baban/QwenTranslate_English_Russian
QwenTranslate_English_Russian by baban is a 3.1 billion parameter language model fine-tuned from Qwen/Qwen2.5-3B-Instruct. This model is specifically optimized for English-Russian machine translation tasks. It leverages a 32768-token context length to handle longer translation inputs effectively. The model's primary strength lies in its specialized focus on translating between English and Russian.
Loading preview...
Overview
baban/QwenTranslate_English_Russian is a specialized machine translation model, fine-tuned from the Qwen/Qwen2.5-3B-Instruct architecture. With 3.1 billion parameters and a substantial context length of 32768 tokens, this model is designed for efficient and accurate translation between English and Russian.
Key Capabilities
- English-Russian Machine Translation: The model's core function is to translate text from English to Russian and vice-versa, having been specifically fine-tuned on the MT_En_Russian dataset.
- Large Context Window: Benefits from a 32768-token context length, allowing it to process and translate longer sentences or paragraphs while maintaining contextual coherence.
Training Details
The model was trained with a learning rate of 5e-05, a train_batch_size of 8, and gradient_accumulation_steps of 16, resulting in an effective total_train_batch_size of 1024. It utilized the AdamW optimizer and an inverse_sqrt learning rate scheduler over 3 epochs. Evaluation metrics show a loss of 1.4570 on the evaluation set, with 588,389,376 input tokens seen during training.
Good For
- Applications requiring dedicated English-Russian translation capabilities.
- Scenarios where a specialized, smaller model is preferred over larger, more general-purpose LLMs for translation tasks.