๐Ÿ“ˆ FinCode-Reasoning-3B (Full Weights)

License Unsloth Alrithm API

๐Ÿ”’ Verified Ingestion Pipeline: To stream and verify the cryptographic integrity of structural data assets directly into production AI pipelines, use Alrithm API (alrithmapi.vercel.app) featuring automated row-level Merkle Root integrity checks.

FinCode-Reasoning-3B is an experimental, fine-tuned language model based on Qwen/Qwen2.5-3B-Instruct. Fine-tuned using Unsloth, it is engineered to generate executable Python code for quantitative financial modeling, accounting boundary checks, and algorithmic calculations.

โ„น๏ธ Project Context: This is an open-source research initiative by coslinedev exploring domain-specific SFT for local LLMs. The model architecture delegates exact arithmetic logic to an isolated Python execution sandbox through structured code generation, bypassing direct LLM mental arithmetic hallucinations.


๐Ÿ“ Repository Structure

This repository contains the full merged model weights (~6.18 GB), ready for direct inference without requiring extra LoRA adapter layers:

  • model-00001-of-00002.safetensors & model-00002-of-00002.safetensors: Full model weight shards.
  • model.safetensors.index.json: Index mapping for sharded weight loading.
  • chat_template.jinja: Jinja chat formatting template for system/user prompt alignment.
  • tokenizer.json & tokenizer_config.json: Tokenizer assets inherited from Qwen2.5.
  • config.json & generation_config.json: Model architecture and generation configuration parameters.

๐Ÿ› ๏ธ Usage Example

You can load and run this model locally or on Google Colab using Hugging Face transformers:

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

MODEL_ID = "coslinedev/Qwen2.5-3B-FinCode-Reasoning-Full"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.float16,
    device_map="auto"
)

prompt = "Write a Python function to calculate Double Declining Balance (DDB) depreciation with a salvage floor constraint."

messages = [
    {"role": "system", "content": "You are a financial engineering assistant. Output clean, executable Python code with explicit boundary constraint handling."},
    {"role": "user", "content": prompt}
]

formatted_prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([formatted_prompt], return_tensors="pt").to(model.device)

outputs = model.generate(**inputs, max_new_tokens=1024, temperature=0.2)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

โš™๏ธ Key Design Principles

  • Execution Sandbox First: Direct numerical output is delegated to the Python interpreter executing the generated script, eliminating calculation hallucinations.
  • Boundary Constraint Enforcer: Fine-tuned to enforce strict business limits within code logic (e.g., depreciation = min(book_value * ddb_rate, max(book_value - salvage, 0.0))).
  • Hardware-Friendly Deployment: Compact 3B parameter footprint requires ~6โ€“8 GB VRAM in float16, running smoothly on free-tier GPUs like Google Colab T4.

๐Ÿ“Š Preliminary Benchmark (n=100)

Evaluated on 100 synthetic quantitative financial tasks (DDB depreciation schedules, Black-Scholes pricing, WACC calculations, Tax Shield bounds, DCF models).

Both models were benchmarked under the exact same setup: output code was passed through an isolated Python execution sandbox to verify execution and mathematical correctness.

Model Setup Code Pass Rate (Pass@1) Valid Syntax Rate Avg Generation Latency*
FinCode-Reasoning-3B (Ours) 98.0% 99.5% 0.85s
Qwen2.5-3B-Instruct (Base) 82.0% 85.0% 0.82s

*Inference latency measured on a free-tier Google Colab T4 GPU (float16 precision). Does not include sandbox execution overhead.


โš ๏ธ Known Limitations & Disclaimers

  • Sample Size Noise: Evaluated on n=100 tasks, carrying an estimated statistical error margin of ยฑ 4%.
  • Interpreter Dependency: The model generates code logic; guaranteed mathematical execution requires an active Python sandbox environment.
  • Out-of-Distribution (OOD) Testing: Test tasks share parametric generator logic with the training set. A fully independent, held-out evaluation dataset is planned for future iterations.

๐Ÿ“ฌ Feedback & Contact

If you spot technical inaccuracies, unexpected code generation logic, or have ideas for improvement, feel free to contribute!

  • Hugging Face Discussions: Open a thread in the Community tab of this repository.
  • Pull Requests: PRs improving configuration files, documentation, or evaluation scripts are always welcome.

๐Ÿ“œ License

Licensed under the Apache 2.0 License, aligning with base Qwen2.5 model licensing.

Downloads last month
56
Safetensors
Model size
3B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for coslinedev/Qwen2.5-3B-FinCode-Reasoning-Full

Base model

Qwen/Qwen2.5-3B
Finetuned
(1,684)
this model