baban/QwenTranslate_English_Hindi
The baban/QwenTranslate_English_Hindi model is a 3.1 billion parameter language model fine-tuned from Qwen/Qwen2.5-3B-Instruct. It is specifically optimized for English-Hindi machine translation tasks, leveraging its base architecture for language understanding and generation. This model is designed for applications requiring translation between English and Hindi, demonstrating a reported loss of 0.3296 on its evaluation set.
Loading preview...
Model Overview
The baban/QwenTranslate_English_Hindi model is a specialized language model fine-tuned from the Qwen/Qwen2.5-3B-Instruct base architecture. With 3.1 billion parameters, this model is engineered for machine translation between English and Hindi.
Key Capabilities
- English-to-Hindi Translation: The primary function of this model is to translate text from English to Hindi.
- Hindi-to-English Translation: While not explicitly stated, fine-tuning on an English-Hindi dataset typically implies bidirectional translation capabilities.
- Qwen2.5-3B-Instruct Foundation: Benefits from the robust language understanding and generation capabilities of its Qwen2.5-3B-Instruct base.
Training Details
The model was fine-tuned on the MT_En_Hindi dataset, achieving an evaluation loss of 0.3296. Key training hyperparameters included a learning rate of 5e-05, a train_batch_size of 8, and a gradient_accumulation_steps of 16, resulting in a total_train_batch_size of 1024 over 3 epochs. The optimizer used was ADAMW_TORCH with an inverse_sqrt learning rate scheduler.
Intended Use Cases
This model is suitable for applications requiring efficient and accurate translation between English and Hindi, such as:
- Content Localization: Translating documents, websites, or user interfaces.
- Cross-lingual Communication: Facilitating understanding between English and Hindi speakers.
- Research and Development: As a component in larger multilingual NLP systems.