Model Card for Ne0
Model Details
Model Name: Ne0
Base Model: Qwen/Qwen2.5-1.5B-Instruct
Training Method: LoRA (Low-Rank Adaptation)
Task: Text generation for Quantum Calibration Analysis
Developer: Shivansh Sagar Pancholi
Intended Use
This model is fine-tuned to analyze quantum experiment data and provide assessments on DRAG calibration success, optimal parameters, and suggested sweep ranges based on experimental descriptions.
Training Data
Fine-tuned on the nvidia/QCalEval dataset (test split), which focuses on quantum calibration evaluation and reasoning.
Training Procedure
Quantization: 4-bit (NF4) using BitsAndBytes.
LoRA Config: rank=8, alpha=16, target modules: q_proj, v_proj.
Optimizer: AdamW
Batch Size: 1 (with gradient accumulation of 8)
Precision: BF16 compute
How to Use---
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("ShivanshSagarPancholi/Ne0")
tokenizer = AutoTokenizer.from_pretrained("ShivanshSagarPancholi/Ne0")
