HY-WorldPlay

Project page Official code

Introduced by Tencent Hunyuan, HY-WorldPlay is a real-time interactive image-to-video (I2V) world model with action + camera-trajectory conditioning and reconstituted-context memory. FlashDreams ships a native port of the distilled WAN-5B variant (Wan 2.2 TI2V-5B backbone, 4-step distilled Euler).

Generated with FlashDreams' native HY-WorldPlay WAN-5B I2V pipeline.

Installation

# from the repo root
uv sync --project integrations_v2/hy_worldplay

Running the method

HY-WorldPlay WAN-5B is image-to-video only. Its model package binds the pipeline directly to the reusable Cam2V v2 application:

uv sync --package flashdreams-hy-worldplay --inexact
uv run --no-sync flashdreams-run-v2 cam2v-hy-worldplay \
    --mode webrtc --host 0.0.0.0 --port 8089 -- --example-data

Use W/S to move, A/D to yaw, Q/E to strafe, and I/K to pitch. The binding converts live camera poses to HY’s latent-rate PRoPE, action, and memory inputs. Application arguments follow --; inspect them with:

uv run --no-sync flashdreams-run-v2 cam2v-hy-worldplay -- --help

Some generated samples from the above commands:

a person walking

Walking through a seaside village

Walking through a snowy forest

Walking toward a castle

Profiling benchmark

Here is the profiling benchmark on total DiT + VAE encode/decode runtime for FlashDreams HY-WorldPlay compared to the official HY-WorldPlay implementation under matched settings.

This chart shows total DiT + VAE-decode runtime per autoregressive chunk (4 diffusion steps) in milliseconds, at steady state (median of the post-warmup chunks), measured at num_chunk=8, 704x1280, seed=0 on a single GB300. For an apples-to-apples comparison, both implementations are forced to use the cuDNN attention backend and torch.compile under matched runtime settings. For the official HY-WorldPlay implementation, see this instruction.