Skip to content
Video

Camera adapter v2 for Wan2.2-TI2V-5B

virtual camera control for video generation

2 min readupdated

This lightweight adapter adds precise virtual camera control to the frozen Wan-AI/Wan2.2-TI2V-5B-Diffusers base model. Feed it any static image and it outputs a 121-frame video at 1280x704 resolution and 24 fps, smoothly panning left or right, tilting up or down, or zooming in and out. A single strength parameter from 0 to 1 controls the intensity of the motion. It is ideal for turning still images into dynamic video clips on high-VRAM GPUs.

Key features

  • Six movement directions: supports pan, tilt, and zoom with adjustable motion strength.
  • Lightweight footprint: only 14.26M trainable parameters (57.1 MB in float32).
  • Diffusers integration: runs on top of the frozen Wan2.2 base without altering its core weights.
  • Padding mode: uses replicate padding to cleanly generate newly revealed areas as the camera moves.

What you can do

  • Bring static landscape photos to life with smooth zooms or pans.
  • Generate short clips with dynamic camera motion for presentations or concept art.
  • Test this early checkpoint (step 415) using a Python script and the diffusers library.

Example

from pathlib import Path
from huggingface_hub import snapshot_download
import zipfile, sys

local = Path(snapshot_download("D3vil-Reap3r/ie643-wan22-camera-adapter-v2", local_dir="ie643_adapter_v2"))
zipfile.ZipFile(local / "code" / "fpcamera_src.zip").extractall("fpcamera_code")
sys.path.insert(0, str(Path("fpcamera_code") / "src"))

from fpcamera import flow_infer
record = flow_infer.generate(
    ".",
    local / "adapter" / "adapter_000415.safetensors",
    "my_photo.jpg",
    "out/run1",
    motion="zoom_in",
    strength=0.6,
    steps=50, seed=42
)
Read moreDetails
• Base model: Wan-AI/Wan2.2-TI2V-5B-Diffusers (revision b8fff7315c768468a5333511427288870b2e9635).
• Hardware requirements: CUDA GPU with around 43 GiB peak VRAM usage.
• Limitations: early checkpoint (step 415 of 1000), uncalibrated strength scale, fixed 1280x704 output resolution.
• Dependencies: torch 2.11.x, diffusers 0.36.0, transformers 4.57.6.
Discuss Telegram 0

Will it run?

Runs in BF16

Full precision does not fit, a quantised version does: quality barely changes down to 8 bits, more noticeably at 4 bits.

  • F32weights 19 GB · ready files24 GB
  • BF16weights 9.4 GB14 GB
  • FP8 / INT8recommendedweights 4.7 GB8.4 GB
  • 4-bitweights 2.7 GB6.1 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 3080 · F32 — from RTX 5090

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

This is a LoRA: counted together with its base model Wan2.2-TI2V-5B-Diffusers.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # D3vil-Reap3r/ie643-wan22-camera-adapter-v2
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download Wan-AI/Wan2.2-TI2V-5B-Diffusers --local-dir ~/models/Wan2.2-TI2V-5B-Diffusers
    hf download D3vil-Reap3r/ie643-wan22-camera-adapter-v2 --local-dir ~/models/ie643-wan22-camera-adapter-v2
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • SageAttention 3v2.2.0from October 28, 2025

    even faster on RTX 50 / Blackwell

    FP4 attention for Blackwell. Faster than SageAttention 2, but the authors suggest v2 when accuracy matters.

    How to enable: Build from the sageattention3_blackwell folder of the repo; in ComfyUI via KJNodes.

    Earlier generations (2)
    • SageAttention 2++ — Quantised (INT8/FP8) attention with no visible quality loss. The main speed-up for image and video generation.
    • SageAttention 1 — First version: INT8 for Q·K. Still works, but 2++ is faster and more accurate.
  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (3)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
    • xFormers — The older way to speed up attention, works almost everywhere. Sage or Flash are faster where available.
  • cache-dit (DBCache)v1.5.6from November 20, 2026new release

    1.5–3× at nearly the same quality

    Reuses block outputs between diffusion steps. Supports most DiT models.

    How to enable: pip install cache-dit; one line cache_dit.enable_cache(pipe) for diffusers.

    Earlier generations (2)
    • MagCache — Skips steps by magnitude of change. More accurate than TeaCache, no calibration.
    • TeaCache — The first popular diffusion step cache. Works, newer methods are more accurate.
  • LightX2V step distill0.6.1from November 9, 2026new release

    5–10× fewer steps

    Distilled LoRAs and an engine: video in 4–8 steps instead of 30–50.

    How to enable: Load their step-distill LoRA, CFG 1.0, 4–8 steps.

  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

Requires the base model

Wan-AI/Wan2.2-TI2V-5B-Diffusers

This model is a LoRA / adapter. Download the base model first, then load this one on top.

New posts about versions of Wan2.2-TI2V-5B-Diffusers (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 1All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 10All on Hugging Face

About the model

D3vil-Reap3r/ie643-wan22-camera-adapter-v2
Author
D3vil-Reap3r
Category
Video
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments