CoolFace
Modelpublic

ApexVOrteX-1/Financial-Qwen-Model

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
1likes16downloads
Model Card

Financial Analyst AI Assistant

A domain-specific Large Language Model adapter fine-tuned for financial question answering and reasoning tasks.

This model is a LoRA adapter built on top of Qwen2.5-1.5B-Instruct and trained using QLoRA with 4-bit quantization to efficiently adapt the model for financial analysis tasks.

Model Details

Model Description

  • —Developed by: Ahmed Elsayed Taha
  • —Model type: Causal Language Model (LLM) LoRA Adapter
  • —Base model: Qwen2.5-1.5B-Instruct
  • —Fine-tuning method: QLoRA (LoRA + 4-bit Quantization)
  • —Language(s): English
  • —License: Apache 2.0 (inherits from base model license)
  • —Finetuned from: Qwen/Qwen2.5-1.5B-Instruct

Model Sources

  • —Base Model: https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct

Uses

Direct Use

This model can be used for:

  • —Financial question answering
  • —Accounting concept explanation
  • —Financial reasoning tasks
  • —Understanding financial contexts
  • —Generating structured financial responses

Downstream Use

Possible applications:

  • —Financial analyst assistants
  • —Document-based financial analysis systems
  • —Financial education tools
  • —Retrieval Augmented Generation (RAG) pipelines
  • —Business intelligence assistants

Out-of-Scope Use

This model should not be used as:

  • —A replacement for professional financial advisors
  • —A source of guaranteed investment decisions
  • —A system for high-stakes financial decisions without human verification

Training Details

Training Data

The model was fine-tuned using:

  • —TheFinAI/Fino1ReasoningPath_FinQA
  • —TheFinAI/Fino1ReasoningPathFinQAv2

Dataset statistics:

DatasetSamples
Fino1ReasoningPath_FinQA5,499
Fino1ReasoningPathFinQAv23,472
Total8,971

The dataset contains:

  • —Open-ended financial questions
  • —Ground-truth answers
  • —Financial reasoning paths
  • —Generated responses

Preprocessing

The dataset was converted into Qwen instruction format:

json
{
  "messages": [
    {
      "role": "system",
      "content": "You are a financial analyst AI assistant."
    },
    {
      "role": "user",
      "content": "Financial question"
    },
    {
      "role": "assistant",
      "content": "Financial answer"
    }
  ]
}