.Tensor, float, int], batch_size: int, sample: torch.Tensor
     ) -> torch.Tensor:
         if not isinstance(timestep, torch.Tensor):
-            timestep = torch.tensor([timestep], device=sample.device, dtype=sample.dtype)
-        timestep = timestep.to(device=sample.device, dtype=sample.dtype)
+            timestep = torch.tensor([timestep], device=sample.device, dtype=torch.float64)
+        timestep = timestep.to(device=sample.device, dtype=torch.float64)
         if timestep.ndim == 0:
             timestep = timestep[None]
         if timestep.shape[0] == 1 and batch_size > 1:

README.md

@@ -107,3 +107,35 @@ image.save("deco_xl_256_demo.png")
 ```
 
 Integer class ids, batched labels, and optional `batch_size` for repeating a single label are also supported.
+
+### 3) Text-to-image (`DeCo-XXL-16-512-t2i` / `t2i_DeCo.ckpt`)
+
+Use the **AdamLM** scheduler defaults from official DeCo (not the c2i 100-step / CFG 5.0 recipe):
+
+```python
+import torch
+from diffusers import DiffusionPipeline
+
+model_path = "./DeCo-XXL-16-512-t2i"
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+pipe = DiffusionPipeline.from_pretrained(
+    model_path,
+    trust_remote_code=True,
+    custom_pipeline=f"{model_path}/pipeline.py",
+    torch_dtype=torch.bfloat16,
+).to(device)
+
+# Bundled ./text_encoder (Qwen3-1.7B weights + tokenizer). Pipeline loads both from that folder.
+# Denoiser runs in float32 during __call__ (matches official GenEval predict).
+
+image = pipe(
+    prompt="a golden retriever playing in the snow, high quality photograph",
+    negative_prompt="Unrealistic, JPEG artifacts.",
+    num_inference_steps=25,
+    guidance_scale=4.0,
+    timeshift=3.0,
+    generator=torch.Generator(device="cpu").manual_seed(42),
+).images[0]
+image.save("deco_t2i_demo.png")
+```