yandex/YaLM-100B
Yandex's 100-billion parameter GPT-like pretrained language model for text generation and processing.
Not currently ranked — collecting fresh signals.
star history
YaLM 100B is a large-scale pretrained language model trained on 1.7TB of online texts, books, and other sources in English and Russian. The repository provides inference code based on DeepSpeed ZeRO-3 and supports multi-GPU tensor parallelism. It includes Docker images and scripts for downloading model weights and vocabulary, enabling deployment on GPU clusters.
Frequently asked
- What is yandex/YaLM-100B?
- Yandex's 100-billion parameter GPT-like pretrained language model for text generation and processing.
- Is YaLM-100B open source?
- Yes — yandex/YaLM-100B is open source, released under the Apache-2.0 license.
- What language is YaLM-100B written in?
- yandex/YaLM-100B is primarily written in Python.
- How popular is YaLM-100B?
- yandex/YaLM-100B has 3.8k stars on GitHub.
- Where can I find YaLM-100B?
- yandex/YaLM-100B is on GitHub at https://github.com/yandex/YaLM-100B.