baban/QwenTranslate_English_Bengali_100K_SFT

TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 12, 2026License:otherArchitecture:Transformer Cold

The baban/QwenTranslate_English_Bengali_100K_SFT model is a 3.1 billion parameter language model, fine-tuned from Qwen/Qwen2.5-3B-Instruct, specifically designed for English-to-Bengali translation. This model leverages a 32768-token context length and is optimized for high-quality translation tasks between these two languages. Its primary strength lies in its specialized training on the MT_En_Bengali dataset, making it highly effective for applications requiring accurate English-Bengali language conversion.

Loading preview...

Model Overview

The baban/QwenTranslate_English_Bengali_100K_SFT model is a specialized language model fine-tuned for English-to-Bengali translation. It is built upon the robust Qwen/Qwen2.5-3B-Instruct architecture, featuring 3.1 billion parameters and a substantial 32768-token context window.

Key Capabilities

  • English-Bengali Translation: The model's core capability is translating text from English to Bengali, having been specifically fine-tuned on the MT_En_Bengali dataset.
  • Instruction Following: Inherits instruction-following capabilities from its base model, Qwen2.5-3B-Instruct, which can be beneficial for structured translation tasks.

Training Details

The model was trained with a learning rate of 5e-05, a total batch size of 1024 (achieved with 8 GPUs and 32 gradient accumulation steps), and for 3 epochs. The training process resulted in a validation loss of 0.4847, indicating effective learning on the translation task.

Ideal Use Cases

  • Localized Content Creation: Translating English content into Bengali for websites, applications, or marketing materials.
  • Cross-Lingual Communication: Facilitating communication between English and Bengali speakers.
  • Research and Development: As a base for further fine-tuning on more specific English-Bengali translation domains.