AureusERP Fine-Tuned Gemma-3 Model

This is a fine-tuned version of the Gemma-3 open-source LLM, specialized for answering questions based on AureusERP developer documentation.

🔧 Model Details

Base Model: gemma-3b
Fine-tuned On: Internal AureusERP Developer Docs
Precision: Float16 (merged & optimized for deployment)
Tokenizer: Same as base model
Training Objective: Supervised fine-tuning with synthetically generated QA pairs for LLM alignment

🧠 Use Case

This model is optimized for answering technical questions related to AureusERP, such as:

Filament integration and actions
Module architecture (e.g., FollowerAction, notifications, dev hooks)
Deployment/configuration steps
Developer-specific usage patterns

📁 Data Source

Training data was automatically generated using Gemini 2.5-flash and then curated.

📄 Source: Internal AureusDevDocs (markdown)
🤖 Process: QA pairs extracted from docs using Gemini
🔢 Format: JSON list of question-answer pairs

🔌 How to Use

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("webkul/Aureus-gemma-finetuned", torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained("webkul/Aureus-gemma-finetuned")

prompt = "How can I add followers using AureusERP FollowerAction?"
inputs = tokenizer(prompt, return_tensors="pt")

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

👥 Maintained By

Webkul AI Research Team https://webkul.com