Wan-AI/Wan2.2-TI2V-5B-Diffusers provider: huggingface license: apache-2.0 notes: Unified Wan 2.2 TI2V 5B Diffusers checkpoint. Supports both text-to-video and image-to-video at 720p/24fps. AbstractVision runs it locally through the MLX-Gen Wan runtime on Apple Silicon; the full Diffusers snapshot remains cataloged as a fallback artifact. downloads: - wan2.2-ti2v-5b engine=mlx-gen target=mlx bits=8 repo=AbstractFramework/wan2.2-ti2v-5b-diffusers-8bit source=abstractframework-mlx-gen notes: AbstractFramework MLX-Gen prepared 8-bit folder for the unified Wan 2.2 TI2V 5B checkpoint. Supports both text-to-video and image-to-video. - wan2.2-ti2v-5b engine=mlx-gen target=mlx bits=16 repo=AbstractFramework/wan2.2-ti2v-5b-diffusers-bf16 source=abstractframework-mlx-gen notes: AbstractFramework MLX-Gen prepared BF16 folder for the unified Wan 2.2 TI2V 5B checkpoint. Supports both text-to-video and image-to-video. - wan2.2-ti2v-5b engine=mlx-gen target=mlx bits=16 repo=Wan-AI/Wan2.2-TI2V-5B-Diffusers source=official notes: Official Wan TI2V snapshot consumed by MLX-Gen 0.18.7+ for both text-to-video and image-to-video. This is not a prepared q4/q8 folder. - wan2.2-ti2v-5b engine=diffusers target=diffusers bits=16 repo=Wan-AI/Wan2.2-TI2V-5B-Diffusers source=official notes: Official Diffusers snapshot for the unified Wan 2.2 TI2V 5B checkpoint. tasks: - image_to_video requires: {"backend": "mlx-gen", "min_runtime_version": "0.18.7"} required params: image optional params: fps, guidance_scale, height, negative_prompt, num_frames, prompt, seed, steps, width defaults: fps=24, guidance_scale=5.0, num_frames=121, steps=50 - text_to_video requires: {"backend": "mlx-gen", "min_runtime_version": "0.18.7"} required params: prompt optional params: flow_shift, fps, guidance_scale, height, negative_prompt, num_frames, seed, steps, width defaults: flow_shift=5.0, fps=24, guidance_scale=5.0, num_frames=121, steps=50 runtime routes: - AbstractFramework/wan2.2-ti2v-5b-diffusers-8bit task: image_to_video supports_lora: True lora_status: validated lora_target_roles: transformer lora_validation_profile: lora_wan_ti2v5b_q8_crushit_i2v_2026_06_11 task: text_to_video supports_lora: True lora_status: validated lora_target_roles: transformer lora_validation_profile: lora_wan_ti2v5b_q8_hstoric_t2v_2026_06_11 - Wan-AI/Wan2.2-TI2V-5B-Diffusers task: image_to_video supports_lora: True lora_status: mapped-unvalidated lora_target_roles: transformer task: text_to_video supports_lora: True lora_status: mapped-unvalidated lora_target_roles: transformer