bytedance-research/Lance · Hugging Face

Source: https://huggingface.co/bytedance-research/Lance

Lance: Unified Multimodal Modeling by Multi-Task Synergy

Fengyi Fu*,Mengqi Huang*,✉,Shaojin Wu*,Yunsheng Jiang*,Yufei Huo,Jianzhu Guo✉,§ Hao Li, Yinghang Song, Fei Ding, Qian He, Zheren Fu, Zhendong Mao, Yongdong Zhang ByteDance *Equal contribution✉Corresponding authors§Project lead

English |简体中文

https://huggingface.co/bytedance-research/Lance#%F0%9F%8C%9F-highlights🌟 Highlights

Lance is a lightweight native unified multimodal model that supportsimage and video understanding, generation, and editingwithin a single framework.

• Efficient at 3B scale.With only3B active parameters, Lance delivers strong performance across image generation, image editing, and video generation benchmarks.
• Trained from scratch.Lance is built with a staged multi-task recipe and trained entirely from scratch within a128-A100-GPUbudget.

https://huggingface.co/bytedance-research/Lance#%F0%9F%8E%A8-demo🎨 Demo

https://huggingface.co/bytedance-research/Lance#text-to-videoText-to-Video

https://huggingface.co/bytedance-research/Lance#video-editingVideo Editing

https://huggingface.co/bytedance-research/Lance#multi-turn-consistency-editingMulti-turn Consistency Editing

https://huggingface.co/bytedance-research/Lance#intelligent-video-generationIntelligent Video Generation

https://huggingface.co/bytedance-research/Lance#video-understandingVideo Understanding

**Question:**How many times did the person launch objects on the table? Options: (A) 3 (B) 2 (C) 4

Response:(A) 3

**Question:**The person makes sets of repeated actions. How many distinct repeated actions did the person do? Options: (A) 2 (B) 3 (C) 4

Response:(A) 2

**Question:**In which direction does the purple sphere move in the video? Options: (A) Down and to the right. (B) Up and to the left. (C) Up and to the right. (D) The object is stationary.

Response:(A) Down and to the right.

**Question:**What is the unrealistic phenomenon displayed in the video? Options: (A) The man can manipulate time via phone. (B) Man grabs an object through a phone screen. (C) Chocolate transforms into different objects. (D) Visible means of propulsion enables flight.

Response:(B) Man grabs an object through a phone screen.

**Question:**Offer a succinct account of the culinary process shown in this video.

**Response:**Add tomato puree and mix it well with chicken pieces.

**Question:**Provide a detailed description of the given video, capturing its key moments.

**Response:**In a sunlit meadow, a small tortoiseshell butterfly rests on a purple flower. A bee, with black and yellow stripes, lands on the same flower. The butterfly flaps its wings gently, while the bee busies itself, collecting nectar. The flower sways slightly in the breeze. The butterfly then takes off, and the bee follows, both heading to the next flower. The scene is a vivid display of insect interaction in a natural setting, with the colors of the butterfly and the bee contrasting against the green background of the meadow. The video captures this peaceful moment in a short 6-second duration.

https://huggingface.co/bytedance-research/Lance#text-to-image-generationText-to-Image Generation

https://huggingface.co/bytedance-research/Lance#image-editingImage Editing

https://huggingface.co/bytedance-research/Lance#image-understandingImage Understanding

**Question:**Is the largest segment greater than sum of all the other segments?

**Response:**Yes, the largest segment in the pie chart is the blue one, which is the most populous segment.

**Question:**What percentage of respondents want better border security?

**Response:**29%

**Question:**What is the license plate number of the car?

**Response:**The license plate number of the car is BX62 BFY.

**Question:**According to the data from the proprietary market research, how much amount was spent on the promotional meetings and events during 1998?

**Response:**According to the data from the proprietary market research, the total amount spent on the promotional meetings and events during 1998 was approximately $1.3 billion.

**Question:**What is the appearance of the Colosseum in Rome, Italy?

**Response:**The Colosseum in Rome, Italy is a magnificent ancient amphitheater made of stone and concrete. It is characterized by its large, arched openings and its semi-circular shape. The Colosseum has been partially restored, and its original color has faded, but the architectural details, like the columns and the semi-circular arches, are still visible. The lighting in the image highlights the grandeur and historical significance of the Colosseum, making it stand out against the backdrop of the evening sky.

**Question:**How does a total solar eclipse look like from Earth?

**Response:**A total solar eclipse looks like a dark circle with a bright, white outer edge. This is a result of the Earth’s rotation and the Earth’s shadow, which partially obscures the sun’s light. The bright white edge of the eclipse is caused by the sun’s high energy and its interaction with the Earth’s atmosphere, while the dark part of the eclipse is due to the Earth’s shadow and the surrounding air currents. The solar eclipse’s shape, with its bright white edge and dark center, is similar to the shape of a full moon or a dark disk. It is a natural phenomenon that occurs in the atmosphere of the Earth and is an important part of the solar system.

https://huggingface.co/bytedance-research/Lance#%F0%9F%9A%80-installation🚀 Installation

https://huggingface.co/bytedance-research/Lance#recommended-environmentRecommended Environment

• **Software:**Python 3.10+, CUDA 12.4+ (required)
• **Hardware:**A GPU with at least 40GB VRAM is required for inference

https://huggingface.co/bytedance-research/Lance#installation-stepsInstallation Steps

bash ./setup_env.sh

https://huggingface.co/bytedance-research/Lance#download-model-weightsDownload Model Weights

Please download all the necessary model checkpoints ofLance-3B (Huggingface Link)and place them in thedownloads/directory.

https://huggingface.co/bytedance-research/Lance#%F0%9F%93%9A-usage📚 Usage

https://huggingface.co/bytedance-research/Lance#inferenceInference

Lance provides a unified command-line interface for all generation / editing / understanding tasks:

bash inference_lance.sh

• Before running, please configure the inference parameters at the top ofinference\_lance\.sh.
• Supported tasks:t2i,t2v,image\_edit,video\_edit,x2t\_image, andx2t\_video. You can modifyTASK\_DEFAULT\_CONFIGSininference\_lance\.pyto customize the default data samples for each task.

https://huggingface.co/bytedance-research/Lance#available-tasksAvailable Tasks

Task NameDescriptionExample JSONt2vText-to-Video generationconfig/examples/t2v\_example\.json``t2iText-to-Image generationconfig/examples/t2i\_example\.json``image\_editImage editingconfig/examples/image\_edit\_example\.json``video\_editVideo editingconfig/examples/video\_edit\_example\.json``x2t\_imageImage understandingconfig/examples/x2t\_image\_example\.json``x2t\_videoVideo understandingconfig/examples/x2t\_video\_example\.json For understanding examples:

• config/examples/x2t\_image\_example\.json: image understanding examples for visual question answering and image-based reasoning.
• config/examples/x2t\_video\_example\.json: video understanding examples for video question answering and video captioning.

https://huggingface.co/bytedance-research/Lance#parametersParameters

You can configure the following hyperparameters at the top of theinference\_lance\.shscript:

ParameterDefault ValueDescriptionMODEL\_PATH``"downloads/lance\_3b"Path to the downloaded Lance model weights.NUM\_GPUS``1Number of GPUs to use for inference.VALIDATION\_NUM\_TIMESTEPS``30Number of denoising steps (e.g., 30 or 50).VALIDATION\_TIMESTEP\_SHIFT``3\.5Timestep shift parameter for flow matching scheduling.CFG\_TEXT\_SCALE``4\.0Classifier-Free Guidance (CFG) scale for text conditioning.VALIDATION\_DATA\_SEED``42Random seed for generation reproducibility.NUM\_FRAMES``50Number of frames for video generation (Max: 121).Unused for image tasks.VIDEO\_HEIGHT/VIDEO\_WIDTH``768Spatial resolution.Unused for editing tasks (determined by input image/video).RESOLUTION``"video\_480p"Base resolution preset (image\_768resorvideo\_480p).

https://huggingface.co/bytedance-research/Lance#gradioGradio

python lance_gradio_t2v_v2t.py --gpus 0 --server-port 7860

https://huggingface.co/bytedance-research/Lance#benchmarksBenchmarks

https://huggingface.co/bytedance-research/Lance#dpg-bench-evaluationDPG-Bench Evaluation

Models# Params.GlobalEntityAttributeRelationOtherOverallGeneration-only ModelsSDXL3.5B83.2782.4380.9186.7680.4174.65DALL-E 3-90.9789.6188.3990.5889.8383.50SD3-Medium2B87.9091.0188.8380.7088.6884.08FLUX.1-dev12B74.3590.0088.9690.8788.3383.84Qwen-Image20B91.3291.5692.0294.3192.7388.32Unified ModelsJanus-Pro-7B7B86.9088.9089.4089.3289.4884.19OmniGen24B88.8188.8390.1889.3790.2783.57Show-o27B89.0091.7889.9691.8191.6486.14BAGEL†7B88.9490.3791.2990.8288.6785.07InternVL-U1.7B90.3990.7890.6890.2988.7785.18TUNA7B90.4291.6890.9491.8790.7386.76TUNA-27B89.5091.4092.0791.9188.8186.54🌟Lance (Ours)3B83.8991.0789.3693.3880.80****84.67 †indicates methods that use LLM rewriters for prompt rewriting before generation.

https://huggingface.co/bytedance-research/Lance#geneval-evaluationGenEval Evaluation

Models# Params.1-Obj.2-Obj.CountColorsPositionAttr.OverallGeneration-only ModelsSDXL3.5B0.980.740.390.850.150.230.55DALL-E 3-0.960.870.470.830.430.450.67SD3-Medium2B0.990.940.720.890.330.600.74FLUX.1-dev12B0.980.930.750.930.680.650.82Qwen-Image20B0.990.920.890.880.760.770.87Unified ModelsJanus-Pro-7B7B0.990.890.590.900.790.660.80OmniGen24B1.000.950.640.880.550.760.80Show-o27B1.000.870.580.920.520.620.76BAGEL†7B0.980.950.840.950.780.770.88Mogao7B1.00****0.970.830.930.840.800.89InternVL-U1.7B0.990.940.740.910.770.740.85TUNA7B1.00****0.970.810.910.880.830.90TUNA-27B0.990.960.800.910.840.760.87🌟Lance (Ours)3B1.000.940.840.970.870.810.90 †indicates methods that use LLM rewriters for prompt rewriting before generation.

https://huggingface.co/bytedance-research/Lance#gedit-bench-evaluationGEdit-Bench Evaluation

Models# Params.BCCAMMMCPBSTSASRSRpTMTTAvg/G_OGeneration-only ModelsGemini 2.0------------6.32GPT Image 1-6.966.857.105.416.747.447.518.738.558.458.697.49Qwen-Image-Edit20B8.238.307.338.057.496.748.578.098.298.488.508.01Unified ModelsLumina-DiMOO8B3.434.273.082.774.745.194.443.804.382.684.203.91Ovis-U11.2B7.496.886.214.795.986.467.497.257.274.486.316.42BAGEL7B7.326.916.384.754.576.157.907.167.027.326.226.52InternVL-U1.7B7.087.056.387.026.036.277.136.556.336.596.856.66InternVL-U (w/ CoT)1.7B7.057.876.506.995.776.107.337.167.127.366.466.88🌟Lance (Ours)3B7.737.747.287.837.50****7.037.647.85****7.714.467.57****7.30

https://huggingface.co/bytedance-research/Lance#vbench-evaluation-video-generationVBench Evaluation (Video Generation)

TypeModel# Params.Total Score ↑Gen. OnlyModelScope1.7B75.75LaVie3B77.08Show-16B78.93AnimateDiff-V2-80.27VideoCrafter-2.0-80.44CogVideoX5B81.61Kling-81.85Open-Sora-2.0-81.71Gen-3-82.32Step-Video-T2V30B81.83Hunyuan Video-83.43Wan2.1-T2V14B83.69UnifiedHaproOmni7B78.10Emu38B80.96VILA-U7B74.01Show-o22B81.34TUNA1.5B84.06🌟Lance (Ours)3B85.11

https://huggingface.co/bytedance-research/Lance#running-benchmarksRunning Benchmarks

Ready-to-run benchmark scripts are provided underbenchmarks/:

BenchmarkModalityScriptGenEVAL (image gen)Imagebenchmarks/image\_gen/GenEVAL/sample\_GenEVAL\.shDPG (image gen)Imagebenchmarks/image\_gen/DPG/sample\_DPG\.shGEdit (image edit)Imagebenchmarks/image\_gen/GEdit/sample\_GEdit\.shVBench (video gen)Videobenchmarks/video\_gen/Vbench/sample\_vbench\.sh

https://huggingface.co/bytedance-research/Lance#%F0%9F%93%84-license📄 License

Copyright 2025 Bytedance Ltd. and/or its affiliates.

https://huggingface.co/bytedance-research/Lance#%F0%9F%92%96-citation💖 Citation

If you findLanceuseful for your project or research, welcome to 🌟 this repo and cite our work using the following BibTeX:

@misc{lance2026,
  title  = {Lance: Unified Multimodal Modeling by Multi-Task Synergy},
  author = {Fengyi Fu and Mengqi Huang and Shaojin Wu and Yunsheng Jiang and Yufei Huo and Jianzhu Guo and Hao Li and Yinghang Song and Fei Ding and Qian He and Zheren Fu and Zhendong Mao and Yongdong Zhang},
  year   = {2026},
  note   = {Manuscript}
}

https://huggingface.co/bytedance-research/Lance#%F0%9F%93%9E-contact📞 Contact

For questions, issues, or collaborations, please contactMengqi HuangandJianzhu Guo.