Bernini V2 ComfyUI weights

Native ComfyUI packages for ByteDance/Bernini-Diffusers-v2. They are used by T8mars/comfyui-bernini-v2-T8 and do not require a Diffusers pipeline at runtime. Native ComfyUI Core support is under review in Comfy-Org/ComfyUI#16001.

Bernini v2 first plans the requested scene and motion with Qwen2.5-VL, then renders the result with two Wan2.2 experts. It supports text-to-image, image-to-image, text-to-video, video-to-video, reference-to-video, and reference-guided video editing.

Available packages

Folder	Storage	Size	Recommended use
`Bernini-v2-balanced-int8/`	ComfyUI `int8_tensorwise` + ConvRot, BF16 fallback	45.62 GiB	Recommended for 24 GB GPUs
`Bernini-v2-bf16-native/`	True BF16	83.03 GiB	Reference-quality and conversion source

The Balanced INT8 package quantizes 1,300 linear layers across the planner, T5, and both Wan experts. Every converted layer was reconstructed and checked; no layer required a quality fallback at cosine 0.99 and relative-error 2% thresholds.

Both packages come from the immutable upstream revision 399cf6a18a4c523b367b2b1ac25a2a61009e7df3. Their manifests contain the source revision, component indexes, tensor dtypes, byte counts, and SHA-256 hashes for every shard.

Installation

Install Bernini v2 (Native) from ComfyUI-Manager, or clone the node repository into ComfyUI/custom_nodes.

Download the recommended package:

hf download t8star/Bernini-V2-Comfy \
  --include "Bernini-v2-balanced-int8/*" \
  --local-dir /path/to/ComfyUI/models/bernini_v2

For BF16 instead:

hf download t8star/Bernini-V2-Comfy \
  --include "Bernini-v2-bf16-native/*" \
  --local-dir /path/to/ComfyUI/models/bernini_v2

Bernini v2 also needs the standard Wan 2.1 VAE at:

ComfyUI/models/vae/wan_2.1_vae.safetensors

The node repository includes a pinned downloader:

python tools/download_vae.py --output /path/to/ComfyUI/models/vae

Open one of the six workflows under the node repository's examples/workflows/ directory after the files are in place.

Validated runtime

The current ComfyUI Core candidate completed the official T2V preset at 640×368, 33 frames, and 16 fps: 50 MaskGIT planning steps, one VIT denoising step, 50 flow-UniPC renderer steps, and the released T2V guidance values. All 33 decoded frames are unique. Peak ComfyUI-visible VRAM was 23.255 GiB on the 24 GB test GPU.

The Balanced INT8 package passed a production-step 640×368, 33-frame T2V run on a 24 GB RTX 5090 Laptop GPU. ComfyUI-visible VRAM peaked at 16.51 GiB. Two consecutive uncached jobs in one server process produced identical decoded frames and returned memory to baseline. Reduced two-second, long-edge-640 T2V, V2V, R2V, and RV2V runs also passed.

Detailed settings, hashes, and visual evidence are available in the quality report and low-memory guide.

License and attribution

The original Bernini v2 weights and code are released under Apache License 2.0. These dtype-converted and quantized packages remain under Apache-2.0.

Original model: https://huggingface.co/ByteDance/Bernini-Diffusers-v2
Original source: https://github.com/bytedance/Bernini
Native ComfyUI nodes: https://github.com/T8mars/comfyui-bernini-v2-T8

Model tree for t8star/Bernini-V2-Comfy

Base model

ByteDance/Bernini-Diffusers-v2

Finetuned

(4)

this model