GaMS-Beta/GaMS-9B-SFT-Translator
GaMS-Beta/GaMS-9B-SFT-Translator is a 9 billion parameter instruction-tuned causal language model developed by GaMS-Beta, specifically fine-tuned for English to Slovene translation. This model builds upon GaMS-9B-Instruct, further specializing in translation tasks using approximately 130k English to Slovene translation pairs. With a context length of 16384 tokens, it excels at generating accurate Slovene translations from English input.
Loading preview...
GaMS-9B-SFT-Translator: English to Slovene Translation
GaMS-9B-SFT-Translator is a 9 billion parameter model developed by GaMS-Beta, specifically designed for high-quality English to Slovene translation. It is an instruction-tuned variant of the GaMS-9B-Instruct model, further specialized through Supervised Fine-Tuning (SFT) on a substantial dataset of approximately 130,000 English to Slovene translation pairs.
Key Capabilities
- Specialized Translation: Optimized for accurate and fluent translation from English into Slovene.
- Instruction-Following: Capable of understanding and executing translation instructions.
- Context Length: Supports a context window of 16384 tokens, allowing for translation of longer texts.
Good For
- English to Slovene Translation: Ideal for applications requiring precise translation between these two languages.
- Integration into Pipelines: Easily runnable via the
transformerslibrary'spipelineAPI for straightforward deployment.
This model was developed as part of the PoVeJMo research program, focusing on adaptive natural language processing for Slovene, with funding from the Slovenian Research and Innovation Agency (ARIS) and NextGenerationEU.