Video · text-to-video · image-to-video

HunyuanVideo

tencent

Tencent-Hunyuan GitHub and tencent Hugging Face repos were both reachable.

Runner verified独立环境源码已固定hunyuanvideo

构建运行命令

选择仓库中已记录的 variant，然后复制对应的准备、检查或运行命令。

worldfoundry.pipeline

Variant

任务text-to-video

环境HunyuanVideo

设备CUDA 12.4

worldfoundry-eval evaluate \
  --mode model \
  --model-id hunyuanvideo-t2v \
  --model-runner worldfoundry:pipeline \
  --model-manifest-dir worldfoundry/data/models/catalog \
  --requests-path tmp/requests.jsonl \
  --output-dir tmp/model_eval/hunyuanvideo-t2v \
  --metric artifact_count \
  --json

兼容性与版本

由 Manifest 生成

集成状态: integrated
Runner 证据: verified
环境: HunyuanVideo
Python: 3.10
CUDA: CUDA 12.4
PyTorch: torch
源码 revision: e260ed40c88d104801a8b1de05d2ab81e965a9ef
Checkpoint revision: 6204ad6aea1a77ff5aba337c88278bb9500eb37d

Runtime profile: hunyuanvideo-t2v
Pipeline binding: hunyuanvideo
Runner: worldfoundry.pipeline
Pipeline target: worldfoundry.pipelines.hunyuan_video.pipeline_hunyuan_video:HunyuanVideoT2VPipeline
Backend stage: official_in_tree_runtime
Runtime 状态: official_in_tree_runtime_ready_checkpoint_required
Driver 状态: compatible
环境类型: 独立环境

安装环境

环境解析器会读取已记录 profile，并选择此处显示的统一或独立环境。

bash scripts/setup/model_env_install.sh --model hunyuanvideo-t2v

依赖版本约束 10

torch
torchvision
diffusers==0.31.0
transformers
accelerate
safetensors
imageio
imageio-ffmpeg
opencv-python
sentencepiece

Conda 依赖 3

python
pip
ffmpeg

Checkpoint 与资产

分配算力前先做本地检查；gated、private 与 license 字段直接来自 checkpoint manifest。

worldfoundry-eval zoo model-download --model-id hunyuanvideo --check-local --json

tencent/HunyuanVideo

Revision: 6204ad6aea1a77ff5aba337c88278bb9500eb37d
License: other
Gated: false
Private: false

tencent/HunyuanVideo-I2V

Revision: 3914f209367854b5e470f062c33159d5ab139e1e
License: other
Gated: false
Private: false

运行与输出

生成的命令通过共享 evaluation 边界运行，并持久保存结果 manifest 与 artifact。

worldfoundry-eval evaluate \
  --mode model \
  --model-id hunyuanvideo-t2v \
  --model-runner worldfoundry:pipeline \
  --model-manifest-dir worldfoundry/data/models/catalog \
  --requests-path tmp/requests.jsonl \
  --output-dir tmp/model_eval/hunyuanvideo-t2v \
  --metric artifact_count \
  --json

输入契约

字段	已记录契约
`prompt`	Required
`image`	Optional
`video`	Optional
`actions`	seed

Artifact 契约

Artifact 类型	文件名 / 路径
`generated_video`	`hunyuanvideo-t2v.mp4`
`expected artifact`	`hunyuanvideo_t2v_official_cat_grass_1280x720_129f.mp4`
`expected artifact`	`hunyuanvideo_t2v_worldfoundry_cat_grass_1280x720_129f.mp4`

证据与来源

Catalog 集成、原生 demo parity 与 runner parity 是三条独立记录。

集成状态: integrated
Runner 证据: verified
原生 Demo 证据: verified
验证 Imports: torch, diffusers, transformers

配方溯源

worldfoundry/data/models/catalog/video/hunyuanvideo.yamlGitHube260ed40c88d104801a8b1de05d2ab81e965a9ef

构建运行命令

输入契约

Artifact 契约

相关配方