Views
No views yet
| Component | Detail |
|---|---|
| Base | Qwen/Qwen-Image-Edit transformer (original release) |
| Graft | MetaView model-2500-best delta: _3D_in projection + per-block prope_attn parallel 3D attention (PRoPE camera-aware rope), kept bf16 |
| Speed | Qwen-Image-Edit-Lightning-8steps-V1.0 LoRA fused at alpha 0.125 → 8 steps, CFG 1.0 |
| Quantization | 841 base layers fp8_e4m3fn, per-tensor scale = amax/448, ComfyUI-native weight_scale convention; sensitive layers (img_in / txt_in / proj_out / norms / biases) and all MetaView layers bf16 |
| Size | 26.5 GB (vs ~41 GB bf16 stack) — runs on a 16 GB GPU via ComfyUI weight streaming |
MetaViewModelLoader, MetaViewDA3Loader, MetaView3DConditioning) — the grafted architecture is not detected by stock loaders. Everything else is stock ComfyUI.| Model | File / dir | Where |
|---|---|---|
| Text encoder | qwen_2.5_vl_7b_fp8_scaled.safetensors → models/text_encoders/ | Comfy-Org/Qwen-Image_ComfyUI |
| VAE | qwen_image_vae.safetensors → models/vae/ | Comfy-Org/Qwen-Image_ComfyUI |
| 3D features | DA3-GIANT-1.1/ (whole dir) → models/depth_anything_3/ | depth-anything/DA3-GIANT-1.1 |
| Depth | DA3NESTED-GIANT-LARGE-1.1/ (whole dir) → models/depth_anything_3/ | depth-anything/DA3NESTED-GIANT-LARGE-1.1 |
models/diffusion_models/.workflows/). Wiring summary:LoadImage ──┬─▶ MetaView3DConditioning (yaw / pitch / radius, auto_radius) ─▶ positive ─▶ SamplerCustom
└─▶ TextEncodeQwenImageEdit (image, no vae) ────────────────────▶ (via conditioning node)
MetaViewDA3Loader ─▶ (da3_model) LATENT from MetaView3DConditioning ─▶ SamplerCustom + MetaViewSigmas
MetaViewModelLoader ─▶ MODEL ─▶ SamplerCustom (euler, CFG 1.0, sigmas from MetaViewSigmas) ─▶ VAEDecode ─▶ SaveSamplerCustom + euler at CFG 1.0, sigmas from the pack's MetaViewSigmas node (8 steps default). Lightning is fused — do not add another speed LoRA, and use MetaViewSigmas rather than a named scheduler (it reproduces the exact Qwen-Image flow-match schedule the model expects).镜头视角转到指定位置 ("move the camera to the specified position"). Extra text after it is best-effort — the model's first-class use is promptless scene mutation.auto_radius derives orbit radius from scene depth (recommended).full_precision_matrix_mult (Comfy-Org scaled-fp8 convention) — weights stay fp8 in RAM/VRAM and dequantize to bf16 for each matmul. No launch flags needed; --fast fp8_matrix_mult does not apply to these layers (this model's outputs degrade under fp8 activation matmul, so the loader opts out per-layer).