Cosmos-UpsamplePrompt1-12B-Transfer

Cosmos | Code | Paper | Paper Website

Description:

Cosmos-UpsamplePrompt1-12B-Transfer is a multimodal model designed to transform original input prompts into more detailed and enriched versions based on the control video. It improves the prompts by adding more details and maintaining a consistent description structure before they are used in a conditional world generation model, which generally leads to higher quality outputs. This model is ready for commercial use.

License:

GOVERNING TERMS: Use of this model is governed by the NVIDIA Open Model License. Additional Information: Apache License Version 2.0.

For a custom license, please contact [email protected].

Under the NVIDIA Open Model License, NVIDIA confirms:

Models are commercially usable.
You are free to create and distribute Derivative Models.
NVIDIA does not claim ownership to any outputs generated using the Models or Derivative Models.

Important Note: If you bypass, disable, reduce the efficacy of, or circumvent any technical limitation, safety guardrail or associated safety guardrail hyperparameter, encryption, security, digital rights management, or authentication mechanism contained in the Model, your rights under NVIDIA Open Model License Agreement will automatically terminate.

Deployment Geography:

Global

Use Case:

It improves the prompts by adding more details and maintaining a consistent description structure before they are used in a conditional world generation model, which generally leads to higher quality outputs. Intended for research and development purposes.

Release Date:

Hugging Face 03/18/2025 via https://huggingface.co/nvidia/Cosmos-UpsamplePrompt1-12B-Transfer

Model Architecture:

Architecture Type: Transformer
Network Architecture: Pixtral 12B

Input:

Input Type(s): Text+Video
Input Format: Text: String, Video: mp4
Input Parameters: Text: One-Dimensional (1D); Video: Three-Dimensional (3D)

Output:

Output Type(s): Text
Output Format: String
Output Parameters: Text: One-dimensional (1D)

Software Integration:

Runtime Engine(s):

Cosmos-Transfer1

Supported Hardware Microarchitecture Compatibility:

NVIDIA Ampere
NVIDIA Hopper

Supported Operating System(s): Linux

Model Version:

The initial release (v1.0) of Cosmos Prompt Upsampler contains the following model:

Cosmos-UpsamplePrompt1-12B-Transfer

Usage

See Cosmos-Transfer1 for on how to use the model.

Example:

Input: "A robot in the kitchen picks up a bottle from the floor and puts it on a table."

condition video

Output: "The video features a kitchen with wooden cabinets and a granite countertop. A robot with a white body, black joints, and a red light on its head is seen performing tasks. It moves its arms and legs to pick up a white bottle with a red label from the floor and place it on the countertop. The robot then moves to a dining area with a wooden table and chairs, where it picks up a white chair and places it back in its original position."

Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

For more detailed information on ethical considerations for this model, please see the subcards of Explainability, Bias, Safety & Security, and Privacy below. Please report security vulnerabilities or NVIDIA AI Concerns here.

Plus Plus (++) Promise

We value you, the datasets, the diversity they represent, and what we have been entrusted with. This model and its associated data have been:

Verified to comply with current applicable disclosure laws, regulations, and industry standards.
Verified to comply with applicable privacy labeling requirements.
Annotated to describe the collector/source (NVIDIA or a third-party).
Characterized for technical limitations.
Reviewed to ensure proper disclosure is accessible to, maintained for, and in compliance with NVIDIA data subjects and their requests.
Reviewed before release.
Tagged for known restrictions and potential safety implications.

Bias

Field	Response
Participation considerations from adversely impacted groups protected classes in model design and testing:	None
Measures taken to mitigate against unwanted bias:	None

Explainability

Field	Response
Intended Application & Domain:	Text prompt enrichment for world generation
Model Type:	Transformer
Intended Users:	Generative AI developers for world generation models
Output:	Text
Describe how the model works:	Transform original input prompts into more detailed and enriched versions
Technical Limitations:	The model may not follow input text prompt accurately and may have incorrect responses or generate incorrect information.
Verified to have met prescribed NVIDIA quality standards:	Yes
Performance Metrics:	Human Evaluation
Potential Known Risks:	The model's output can generate all forms of text, including what may be considered toxic, offensive, or indecent.
Licensing:	Abide by NVIDIA Open Model License. Additional Information: Apache License Version 2.0.

Privacy

Field	Response
Generatable or reverse engineerable personal information?	None Known
Protected class data used to create this model?	None Known
Was consent obtained for any personal data used?	None Known
How often is dataset reviewed?	Before Release
Is there provenance for all datasets used in training?	Yes
Does data labeling (annotation, metadata) comply with privacy laws?	Yes

Safety

Field	Response
Model Application(s):	Prompt enrichment for world generation
Describe the life critical impact (if present).	None Known
Use Case Restrictions:	Abide by NVIDIA Open Model License. Additional Information: Apache License Version 2.0.
Model and dataset restrictions:	The Principle of least privilege (PoLP) is applied limiting access for dataset generation and model development. Restrictions enforce dataset access during training, and dataset license constraints adhered to. Model checkpoints are made available on Hugging Face, and may become available on cloud providers' model catalog.