Submission Templates
Model Card Template - Metadata
### Model Card Template - Metadata
# How to use this YAML file:
# Fill in your information within the quotations ("") e.g. model_display_name: "ModelX"
# ~ or ~
# Add a info block for each item that you would like to add to a category
# e.g. authors:
# - name: "Jane Doe"
# type: "individual"
# affiliate: "Chan Zuckerberg Initiative"
# - name: "Chan Zuckerberg Initiative"
# type: "organization"
# ~ or ~
# Choose from a list provided by deleting the irrelevant items leaving only the list items you wish to select
# You can also add your own items to these lists by adding a list item with a dash
# e.g. tasks_performed_by_model:
# - Cell Clustering
# - Cell Labeling
# - Cross Species Integration
#. - Special Task Added To List
### Basic Info
model_display_name: ""
model_version: "" # vX.X.X or YYYY-MM-DD format
primary_contact_email: ""
repository_link: ""
publication_preprint_link: ""
release_date: "" # YYYY-MM-DD
model_domain: "" # one of "Imaging" or "Transcriptomic" or "Text"
model_description: ""
short_description: "" # Max 140 characters
authors: # List one or more individuals or organizations
- name: ""
type: "" # one of "individual" or "organization"
affiliate: "" # (optional) if it is an "individual", add an affiliated organization name
licenses: # Add one or more licenses
- type: "" # e.g "CC BY 4.0"
url: "" # e.g "https://creativecommons.org/licenses/by/4.0/deed.en"
compute_requirement: "" # one of "GPU" or "CPU"
system_requirement: "" # (optional) Any other system constraints or requirements you would like to describe.
### Model Details
model_architecture_type: # Choose from any architectures or packages below and/or add onto the list
- PyTorch
- TensorFlow
- JAX
- Safetensors
- Transformers
- PEFT
- TensorBoard
- GGUF
- Diffusers
- ONNX
- stable-baselines3
- sentence-transformers
- ml-agents
- MLX
- TF-Keras
- Adapters
- Keras
- setfit
- timm
- sample-factory
- Transformers.js
tasks_performed_by_model: # Choose from any tasks below and/or add onto the list
- Batch Correction
- Causal Inference
- Cell Clustering
- Cell Labeling
- Cell Morphology Profiling
- Cell Type Annotation
- Cell Type Classification
- Contrast Transfer Function (CTF) Estimation
- Cross Species Integration
- CryoET Particle Picking
- Data Integration
- Differential Expression
- Feature Extraction
- Frame Alignment & Motion Correction
- Gene Co-expression Prediction
- Gene Network Inference
- Hypothesis Generation
- Image Segmentation
- Imputation
- Perturbation Detection/Prediction
- Protein Localization
- Synthetic Data Generation
- Tomogram Alignment / Reconstruction
- Virtual Staining
finetuned_from: # (optional) If applicable. Add one or more.
- model_name: ""
model_url: ""
model_variants: # (optional) If applicable. Add one or more.
- variant_name: ""
variant_description: ""
variant_url: ""
### Training Details
training_date: "" # (optional) YYYY-MM-DD
uses_synthetic_data: "" # one of "Yes" or "No". Was synthetic data used in developing the model?
uses_purchased_licensed_data: "" # one of "Yes" or "No". Were any datasets used for model training purchased or licensed?
input_data_type: # Choose from any data types below and/or add onto the list
- pdf
- svg
- png
- jpg
- zarr
- czi
- tiff
- mrc
- hdf5
- csv
- docx
- txt
output_data_type: # Choose from any data types below and/or add onto the list
- embeddings
- cell types
- images
# To gather Training Dataset data, we use the ~Cross Modality Schema~ which standardizes biological information.
# https://github.com/chanzuckerberg/data-guidance/blob/main/standards/cross-modality/1.1.0/schema.md
training_datasets:
- dataset_name: ""
dataset_license: ""
dataset_version: ""
dataset_url: ""
dataset_datePublished: "" # YYYY-MM-DD
dataset_collectionTimeframe: "" # YYYY-MM-DD/YYYY-MM-DD
dataset_firstUsedInModelDev: "" # YYYY-MM When was the dataset first used during model development?
includes_consumer_info: "" # one of "Yes" or "No". Does the dataset include any aggregate consumer information?
modality: "" # one of "Imaging" or "Transcriptomic" or "Text"
# If modality = Transcriptomic:
assay_ontology_term_id: "" # e.g. "EFO:0009922" Search the Experimental Factor Ontology (https://www.ebi.ac.uk/ols4/ontologies/efo/) for dataset assay like "10x 3' v3" or "Perturb-Seq" and copy the id.
# If modality = Imaging
imaging_technology: # Choose imaging technology type from the list below and/or write in a different technology
- Brightfield Microscopy
- Phase Contrast Microscopy
- Differential Interference Contrast (DIC)
- Fluorescence Microscopy
- Confocal Microscopy
- Spinning Disk Confocal Microscopy
- Two-Photon Microscopy
- Total Internal Reflection Fluorescence (TIRF)
- STED
- SIM
- PALM
- STORM
- Transmission Electron Microscopy (TEM)
- Scanning Electron Microscopy (SEM)
- Cryo-Electron Microscopy (Cryo-EM)
- Cryo-Electron Tomography (Cryo-ET)
- Imaging Mass Cytometry (IMC)
- Multiplexed Ion Beam Imaging (MIBI)
- CODEX
- Cyclic Immunofluorescence (CycIF)
- Multiplexed Immunohistochemistry (mIHC)
- seqFISH / MERFISH
- Hematoxylin & Eosin (H&E) Staining
- Immunohistochemistry (IHC)
- Immunofluorescence (IF)
- Whole Slide Imaging (WSI)
- Electron microscopy
- Cryo-electron microscopy
- Fluorescence light microscopy
- Transmission light microscopy
- Super resolution microscopy
- X-ray microscopy
- Force microscopy
- High-contrast microscopy
- Whole-slide microscopy
disease_ontology_term_id: "" # e.g "PATO:0000461" for normal or healthy. Please follow these instructions for diseased and injured: https://github.com/chanzuckerberg/data-guidance/blob/main/standards/cross-modality/1.1.0/schema.md#disease_ontology_term_id
organism_ontology_term_id: "" # e.g. "NCBITaxon:9606" for Human. Please follow these instructions: https://github.com/chanzuckerberg/data-guidance/blob/main/standards/cross-modality/1.1.0/schema.md#organism_ontology_term_id
development_stage_ontology_term_id: "" # Please follow these instructions: https://github.com/chanzuckerberg/data-guidance/blob/main/standards/cross-modality/1.1.0/schema.md#development_stage_ontology_term_id
tissue_type: "" # one of "tissue", "organoid", "cell culture", "cell line" or "organelle"
tissue_ontology_term_id: "" # Please follow these instructions: https://github.com/chanzuckerberg/data-guidance/blob/main/standards/cross-modality/1.1.0/schema.md#tissue_ontology_term_id
num_of_cells: "" # e.g. "1000000"
### Other Resources
related_models: # (optional) Link any other models your model should be associated with.
- model_name: ""
model_url: ""
related_datasets: # (optional) Link any other datasets your model should be associated with.
- dataset_name: ""
dataset_url: ""
# See the Model Contribution Docs for detailed instructions on providing a Quickstart and Tutorial link: https://chanzuckerberg.github.io/virtual-cells-platform/
quickstart_colab_link: "" # template -> https://colab.research.google.com/drive/1VfrAM-BxXwUveDqhdwViD8BOH5yk_FU6
tutorial_colab_link: "" # (optional but encouraged) template -> https://colab.research.google.com/drive/1DrRY_mJyYkx3Bg-Y9Ev9X209jGwt8wVF?usp=sharing
model_download_link: "" # e.g. s3://.. (not needed if model is packaged in MLFlow)
Model Card Template - Details
# Model Card Template - Details
<!-- You can use standard [Markdown](https://www.markdownguide.org/basic-syntax/) in this file to format your responses including lists, tables, links, and headings.
To include images in your model card, place them in the `model_card_images/` folder and reference them like so:
Write descriptive alt text that explains what's in the image for screen readers. This is crucial for users with visual impairments. For guidance:
- [WebAIM Alt Text Guide](https://webaim.org/techniques/alttext/)
- [W3C Image Decision Tree](https://www.w3.org/WAI/tutorials/images/decision-tree/) -->
<!-- MODEL DETAILS SECTION -->
## Model Details
### Model Architecture
<!-- Brief description of the model architecture (e.g., number of layers and attention heads, embedding dimensions, input size or context length) and rationale behind it -->
...
### Parameters
<!-- Number of parameters (e.g., 15 million) -->
...
### Citation
<!-- Provide citation information for users of the model -->
...
<!-- INTENDED USE SECTION -->
## Intended Use
<!-- This section addresses questions around how the model is intended to be used in different applied contexts, discusses the foreseeable users of the model (including those affected by the model), and describes uses that are considered out of scope or misuse of the model. -->
### Primary Use Cases
<!-- List primary use cases (e.g., cell type classification, perturbation prediction, protein localization, cell morphology profiling). You can include the 'tasks_performed_by_model' that you provided in model_card_details.yaml as a starting point. -->
...
### Out-of-Scope or Unauthorized Use Cases
<!-- Suggested Text:
"Do not use the model for the following purposes:
- Use that violates applicable laws, regulations (including trade compliance laws), or third party rights such as
privacy or intellectual property rights.
- Any use that is prohibited by the [link to model license] license.
- Any use that is prohibited by the Acceptable Use Policy."
[Please include other specific out-of-scope use cases that may be relevant for this model, as applicable]
-->
...
<!-- TRAINING DETAILS SECTION -->
## Training Details
<!-- This section provides information to describe and replicate training, including the training data and the speed and size of training elements. -->
...
### Training Procedure
<!-- Briefly describe the training approach including data pre-processing steps (e.g., steps taken to clean and preprocess the data, detail tokenization, modality dependent resizing/rewriting) -->
...
### Training Code
<!-- (optional, but strongly encouraged) Provide links to training scripts -->
...
### Speeds, Sizes, Times
<!-- (optional, include if available) Provide information about throughput, start/end time, checkpoint size if relevant, etc. (optional, include if available) -->
...
### Training Hyperparameters
<!-- (optional, include if available) Examples: fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
...
<!-- PERFORMANCE METRICS SECTION -->
## Performance Metrics
<!-- This section describes the evaluation protocols, what is being measured in the evaluation, and provides the results. -->
### Metrics
<!-- List evaluation metrics used and the rationale for using them along with links where applicable. If this model was benchmarked against existing models, list them here and explain the rationale for comparison.
For example:
"The model was evaluated using a range of benchmarks to measure its performance.
Key metrics include: [metrics]." -->
...
### Evaluation Datasets
<!-- List the evaluation datasets along with links to the data where possible. Link to evaluation datasets processing code if available. -->
...
### Evaluation Results
<!-- Provide table and/or figures summarizing evaluation results -->
...
<!-- BIASES, RISKS, AND LIMITATIONS SECTION -->
## Biases, Risks, and Limitations
<!-- This section identifies potential harms, misunderstandings, and technical and limitations. It also provides
information on warnings and potential mitigations. Suggestions are provided below. -->
### Potential Biases
<!-- Suggested Text:
"- The model may reflect biases present in the training data.
- Certain demographic groups may be underrepresented."
[Please include other specific biases that may be relevant for this model, as applicable] -->
...
### Risks
<!-- Suggested Text:
"Areas of risk may include but are not limited to:
- Inaccurate outputs or hallucinations
- Potential misuse for incorrect biological interpretations."
[Please include other specific risks that may be relevant for this model, as applicable] -->
...
### Limitations
<!--
Suggested Text:
"- The model may not perform well on general tasks."
[Please include other specific limitations that may be relevant for this model, as applicable]
-->
...
### Caveats and Recommendations
<!-- (optional)
Suggested Text:
"- Review and validate outputs generated by the model.
- We are committed to advancing the responsible development and use of artificial intelligence. Please follow our Acceptable Use Policy when using the model."
For CZI and CZ Biohub models:
"- Should you have any security or privacy issues or questions related to the model, please reach out to our team at security@chanzuckerberg.com or privacy@chanzuckerberg.com, respectively."
[Please include other recommendations that may be relevant for users of this model, as applicable]
-->
...
<!-- ACKNOWLEDGEMENTS -->
## Acknowledgements
<!-- (optional) This section is for providing acknowledgement of other contributors or supporting organizations. -->
...