Extract pinned native visual components; preserve tensor bytes and validate strict loading and native interfaces

28555c1 verified 12 days ago

825 Bytes

{
  "model_type": "native_visual_component_bundle",
  "architectures": [
    "NativeVisualComponents"
  ],
  "loader": "component.load_model",
  "source_repo": "tencent/Unicom-Unified-Multimodal-Modeling-via-Compressed-Continuous-Semantic-Representations",
  "source_revision": "78e43059d8358a045e64fca35698fda86c476154",
  "scope": "Preserves 1152\u219264 semantic compression, 64\u21921152 expansion, vision aligner and latent/time input/output modules. The separate 23.8GB pixel decoder, FLUX VAE and language backbone are not included. IMPORTANT: the published checkpoint aligner requires 1216 input features, while the public direct vision helper emits 1152; this wrapper explicitly rejects implicit adaptation. No speculative concatenation is used. Compression and flow interfaces work independently and are tested."
}

	{
	"model_type": "native_visual_component_bundle",
	"architectures": [
	"NativeVisualComponents"
	],
	"loader": "component.load_model",
	"source_repo": "tencent/Unicom-Unified-Multimodal-Modeling-via-Compressed-Continuous-Semantic-Representations",
	"source_revision": "78e43059d8358a045e64fca35698fda86c476154",
	"scope": "Preserves 1152\u219264 semantic compression, 64\u21921152 expansion, vision aligner and latent/time input/output modules. The separate 23.8GB pixel decoder, FLUX VAE and language backbone are not included. IMPORTANT: the published checkpoint aligner requires 1216 input features, while the public direct vision helper emits 1152; this wrapper explicitly rejects implicit adaptation. No speculative concatenation is used. Compression and flow interfaces work independently and are tested."
	}