lucataco/hunyuan-heygen-joshua:e82a4154 – Run with an API on Replicate

Version

Fast booting

You're looking at a specific version of this model. Jump to the model overview.

lucataco /hunyuan-heygen-joshua:e82a4154

Playground API

Input

prompt

string

Shift + Return to add a new line

HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffeeHGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee

The text prompt describing your video scene.

Default: ""

lora_url

string

Shift + Return to add a new line

A URL pointing to your LoRA .safetensors file or a Hugging Face repo (e.g. 'user/repo' - uses the first .safetensors file).

Default: ""

lora_strength

number

(minimum: -10, maximum: 10)

Scale/strength for your LoRA.

Default: 1

scheduler

string

Algorithm used to generate the video frames.

Default: "DPMSolverMultistepScheduler"

steps

integer

(minimum: 1, maximum: 150)

Number of diffusion steps.

Default: 50

guidance_scale

number

(minimum: 0, maximum: 30)

Overall influence of text vs. model.

Default: 6

flow_shift

integer

(minimum: 0, maximum: 20)

Video continuity factor (flow).

Default: 9

num_frames

integer

(minimum: 1, maximum: 1440)

How many frames (duration) in the resulting video.

Default: 33

width

integer

(minimum: 64, maximum: 1536)

Width for the generated video.

Default: 640

height

integer

(minimum: 64, maximum: 1024)

Height for the generated video.

Default: 360

denoise_strength

number

(minimum: 0, maximum: 2)

Controls how strongly noise is applied each step.

Default: 1

force_offload

boolean

Whether to force model layers offloaded to CPU.

Default: true

frame_rate

integer

(minimum: 1, maximum: 60)

Video frame rate.

Default: 16

crf

integer

(minimum: 0, maximum: 51)

CRF (quality) for H264 encoding. Lower values = higher quality.

Default: 19

enhance_weight

number

(minimum: 0, maximum: 2)

Strength of the video enhancement effect.

Default: 0.3

enhance_single

boolean

Apply enhancement to individual frames.

Default: true

enhance_double

boolean

Apply enhancement across frame pairs.

Default: true

enhance_start

number

(minimum: 0, maximum: 1)

When to start enhancement in the video. Must be less than enhance_end.

Default: 0

enhance_end

number

(minimum: 0, maximum: 1)

When to end enhancement in the video. Must be greater than enhance_start.

Default: 1

seed

integer

Set a seed for reproducibility. Random by default.

Run this model in Node.js with one line of code:

npx create-replicate --model=lucataco/hunyuan-heygen-joshua

or set up a project from scratch

Install Replicate’s Node.js client library

npm install replicate

Set the REPLICATE_API_TOKEN environment variable

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run lucataco/hunyuan-heygen-joshua using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "lucataco/hunyuan-heygen-joshua:e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f",
  {
    input: {
      crf: 19,
      steps: 30,
      width: 960,
      height: 544,
      prompt: "HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee",
      lora_url: "",
      scheduler: "DPMSolverMultistepScheduler",
      flow_shift: 9,
      frame_rate: 15,
      num_frames: 33,
      enhance_end: 1,
      enhance_start: 0,
      force_offload: true,
      lora_strength: 0.8,
      enhance_double: true,
      enhance_single: true,
      enhance_weight: 0.3,
      guidance_scale: 6,
      denoise_strength: 1
    }
  }
);
console.log(output);

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library

pip install replicate

Set the REPLICATE_API_TOKEN environment variable

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client

import replicate

Run lucataco/hunyuan-heygen-joshua using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "lucataco/hunyuan-heygen-joshua:e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f",
    input={
        "crf": 19,
        "steps": 30,
        "width": 960,
        "height": 544,
        "prompt": "HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee",
        "lora_url": "",
        "scheduler": "DPMSolverMultistepScheduler",
        "flow_shift": 9,
        "frame_rate": 15,
        "num_frames": 33,
        "enhance_end": 1,
        "enhance_start": 0,
        "force_offload": True,
        "lora_strength": 0.8,
        "enhance_double": True,
        "enhance_single": True,
        "enhance_weight": 0.3,
        "guidance_scale": 6,
        "denoise_strength": 1
    }
)
print(output)

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run lucataco/hunyuan-heygen-joshua using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "version": "e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f",
    "input": {
      "crf": 19,
      "steps": 30,
      "width": 960,
      "height": 544,
      "prompt": "HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee",
      "lora_url": "",
      "scheduler": "DPMSolverMultistepScheduler",
      "flow_shift": 9,
      "frame_rate": 15,
      "num_frames": 33,
      "enhance_end": 1,
      "enhance_start": 0,
      "force_offload": true,
      "lora_strength": 0.8,
      "enhance_double": true,
      "enhance_single": true,
      "enhance_weight": 0.3,
      "guidance_scale": 6,
      "denoise_strength": 1
    }
  }' \
  https://api.replicate.com/v1/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Install Cog

brew install cog

If you don’t have Homebrew, there are other installation options available.

Pull and run lucataco/hunyuan-heygen-joshua using Cog (this will download the full model and run it in your local environment):

cog predict r8.im/lucataco/hunyuan-heygen-joshua@sha256:e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f \
  -i 'crf=19' \
  -i 'steps=30' \
  -i 'width=960' \
  -i 'height=544' \
  -i 'prompt="HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee"' \
  -i 'lora_url=""' \
  -i 'scheduler="DPMSolverMultistepScheduler"' \
  -i 'flow_shift=9' \
  -i 'frame_rate=15' \
  -i 'num_frames=33' \
  -i 'enhance_end=1' \
  -i 'enhance_start=0' \
  -i 'force_offload=true' \
  -i 'lora_strength=0.8' \
  -i 'enhance_double=true' \
  -i 'enhance_single=true' \
  -i 'enhance_weight=0.3' \
  -i 'guidance_scale=6' \
  -i 'denoise_strength=1'

To learn more, take a look at the Cog documentation.

Pull and run lucataco/hunyuan-heygen-joshua using Docker (this will download the full model and run it in your local environment):

docker run -d -p 5000:5000 --gpus=all r8.im/lucataco/hunyuan-heygen-joshua@sha256:e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "crf": 19,
      "steps": 30,
      "width": 960,
      "height": 544,
      "prompt": "HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee",
      "lora_url": "",
      "scheduler": "DPMSolverMultistepScheduler",
      "flow_shift": 9,
      "frame_rate": 15,
      "num_frames": 33,
      "enhance_end": 1,
      "enhance_start": 0,
      "force_offload": true,
      "lora_strength": 0.8,
      "enhance_double": true,
      "enhance_single": true,
      "enhance_weight": 0.3,
      "guidance_scale": 6,
      "denoise_strength": 1
    }
  }' \
  http://localhost:5000/predictions

Output

{
  "completed_at": "2025-01-09T21:24:57.183687Z",
  "created_at": "2025-01-09T21:22:55.843000Z",
  "data_removed": false,
  "error": null,
  "id": "3kve7qfawdrme0cm9ks83heydr",
  "input": {
    "crf": 19,
    "steps": 30,
    "width": 960,
    "height": 544,
    "prompt": "HGM1 man standing indoors, likely in a modern, well-lit space. He is wearing a white, long-sleeved shirt. The man is drinking coffee",
    "lora_url": "",
    "flow_shift": 9,
    "frame_rate": 15,
    "num_frames": 33,
    "force_offload": true,
    "lora_strength": 0.8,
    "guidance_scale": 6,
    "denoise_strength": 1
  },
  "logs": "Random seed set to: 3931337327\nChecking inputs\n====================================\nChecking weights\n✅ hunyuan_video_720_fp8_e4m3fn.safetensors exists in ComfyUI/models/diffusion_models\n✅ hunyuan_video_vae_bf16.safetensors exists in ComfyUI/models/vae\n====================================\nRunning workflow\n[ComfyUI] got prompt\nExecuting node 30, title: HunyuanVideo TextEncode, class type: HyVideoTextEncode\n[ComfyUI] llm prompt attention_mask shape: torch.Size([1, 161]), masked tokens: 35\n[ComfyUI] clipL prompt attention_mask shape: torch.Size([1, 77]), masked tokens: 36\n[ComfyUI] Input (height, width, video_length) = (544, 960, 33)\nExecuting node 3, title: HunyuanVideo Sampler, class type: HyVideoSampler\n[ComfyUI] Sampling 33 frames in 9 latents at 960x544 with 30 inference steps\n[ComfyUI]\n[ComfyUI] 0%|          | 0/30 [00:00<?, ?it/s]\n[ComfyUI] 3%|▎         | 1/30 [00:02<01:02,  2.14s/it]\n[ComfyUI] 7%|▋         | 2/30 [00:05<01:13,  2.62s/it]\n[ComfyUI] 10%|█         | 3/30 [00:08<01:14,  2.77s/it]\n[ComfyUI] 13%|█▎        | 4/30 [00:10<01:13,  2.84s/it]\n[ComfyUI] 17%|█▋        | 5/30 [00:13<01:11,  2.88s/it]\n[ComfyUI] 20%|██        | 6/30 [00:16<01:09,  2.90s/it]\n[ComfyUI] 23%|██▎       | 7/30 [00:19<01:06,  2.91s/it]\n[ComfyUI] 27%|██▋       | 8/30 [00:22<01:04,  2.92s/it]\n[ComfyUI] 30%|███       | 9/30 [00:25<01:01,  2.93s/it]\n[ComfyUI] 33%|███▎      | 10/30 [00:28<00:58,  2.93s/it]\n[ComfyUI] 37%|███▋      | 11/30 [00:31<00:55,  2.93s/it]\n[ComfyUI] 40%|████      | 12/30 [00:34<00:52,  2.94s/it]\n[ComfyUI] 43%|████▎     | 13/30 [00:37<00:49,  2.94s/it]\n[ComfyUI] 47%|████▋     | 14/30 [00:40<00:47,  2.94s/it]\n[ComfyUI] 50%|█████     | 15/30 [00:43<00:44,  2.94s/it]\n[ComfyUI] 53%|█████▎    | 16/30 [00:46<00:41,  2.94s/it]\n[ComfyUI] 57%|█████▋    | 17/30 [00:49<00:38,  2.94s/it]\n[ComfyUI] 60%|██████    | 18/30 [00:52<00:35,  2.94s/it]\n[ComfyUI] 63%|██████▎   | 19/30 [00:55<00:32,  2.94s/it]\n[ComfyUI] 67%|██████▋   | 20/30 [00:58<00:29,  2.94s/it]\n[ComfyUI] 70%|███████   | 21/30 [01:00<00:26,  2.94s/it]\n[ComfyUI] 73%|███████▎  | 22/30 [01:03<00:23,  2.94s/it]\n[ComfyUI] 77%|███████▋  | 23/30 [01:06<00:20,  2.94s/it]\n[ComfyUI] 80%|████████  | 24/30 [01:09<00:17,  2.94s/it]\n[ComfyUI] 83%|████████▎ | 25/30 [01:12<00:14,  2.94s/it]\n[ComfyUI] 87%|████████▋ | 26/30 [01:15<00:11,  2.94s/it]\n[ComfyUI] 90%|█████████ | 27/30 [01:18<00:08,  2.94s/it]\n[ComfyUI] 93%|█████████▎| 28/30 [01:21<00:05,  2.94s/it]\n[ComfyUI] 97%|█████████▋| 29/30 [01:24<00:02,  2.94s/it]\n[ComfyUI] 100%|██████████| 30/30 [01:27<00:00,  2.94s/it]\n[ComfyUI] 100%|██████████| 30/30 [01:27<00:00,  2.91s/it]\n[ComfyUI] Allocated memory: memory=12.299 GB\n[ComfyUI] Max allocated memory: max_memory=15.815 GB\n[ComfyUI] Max reserved memory: max_reserved=17.625 GB\nExecuting node 5, title: HunyuanVideo Decode, class type: HyVideoDecode\n[ComfyUI]\n[ComfyUI] Decoding rows:   0%|          | 0/3 [00:00<?, ?it/s]\n[ComfyUI] Decoding rows:  33%|███▎      | 1/3 [00:00<00:01,  1.00it/s]\n[ComfyUI] Decoding rows:  67%|██████▋   | 2/3 [00:02<00:01,  1.08s/it]\n[ComfyUI] Decoding rows: 100%|██████████| 3/3 [00:02<00:00,  1.07it/s]\n[ComfyUI] Decoding rows: 100%|██████████| 3/3 [00:02<00:00,  1.03it/s]\n[ComfyUI]\n[ComfyUI] Blending tiles:   0%|          | 0/3 [00:00<?, ?it/s]\n[ComfyUI] Blending tiles:  67%|██████▋   | 2/3 [00:00<00:00, 18.25it/s]\nExecuting node 34, title: Video Combine 🎥🅥🅗🅢, class type: VHS_VideoCombine\n[ComfyUI] Blending tiles: 100%|██████████| 3/3 [00:00<00:00, 23.35it/s]\n[ComfyUI] Prompt executed in 105.92 seconds\noutputs:  {'34': {'gifs': [{'filename': 'HunyuanVideo_00001.mp4', 'subfolder': '', 'type': 'output', 'format': 'video/h264-mp4', 'frame_rate': 15.0, 'workflow': 'HunyuanVideo_00001.png', 'fullpath': '/tmp/outputs/HunyuanVideo_00001.mp4'}]}}\n====================================\nHunyuanVideo_00001.png\nHunyuanVideo_00001.mp4",
  "metrics": {
    "predict_time": 117.059010205,
    "total_time": 121.340687
  },
  "output": "https://replicate.delivery/xezq/smD48Q5XpoJJA9EIZpi5KPX90GLO8g16BjHdzsMqEFVqw3AF/HunyuanVideo_00001.mp4",
  "started_at": "2025-01-09T21:23:00.124677Z",
  "status": "succeeded",
  "urls": {
    "stream": "https://stream.replicate.com/v1/files/bcwr-xhpsx4vcf4gabwnwtqzr3hq3chmx4hshusskz2izhsbtker5fp7q",
    "get": "https://api.replicate.com/v1/predictions/3kve7qfawdrme0cm9ks83heydr",
    "cancel": "https://api.replicate.com/v1/predictions/3kve7qfawdrme0cm9ks83heydr/cancel"
  },
  "version": "e82a415498ba81104f2c1f41833609e546cd53d2734b37740d685b122c93053f"
}

Generated in

1 minute 57 seconds

Tweak it ShareReport

Random seed set to: 3931337327
Checking inputs
====================================
Checking weights
✅ hunyuan_video_720_fp8_e4m3fn.safetensors exists in ComfyUI/models/diffusion_models
✅ hunyuan_video_vae_bf16.safetensors exists in ComfyUI/models/vae
====================================
Running workflow
[ComfyUI] got prompt
Executing node 30, title: HunyuanVideo TextEncode, class type: HyVideoTextEncode
[ComfyUI] llm prompt attention_mask shape: torch.Size([1, 161]), masked tokens: 35
[ComfyUI] clipL prompt attention_mask shape: torch.Size([1, 77]), masked tokens: 36
[ComfyUI] Input (height, width, video_length) = (544, 960, 33)
Executing node 3, title: HunyuanVideo Sampler, class type: HyVideoSampler
[ComfyUI] Sampling 33 frames in 9 latents at 960x544 with 30 inference steps
[ComfyUI]
[ComfyUI] 0%|          | 0/30 [00:00<?, ?it/s]
[ComfyUI] 3%|▎         | 1/30 [00:02<01:02,  2.14s/it]
[ComfyUI] 7%|▋         | 2/30 [00:05<01:13,  2.62s/it]
[ComfyUI] 10%|█         | 3/30 [00:08<01:14,  2.77s/it]
[ComfyUI] 13%|█▎        | 4/30 [00:10<01:13,  2.84s/it]
[ComfyUI] 17%|█▋        | 5/30 [00:13<01:11,  2.88s/it]
[ComfyUI] 20%|██        | 6/30 [00:16<01:09,  2.90s/it]
[ComfyUI] 23%|██▎       | 7/30 [00:19<01:06,  2.91s/it]
[ComfyUI] 27%|██▋       | 8/30 [00:22<01:04,  2.92s/it]
[ComfyUI] 30%|███       | 9/30 [00:25<01:01,  2.93s/it]
[ComfyUI] 33%|███▎      | 10/30 [00:28<00:58,  2.93s/it]
[ComfyUI] 37%|███▋      | 11/30 [00:31<00:55,  2.93s/it]
[ComfyUI] 40%|████      | 12/30 [00:34<00:52,  2.94s/it]
[ComfyUI] 43%|████▎     | 13/30 [00:37<00:49,  2.94s/it]
[ComfyUI] 47%|████▋     | 14/30 [00:40<00:47,  2.94s/it]
[ComfyUI] 50%|█████     | 15/30 [00:43<00:44,  2.94s/it]
[ComfyUI] 53%|█████▎    | 16/30 [00:46<00:41,  2.94s/it]
[ComfyUI] 57%|█████▋    | 17/30 [00:49<00:38,  2.94s/it]
[ComfyUI] 60%|██████    | 18/30 [00:52<00:35,  2.94s/it]
[ComfyUI] 63%|██████▎   | 19/30 [00:55<00:32,  2.94s/it]
[ComfyUI] 67%|██████▋   | 20/30 [00:58<00:29,  2.94s/it]
[ComfyUI] 70%|███████   | 21/30 [01:00<00:26,  2.94s/it]
[ComfyUI] 73%|███████▎  | 22/30 [01:03<00:23,  2.94s/it]
[ComfyUI] 77%|███████▋  | 23/30 [01:06<00:20,  2.94s/it]
[ComfyUI] 80%|████████  | 24/30 [01:09<00:17,  2.94s/it]
[ComfyUI] 83%|████████▎ | 25/30 [01:12<00:14,  2.94s/it]
[ComfyUI] 87%|████████▋ | 26/30 [01:15<00:11,  2.94s/it]
[ComfyUI] 90%|█████████ | 27/30 [01:18<00:08,  2.94s/it]
[ComfyUI] 93%|█████████▎| 28/30 [01:21<00:05,  2.94s/it]
[ComfyUI] 97%|█████████▋| 29/30 [01:24<00:02,  2.94s/it]
[ComfyUI] 100%|██████████| 30/30 [01:27<00:00,  2.94s/it]
[ComfyUI] 100%|██████████| 30/30 [01:27<00:00,  2.91s/it]
[ComfyUI] Allocated memory: memory=12.299 GB
[ComfyUI] Max allocated memory: max_memory=15.815 GB
[ComfyUI] Max reserved memory: max_reserved=17.625 GB
Executing node 5, title: HunyuanVideo Decode, class type: HyVideoDecode
[ComfyUI]
[ComfyUI] Decoding rows:   0%|          | 0/3 [00:00<?, ?it/s]
[ComfyUI] Decoding rows:  33%|███▎      | 1/3 [00:00<00:01,  1.00it/s]
[ComfyUI] Decoding rows:  67%|██████▋   | 2/3 [00:02<00:01,  1.08s/it]
[ComfyUI] Decoding rows: 100%|██████████| 3/3 [00:02<00:00,  1.07it/s]
[ComfyUI] Decoding rows: 100%|██████████| 3/3 [00:02<00:00,  1.03it/s]
[ComfyUI]
[ComfyUI] Blending tiles:   0%|          | 0/3 [00:00<?, ?it/s]
[ComfyUI] Blending tiles:  67%|██████▋   | 2/3 [00:00<00:00, 18.25it/s]
Executing node 34, title: Video Combine 🎥🅥🅗🅢, class type: VHS_VideoCombine
[ComfyUI] Blending tiles: 100%|██████████| 3/3 [00:00<00:00, 23.35it/s]
[ComfyUI] Prompt executed in 105.92 seconds
outputs:  {'34': {'gifs': [{'filename': 'HunyuanVideo_00001.mp4', 'subfolder': '', 'type': 'output', 'format': 'video/h264-mp4', 'frame_rate': 15.0, 'workflow': 'HunyuanVideo_00001.png', 'fullpath': '/tmp/outputs/HunyuanVideo_00001.mp4'}]}}
====================================
HunyuanVideo_00001.png
HunyuanVideo_00001.mp4