Mistral-7B QLoRA Adapter — ECTSum

Base Model

• Base model: mistralai/Mistral-7B-v0.1
• Quantization: 4-bit (bnb)
• Fine-tuning method: QLoRA (LoRA adapters only)

Training Data

ECTSum: A New Benchmark for Executive Summarization of Earnings Calls

• The full dataset is not redistributed here.
• Only processed subsets were used during experimentation.
• All rights to the dataset belong to the original authors.

Intended Use

• research and experimentation
• financial document summarization
• studying lightweight fine-tuning with long-context inputs

Evaluation Summary

• ROUGE/BLEU improvements were marginal due to the highly abstractive nature of the task.
• BERTScore indicates that the base model already captures general financial semantics.
• Qualitative inspection suggests improvements in summary structure and tone.

GitHub Repository

Usage