OLMo
Language models that are open end to end, from training data to intermediate checkpoints.
OLMo is Ai2's family of open language models, designed "by scientists, for scientists". The training data mixes are named and available, and checkpoints are published at least every 1,000 training steps, for models from 1 to 32 billion parameters.
And where does AI come in?
The project gives researchers models they can study end to end: training, evaluation and inference code, data and intermediate states. The published configurations make it possible to reproduce a training run.
The AI label
- Model or technique
- Not filled in
- Model published
- Not filled in
- Where it runs
- Not filled in
- Data
- Not filled in
- Role of people
- Not filled in
- Frugality
- Not filled in
“Not filled in”: MAUB could not check it with the project. Know more? Write to us.
Run this project? Show the “Listed on MAUB” badge
Paste this code on your site: an image and a link to this page. No script, no tracker: MAUB cannot tell who displays it.
<a href="https://maub.subalpine.it/en/projets/olmo"><img src="https://maub.subalpine.it/statique/img/badge-recense-en.svg" alt="Listed on MAUB" width="180" height="48"></a>Something wrong or out of date? Write to contact@subalpine.it.
The monthly newsletter
Once a month, what moved on MAUB, in your inbox.