nvidia/Cosmos-Tokenize1-DV8x16x16-720p

Cosmos-Tokenize1: A suite of image and video tokenizers

Cosmos | Code | Paper | Paper Website

Model Overview

Description:

Cosmos-Tokenize1 is a suite of visual tokenizers for images and videos that delivers various compression rates while maintaining high reconstruction quality. Cosmos Tokenizer can serve as an effective and efficient building block in both diffusion-based and autoregressive models for image and video generation. This model is ready for commercial use.

Our tokenizers come in two types: Continuous (C) and Discrete (D), each with Image (I) and Video (V) variants:

• Continuous tokenizers encode visual data into continuous latent embeddings, as shown in latent diffusion models like Stable Diffusion. These embeddings are suitable for models that generate data by sampling from continuous distributions.
• Discrete tokenizers encode visual data into discrete latent codes, mapping them into quantized indices, as seen in autoregressive transformers such as VideoPoet. This discretization is required for models that generate data by optimizing the cross-entropy loss, such as the GPT models.

	Continuous ( C )	Discrete ( D )
Images ( I )	Cosmos-Tokenizer-CI	Cosmos-Tokenizer-DI
Videos ( V )	Cosmos-Tokenizer-CV	Cosmos-Tokenizer-DV

Model: Cosmos-Tokenize1-DV8x16x16-720p, a discrete video tokenizer with 8x temporal compression, and 16x16 spatial compression rate.

Model Developer: NVIDIA

Model Versions

The Cosmos-Tokenize1 includes the following tokenizers:

• Continuous Tokenizers

  • Cosmos-Tokenize1-CI8x8-360p (8x8 spatial compression, 360p and above)
  • Cosmos-Tokenize1-CI16x16-360p (16x16 spatial compression, 360p and above)
  • Cosmos-Tokenize1-CV4x8x8-360p (4x temporal compression, 8x8 spatial compression, 360p and above, 49 frames context)
  • Cosmos-Tokenize1-CV8x8x8-720p (8x temporal compression, 8x8 spatial compression, 720p and above, 121 frames context)

• Discrete Tokenizers

  • Cosmos-Tokenize1-DI8x8-360p (8x8 spatial compression, 360p and above)
  • Cosmos-Tokenize1-DI16x16-360p (16x16 spatial compression, 360p and above)
  • Cosmos-Tokenize1-DV4x8x8-360p (4x temporal compression, 8x8 spatial compression, 360p and above, 49 frames context)
  • Cosmos-Tokenize1-DV8x16x16-720p (8x temporal compression, 16x16 spatial compression, 720p and above, 49 frames context)

License:

This model is released under the NVIDIA Open Model License. For a custom license, please contact cosmos-license@nvidia.com.

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.

Model Architecture:

Cosmos-Tokenize1-DV8x16x16-720p is a lightweight and computationally efficient architecture, featuring a temporally causal design. Specifically, we employ causal temporal convolution and causal temporal attention layers to preserve the natural temporal order of video frames, ensuring seamless tokenization of images and videos using a single unified network architecture. The encoder and decoder form a symmetrical pair, which are mirrors of each other. The encoder starts with a 2-level Haar wavelet transform layer, which down-samples inputs by a factor of 4 in both spatial and temporal dimensions. Likewise, the decoder ends with an inverse wavelet transform. We employ the vanilla autoencoder (AE) formulation to model the latent space for continuous tokenizers. For discrete tokenizers, we adopt the Finite-Scalar-Quantization (FSQ) as the latent space quantizer.

Input/Output Specifications

Encoder

• Input

  • Type: Images or Videos
  • Format: RGB (Red, Green, Blue)
  • Properties:
    • Resolution: Minimum: 256px (shorter side). Maximum: Up to 4K
    • Video Length: Up to 8 seconds for 1080p videos (bounded by A100 80G GPU memory; higher resolutions will have shorter supported durations)

• Output

  • Type: Tokens
  • Properties:
    • Integer indices ranging from 0 to 63,999

Decoder

• Input

  • Type: Tokens
  • Properties:
    • Integer indices ranging from 0 to 63,999

• Output

  • Type Images or Videos (matching input type)
  • Format: RGB (Red, Green, Blue)
  • Properties:
    • Resolution: Same as input resolution
    • Video Length: Same as input video length

Software Integration (Required For NVIDIA Models Only):

Runtime Engine(s):

• Cosmos-Predict1

Supported Hardware Microarchitecture Compatibility:

• NVIDIA Ampere (e.g., A100)
• NVIDIA Hopper (e.g., H100)

Note: We have only tested Cosmos Tokenizer with BF16 precision on Ampere and Hopper GPUs. If you are using older versions of NVIDIA GPUs (e.g., NVIDIA Volta GPUs), you may need to switch to FP32 precision.

Operating System(s):

• Linux (We have not tested on other operating systems.)

Usage

Inference Engines:

• See Cosmos-Predict1 (PyTorch) for details.

Evaluation

Tokenization Comparison

We have extensively evaluated the additional Cosmos Tokenizer models on DAVIS video benchmark dataset.

Tokenizer	Compression Ratio	Height	Num. of Frames	Quantization	PSNR (DAVIS)	SSIM (DAVIS)	rFVD (DAVIS)
VideoGPT	4×4×4	-	-	VQ	32.23	0.850	72.33
OmniTokenizer	4×8×8	-	-	VQ	28.44	0.712	188.60
Cosmos-Tokenizer-DV	4×8×8	720	17	FSQ	32.98	0.818	37.36
Cosmos-Tokenizer-DV	8×8×8	720	17	FSQ	32.11	0.775	100.15
Cosmos-Tokenizer-DV	8×16×16	720	17	FSQ	31.42	0.716	241.52
Cosmos-Tokenizer-DV	8×16×16	720	49	FSQ	31.59	0.719	259.33

• We compare with the state-of-the-art discrete video tokenizer, OmniTokenizer.
• Evaluation metrics:
  • Peak Signal-to-Noise Ratio (PSNR)
  • Structural Similarity (SSIM)
  • Reconstruction Fréchet Video Distance (rFVD)

Runtime Comparison

The following table shows the number of parameters and the averaged encoding and decoding times per image or video frame, measured on a single A100 80GB GPU. For comparison, we also list the parameters and average speeds of prior state-of-the-art tokenizer(s) with the same compression ratio.

Tokenizer	Resolution	Compression Ratio	Parameters	Time (ms)
OmniTokenizer	720x1280	4×8×8	54M	53.2
Cosmos-DV	720x1280	4×8×8	105M	51.5

Note: We benchmarked the runtime for images under the 8x8 compression and videos under the 4×8×8 compression. Tokenizers with different compression ratios are not included in this comparison.

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.

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.

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.

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:	Tokenization of images and videos
Model Type:	Auto-Encoder
Intended Users:	Generative AI developers for image and video generation models
Output:	Images/Videos and Latent Tokens
Describe how the model works:	Compresses and decompresses visual input (image/video).
Technical Limitations:	Due to tokenizer compression limitations, some visual information (such as small text and other structured fine details) may not be reconstructed accurately. The tokenizers may not produce as high of a reconstruction results for videos with low resolution, e.g. less than 320p.
Verified to have met prescribed NVIDIA quality standards:	Yes
Performance Metrics:	Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), Reconstruction Fréchet Video Distance (rFVD), Reconstruction Fréchet Inception Distance (rFID), Latency
Potential Known Risks:	Tokenizer's output can parse all forms of input, including what may be considered toxic, offensive, or indecent.
Licensing:	NVIDIA Open Model License

Privacy

Field	Response
Generatable or reverse engineerable personal information?	No
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):	Tokenization of images and videos
Describe the life critical impact (if present).	None Known
Use Case Restrictions:	NVIDIA Open Model License
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.