familorujov/llama-3.2-3B-financial-sentiment-finetuned-GGUF

Llama-3.2-3B Financial Sentiment Fine tuned GGUF (Q4KM)

Summary

familorujov/llama-3.2-3B-financial-sentiment-finetuned-Q4_K_M-GGUF is the GGUF quantized version of my financial sentiment model.

It is intended for local inference with llama.cpp and compatible runtimes (for example llama-cpp-python). Given a short finance related text, it should output one label:

• —Positive
• —Negative
• —Neutral

This repo is for inference only. For full training details, evaluation, and the original model card, see:

• —familorujov/llama-3.2-3B-financial-sentiment-finetuned

Source model

Converted from:

• —familorujov/llama-3.2-3B-financial-sentiment-finetuned

Quantization

• —Format: GGUF
• —Quantization: Q4_K_M (4 bit K-quant)
• —Model size: 3B parameters (Llama architecture)

This quantized build targets a strong quality to size tradeoff for running locally on consumer hardware.

Dataset

Trained on the Kaggle dataset:

• —sbhatti/financial-sentiment-analysis

Intended use

Use this model for:

• —sentiment labeling of financial headlines, short news, earnings commentary, and finance related statements
• —sentiment as a component inside analytics, RAG, and agent pipelines (as a classifier step)

Not intended for:

• —financial advice, trading decisions, or automated execution without human oversight
• —predicting price movements, volatility, or returns
• —general sentiment outside finance (movies, products, daily social posts)

Responsible use

This model is not financial advice. Do not use it as the only signal for decisions that impact money, safety, or people. Keep humans in the loop for any high stakes use.

License

This repo contains a quantized derivative intended for inference. You must comply with the license terms of the original model and the upstream base model.