mashriram/Qwen3-4B-Instruct-TableLLM-SFT

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Oct 11, 2025License:apache-2.0Architecture:Transformer Open Weights Warm

mashriram/Qwen3-4B-Instruct-TableLLM-SFT is a 4 billion parameter instruction-tuned language model, based on Qwen3-4B-Instruct-2507, developed by mashriram. It is specifically fine-tuned using the RUCKBReasoning/TableLLM-SFT dataset, making it highly optimized for tasks involving table-based reasoning and understanding. With a context length of 40960 tokens, this model excels at processing and interpreting structured data within tables across multiple languages including English, Spanish, Hindi, French, and German.

Loading preview...

Overview

mashriram/Qwen3-4B-Instruct-TableLLM-SFT is a 4 billion parameter instruction-tuned language model, built upon the Qwen3-4B-Instruct-2507 base model. Developed by mashriram, this model is uniquely specialized for tasks requiring table-based reasoning and understanding.

Key Capabilities

  • Table Data Processing: Highly proficient in interpreting and extracting information from structured table data.
  • Instruction Following: Designed to accurately follow instructions for table-related queries.
  • Multilingual Support: Capable of handling table-based tasks in English, Spanish, Hindi, French, and German.
  • Extended Context: Features a substantial context window of 40960 tokens, allowing for the processing of large tables or multiple related tables.

Good For

  • Data Extraction from Tables: Automating the retrieval of specific information from tabular datasets.
  • Table Question Answering: Answering complex questions where the answer is derived from table content.
  • Structured Data Analysis: Assisting in the interpretation and summarization of data presented in tables.
  • Multilingual Table Processing: Applications requiring table understanding across its supported languages.

Finetunes
1 models