tartuNLP/Llammas
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Nov 29, 2023License:llama2Architecture:Transformer0.0K Open Weights Cold

Llammas is a 7 billion parameter Llama-2-based instruction-tuned language model developed by tartuNLP, specifically optimized for the Estonian language. It underwent continued pre-training on 5 billion tokens from CulturaX, with 75% Estonian and 25% English documents, followed by instruction-tuning using a mix of English and Estonian datasets including Alpaca-est. This model is the first open-source instruction-following LLM for Estonian, excelling in tasks requiring understanding and generation in the language.

Loading preview...