Ljy2004/mapfinben-qwen35-9b-merged-unified-v3
Ljy2004/mapfinben-qwen35-9b-merged-unified-v3 is a 9.41 billion parameter language model, fine-tuned from Qwen3.5-9B-Base using LoRA SFT. This model is specifically optimized for financial benchmark tasks, trained on the unified MapFinBen dataset. It is designed to excel in financial reasoning and analysis, making it suitable for applications requiring specialized financial understanding.
Loading preview...
Overview
Ljy2004/mapfinben-qwen35-9b-merged-unified-v3 is a specialized language model developed by Ljy2004, built upon the Qwen3.5-9B-Base architecture. It features approximately 9.41 billion parameters and has been fine-tuned using the LoRA SFT method (rank=16, 1.5 epoch, lr=1e-4, cutoff_len=2048).
Key Capabilities
- Financial Domain Expertise: Specifically trained on the unified MapFinBen dataset, comprising 51,064 samples, to enhance performance in financial benchmarks.
- Optimized for Financial Reasoning: Designed to address tasks within the CCL26-Eval-MapFinBen evaluation framework.
- Efficient Fine-tuning: Utilizes LoRA (Low-Rank Adaptation) for efficient adaptation of the base model.
When to Use This Model
- Financial Analysis: Ideal for applications requiring deep understanding and processing of financial data and queries.
- Benchmark Performance: Suited for tasks and evaluations related to the MapFinBen benchmark.
- Specialized Financial LLM: A strong candidate for use cases where a general-purpose LLM might lack the necessary domain-specific knowledge in finance.
Users should employ the Qwen3.5 chat template with enable_thinking=False for optimal performance, consistent with its training and official evaluation methodology.