prompt

Optional. The number of inference steps. Defaults to 40

n_steps

Optional. The proportion of `n_steps` to run the base model for. Must be a number between 0 and 1. Defaults to 0.8

high_noise_frac


Stable Diffusion XL 1.0 base combined with the refiner as an ensemble of experts.

![](https://raw.githubusercontent.com/Stability-AI/generative-models/477d8b9a7730d9b2e92b326a770c0420d00308c9/assets/000.jpg)

Options that can be passed when loading the model:

 - `model.compile` (bool): whether or not to use `torch.compile`. Defaults to false
 - `model.enable_cpu_offload` (bool): whether or not to use [`enable_model_cpu_offload()`](https://huggingface.co/docs/diffusers/main/en/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLPipeline.enable_model_cpu_offload). Defaults to false


image

quickstart

Stable Diffusion XL 1.0 + refiner as an Ensemble of Experts

input

See [here](https://github.com/openai/gpt-2) for more details.

The continued strings in the same shape as `input`

output


DPT-Hybrid is model trained on 1.4 million images for monocular depth estimation. It was introduced in the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by Ranftl et al. (2021).  DPT uses the Vision Transformer (ViT) as backbone and adds a neck + head on top for monocular depth estimation.


depth

DPT hybrid is a monocular depth estimation model


Bark is a transformer-based text-to-audio model created by [Suno](https://suno.ai). Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying.


audio

sample_rate

The sentences to fill `[MASK]` tokens in.

The maximum number of tokens to predict for each mask. Optional, defaults to 1.

max_tokens

This model can predict masked tokens in a sentence. For example, it might predict `capital` given `Paris is the [MASK] of France.`.

See [here](https://huggingface.co/bert-base-uncased) for more details.

The predicted tokens for each input sentence. This will have shape `[N, max_tokens]`, but some cells may be empty.

tokens

The scores for each predicted token. This will have shape `[N, max_tokens]`, but some cells may have a score of zero.

scores

A language model that can fill masked tokens in sentences

See [here](https://github.com/facebookresearch/fairseq/blob/main/examples/bart/README.md) for more details.

The summarized strings in the same shape as `input`

A BART model fine-tuned on CNN/Daily Mail to summarize text.

candidate_labels

An optional template string for the model to use. Defaults to 'This example is about {}.'

template

An optional max_length to pass to the model. Defaults to 128.

max_length

See [here](https://huggingface.co/facebook/bart-large-mnli) for more details.

Note: This model performs multi-label classification (i.e. zero or more labels may be true for each input).

Scores between 0 and 1 for each element of `input` for each label in `candidate_labels`

BART Large MNLI is a model that can do zero shot classificiation.

Translation from Hindi to French and Chinese to English

The strings to translate as batches grouped by language

The source language (or empty string) for every batch item

source_language

target_language

See [here](https://about.fb.com/news/2020/10/first-multilingual-machine-translation-model/) for more details. M2M100 supports the following languages:
- Afrikaans
- Danish
- Dutch
- German
- English
- Icelandic
- Luxembourgish
- Norwegian
- Swedish
- WesternFrisian
- Yiddish
- Asturian
- Catalan
- French
- Galician
- Italian
- Occitan
- Portuguese
- Romanian
- Spanish
- Belarusian
- Bosnian
- Bulgarian
- Croatian
- Czech
- Macedonian
- Polish
- Russian
- Serbian
- Slovak
- Slovenian
- Ukrainian
- Estonian
- Finnish
- Hungarian
- Latvian
- Lithuanian
- Albanian
- Armenian
- Georgian
- Greek
- Breton
- Irish
- ScottishGaelic
- Welsh
- Azerbaijani
- Bashkir
- Kazakh
- Turkish
- Uzbek
- Japanese
- Korean
- Vietnamese
- Chinese
- Bengali
- Gujarati
- Hindi
- Kannada
- Marathi
- Nepali
- Oriya
- Panjabi
- Sindhi
- Sinhala
- Urdu
- Tamil
- Cebuano
- Iloko
- Indonesian
- Javanese
- Malagasy
- Malay
- Malayalam
- Sundanese
- Tagalog
- Burmese
- CentralKhmer
- Lao
- Thai
- Mongolian
- Arabic
- Hebrew
- Pashto
- Farsi
- Amharic
- Fulah
- Hausa
- Igbo
- Lingala
- Luganda
- NorthernSotho
- Somali
- Swahili
- Swati
- Tswana
- Wolof
- Xhosa
- Yoruba
- Zulu
- HaitianCreole

The translated strings in the same shape as `input`

M2M100 is a model that can translate directly between any pair of 100 languages.

question

Context for each of the questions. In the same shape as `question`

context

See [here](https://huggingface.co/distilbert-base-cased-distilled-squad) for more details.

Answers to the questions in the same shape as `question`

answer

A DistilBERT model fine tuned for question answering.

See [here](https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english) for more details.

Scores between -1 and 1 for each element of `input`. Negative scores correspond to a negative sentiment.

DistilBERT base uncased finetuned SST-2 is a model that can do sentiment analysis.

an_optional_name


Some description of the model that can be as detailed as you want it to be

This can span multiple lines. It can use markdown, but shouldn't use arbitrary HTML
(which will usually be treated as text instead of HTML when this is rendered)

You can use images and links, but the paths must either be complete https
URLs (e.g. https://example.com/image.png) or must reference a file in the misc folder (e.g `@misc/file.png`).
See the `misc_files` section of this document for more details

![Model Architecture](@misc/model_architecture.png)


Some description of the output with emoji! 🥳

a_self_test

A short description that should be less than or equal to 100 characters.

Carton

Run any ML model from any programming language.

Explore public models from the community.

Featured

stabilityai / sdxl

openai / gpt2-medium

intel / dpt-hybrid-midas

suno-ai / bark

Latest

google-research / bert-base-uncased

facebook / bart-large-cnn

facebook / bart-large-mnli

facebook / m2m100_418M

huggingface / distilbert-base-cased-distilled-squad

Categories

Computer Vision >

Natural Language Processing >

Audio >