Overview

myfi/parser_model_ner_4.00 is a 4 billion parameter instruction-tuned language model developed by myfi. It is finetuned from the unsloth/Qwen3-4B-Instruct-2507 base model, leveraging the Qwen3 architecture. This model was trained with significant speed improvements using the Unsloth library and Huggingface's TRL library, enabling faster iteration and deployment.

Key Capabilities

• Efficient Finetuning: Benefits from Unsloth's optimizations for 2x faster training.
• Qwen3 Architecture: Inherits the robust capabilities of the Qwen3 model family.
• Extended Context: Supports a substantial context length of 32768 tokens, suitable for processing longer documents.

Good For

• Named Entity Recognition (NER): Designed for tasks requiring the identification and classification of entities within text.
• Text Parsing: Ideal for applications that involve structured extraction of information from unstructured text.
• Resource-Efficient Deployment: Its 4B parameter size, combined with Unsloth's training efficiency, makes it suitable for scenarios where faster training and inference are desired without sacrificing significant performance.