Last updated:

Model Card

What is a Model Card

A Model Card is the README.md file in a model repository that describes the model's basic information, use cases, training process, and evaluation results. A well-written Model Card helps other users quickly understand the model's capabilities and applicable scenarios.

A Model Card consists of two parts:

YAML metadata header: Located at the top of the file, wrapped by --- delimiters, defining the model's tags and classification information.
Body content: Detailed model description written in Markdown format.

What a Model Card Should Contain

A complete Model Card should include the following sections:

Section	Description
Model Name	The official name of the model
Model Overview	Brief description of model type, architecture, and core capabilities
Usage	Code examples or invocation instructions
Use Cases	Tasks or problems the model is suited for
Training Data	Description of the dataset used for training
Training Process	Training framework, hyperparameters, hardware environment, etc.
Evaluation Results	Performance metrics on benchmark datasets
Limitations & Biases	Known limitations, biases, and unsuitable scenarios

Metadata Format

The YAML metadata is placed at the top of the README.md file in the following format:

---
language:
  - zh
  - en
tags:
  - text-generation
  - conversational
license: apache-2.0
library_name: transformers
pipeline_tag: text-generation
industry_tags:
  - general
---

Key fields:

Field	Description
`language`	List of languages the model supports
`tags`	Task tags for classification and search
`license`	Open-source license used by the model
`library_name`	Framework used by the model (e.g., `transformers`, `pytorch`)
`pipeline_tag`	Primary task type of the model
`industry_tags`	Industry tags the model applies to

Model Card Template

Below is a complete Model Card template:

---
language:
  - zh
  - en
tags:
  - text-generation
license: apache-2.0
library_name: transformers
pipeline_tag: text-generation
---

# Model Name

## Model Overview

Briefly describe the model architecture, parameter scale, and core capabilities.

## Usage

### Quick Start

Load and use the model with the transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("namespace/model-name")
tokenizer = AutoTokenizer.from_pretrained("namespace/model-name")

inputs = tokenizer("Hello", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

## Use Cases

- Use case one description
- Use case two description

## Training Data

Describe the training dataset and its scale.

## Training Process

Describe the training framework, hyperparameter configuration, and hardware resources.

## Evaluation Results

| Benchmark | Metric | Score |
|-----------|--------|-------|
| Dataset A | Accuracy | 0.95 |
| Dataset B | F1 | 0.92 |

## Limitations & Biases

Describe the model's known limitations and potential biases.

Supported Task Tags

The platform supports the following task tags for model classification:

Tag	Description
`text-generation`	Text Generation
`text-classification`	Text Classification
`token-classification`	Token Classification (e.g., NER)
`question-answering`	Question Answering
`translation`	Translation
`summarization`	Summarization
`conversational`	Conversational
`fill-mask`	Fill Mask
`feature-extraction`	Feature Extraction
`image-classification`	Image Classification
`object-detection`	Object Detection
`image-segmentation`	Image Segmentation
`text-to-image`	Text to Image
`image-text-to-text`	Image-Text to Text
`automatic-speech-recognition`	Automatic Speech Recognition
`text-to-speech`	Text to Speech
`audio-classification`	Audio Classification
`sentence-similarity`	Sentence Similarity
`zero-shot-classification`	Zero-Shot Classification
`reinforcement-learning`	Reinforcement Learning

Supported Industry Tags

Tag	Description
`general`	General
`finance`	Finance
`healthcare`	Healthcare
`education`	Education
`legal`	Legal
`manufacturing`	Manufacturing
`energy`	Energy
`transportation`	Transportation
`agriculture`	Agriculture
`telecom`	Telecommunications
`government`	Government

Note

Model Card metadata tags are used on the platform for model classification display and search filtering. Setting accurate tags improves model discoverability.