asifdotpy/vetta-granite-2b-v3

Vetta Granite Merged Model v3

This repository contains the full merged Vetta AI interviewer model, fine-tuned on Granite 3.0 2B Instruct with LoRA weights integrated.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained(
    "asifdotpy/vetta-granite-2b-v3",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("asifdotpy/vetta-granite-2b-v3")

# Generate
inputs = tokenizer("Begin a technical interview...", return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=256, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Training Details

• Base Model: ibm-granite/granite-3.0-2b-instruct
• Training Method: LoRA fine-tuning with merged weights
• Dataset: Custom interview conversation dataset
• Training Steps: 2250
• Final Loss: 0.1124
• Precision: 16-bit

Intended Use

This model is designed to conduct professional AI-powered interviews, providing empathetic and technically accurate responses.