# AI Datasets by Task
Source: https://savrn.com/datasets/tasks
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

SAVRN Model Hub

# AI Datasets by Task

Every task in the library, with how many datasets each one covers. Each page lists them by demand, with their size, license, files and the GPUs they need.

Every dataset in the Model Hub, 859 of them, is filed here under the task it serves, 32 tasks in all. Text generation is the deepest at 338 entries, more than image and text to text at 138, robotics at 81 and speech recognition at 78 combined, with the list thinning down to text to speech at 64. We keep this index because a model in your own facility is only as good as the data you can tune and test it with, and data is the scarcer asset: 859 datasets against 2,760 models from 1,692 publishers. Of the eight publishers with the most models, only four release any data, NVIDIA with 6, AI at Meta with 5, Google with 3 and Helsinki-NLP with 1; Qwen, first in models at 91, publishes none.

Start from the task your workload matches, not from a model name. Read the license before anything else, since 18 different licenses cover these datasets and the terms change from one to the next, then cross to the models filed under the same task. Data before model is the order we use when scoping a deployment; it turns up gaps before hardware is on order.

- [Text Generation](https://savrn.com/datasets/tasks/text-generation) 140
- [Robotics](https://savrn.com/datasets/tasks/robotics) 120
- [Other](https://savrn.com/datasets/tasks/other) 50
- [Text Classification](https://savrn.com/datasets/tasks/text-classification) 50
- [Text Retrieval](https://savrn.com/datasets/tasks/text-retrieval) 44
- [Question Answering](https://savrn.com/datasets/tasks/question-answering) 39
- [Speech Recognition](https://savrn.com/datasets/tasks/speech-recognition) 34
- [Image to Text](https://savrn.com/datasets/tasks/image-to-text) 29
- [Time Series Forecasting](https://savrn.com/datasets/tasks/time-series-forecasting) 23
- [Tabular Classification](https://savrn.com/datasets/tasks/tabular-classification) 22
- [Image Classification](https://savrn.com/datasets/tasks/image-classification) 18
- [Object Detection](https://savrn.com/datasets/tasks/object-detection) 16
- [Visual Question Answering](https://savrn.com/datasets/tasks/visual-question-answering) 16
- [Text to Speech](https://savrn.com/datasets/tasks/text-to-speech) 14
- [Image to Image](https://savrn.com/datasets/tasks/image-to-image) 13
- [Feature Extraction](https://savrn.com/datasets/tasks/feature-extraction) 12
- [Image Segmentation](https://savrn.com/datasets/tasks/image-segmentation) 9
- [Image to 3D](https://savrn.com/datasets/tasks/image-to-3d) 9
- [Translation](https://savrn.com/datasets/tasks/translation) 9
- [Video Classification](https://savrn.com/datasets/tasks/video-classification) 9
- [Audio Classification](https://savrn.com/datasets/tasks/audio-classification) 8
- [Audio to Audio](https://savrn.com/datasets/tasks/audio-to-audio) 7
- [Reinforcement Learning](https://savrn.com/datasets/tasks/reinforcement-learning) 7
- [Text to Image](https://savrn.com/datasets/tasks/text-to-image) 7
- [Text to Video](https://savrn.com/datasets/tasks/text-to-video) 7
- [Visual Document Retrieval](https://savrn.com/datasets/tasks/visual-document-retrieval) 7
- [Image and Text to Text](https://savrn.com/datasets/tasks/image-and-text-to-text) 6
- [Image Feature Extraction](https://savrn.com/datasets/tasks/image-feature-extraction) 5
- [Tabular Regression](https://savrn.com/datasets/tasks/tabular-regression) 5
- [Video Text to Text](https://savrn.com/datasets/tasks/video-text-to-text) 5
- [Image to Video](https://savrn.com/datasets/tasks/image-to-video) 4
- [Multiple Choice](https://savrn.com/datasets/tasks/multiple-choice) 4
- [Text to 3D](https://savrn.com/datasets/tasks/text-to-3d) 4
- [Depth Estimation](https://savrn.com/datasets/tasks/depth-estimation) 3
- [Video to Video](https://savrn.com/datasets/tasks/video-to-video) 3
- [Summarization](https://savrn.com/datasets/tasks/summarization) 2
- [Table Question Answering](https://savrn.com/datasets/tasks/table-question-answering) 2
- [Text to Text](https://savrn.com/datasets/tasks/text-to-text) 2
- [Video Generation](https://savrn.com/datasets/tasks/video-generation) 2
- [Document Question Answering](https://savrn.com/datasets/tasks/document-question-answering) 1
- [Fill Mask](https://savrn.com/datasets/tasks/fill-mask) 1
- [Graph Ml](https://savrn.com/datasets/tasks/graph-ml) 1
- [Image Generation](https://savrn.com/datasets/tasks/image-generation) 1
- [Keypoint Detection](https://savrn.com/datasets/tasks/keypoint-detection) 1
- [Tabular to Text](https://savrn.com/datasets/tasks/tabular-to-text) 1
- [Text to Code](https://savrn.com/datasets/tasks/text-to-code) 1
- [Voice Activity Detection](https://savrn.com/datasets/tasks/voice-activity-detection) 1
- [Zero Shot Object Detection](https://savrn.com/datasets/tasks/zero-shot-object-detection) 1
- [Zero-Shot Classification](https://savrn.com/datasets/tasks/zero-shot-classification) 1