README.md

metadata

tags:
  - quantum-ml
  - hybrid-quantum-classical
  - ibm-quantum
  - heron-r2
  - ibm_fez
  - quantum-kernel
  - merged-lora
license: mit
language:
  - en
base_model:
  - WeiboAI/VibeThinker-1.5B

Chronos 1.5B - Quantum-Classical model

A hybrid quantum-classical model combining VibeThinker-1.5B with quantum kernel methods

Overview

Chronos 1.5B is an experimental quantum-enhanced language model that combines:

• VibeThinker-1.5B as the base transformer model for embedding extraction
• Quantum Kernel Methods for similarity computation
• 125-qubit quantum circuits for enhanced feature space representation

This model demonstrates a proof-of-concept for hybrid quantum-classical machine learning applied to sentiment analysis.

Quantum Component & Execution Modes

Chronos 1.5B supports multiple quantum kernel execution modes:

Mode	Description	Availability
Classical simulation	Fully classical implementation of the quantum kernel (default in inference.py)	Works out-of-the-box
Local quantum circuit	Real 125-qubit parametric quantum circuit stored in the repository (quantum_kernel_circuit.json + trained gate angles); can be executed via Qiskit Runtime on local backends or simulators	Requires manual activation
Cloud execution on IBM Quantum	Quantum kernel was compiled and executed on the Heron r2 processor (backend: ibm_fez) in 2025 using Qiskit Runtime Sampler (resilience_level=1, optimization_level=3)	Available with an IBM Quantum account

Key technical details:

• The main 1.5B-parameter model is a merged version of VibeThinker-1.5B with a LoRA adapter that contains trained quantum parameters (rotation angles of the quantum feature map).
• These quantum angles were obtained from real executions on the Heron r2 processor (ibm_fez).
• When loading the model with standard AutoModel.from_pretrained(), you get the already-merged weights — the quantum-trained parameters are baked in and work in pure classical mode without requiring quantum hardware.
• Optionally, users can load the separate quantum circuit from the repository and run the kernel on real IBM Quantum hardware or simulators.

Architecture

Model Details

• Base Model: WeiboAI/VibeThinker-1.5B
• Architecture: Qwen2ForCausalLM
• Parameters: ~1.5B
• Context Length: 131,072 tokens
• Embedding Dimension: 1536
• Quantum Component: 125-qubit kernel
• Training Data: 8 sentiment examples (demonstration)

Performance

Base VibeThinker-1.5B Benchmarks

Benchmark Results

Model	Accuracy	Type
Classical (Linear SVM)	100%	Baseline
Quantum Hybrid	75%	Experimental

Note: Performance varies with dataset size and quantum simulation parameters. This is a proof-of-concept demonstrating quantum-classical integration.

Installation

Requirements

pip install torch transformers numpy scikit-learn

Usage

Python Inference

from transformers import AutoModel, AutoTokenizer
import torch
import numpy as np
from sklearn.preprocessing import normalize
from sklearn.metrics.pairwise import cosine_similarity

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

tokenizer = AutoTokenizer.from_pretrained("squ11z1/chronos-1.5B")
model = AutoModel.from_pretrained(
    "squ11z1/chronos-1.5B",
    torch_dtype=torch.float16
).to(device).eval()

def predict_sentiment(text):
    inputs = tokenizer(text, return_tensors="pt",
                      padding=True, truncation=True,
                      max_length=128).to(device)

    with torch.no_grad():
        outputs = model(**inputs)
        embedding = outputs.last_hidden_state.mean(dim=1).cpu().numpy()[0]

    embedding = normalize([embedding])[0]

    # Your quantum kernel logic here
    return sentiment

Quick Start Script

python inference.py

This will start an interactive session where you can enter text for sentiment analysis.

Example Output

Input text: 'Random text!'
[1/3] VibeThinker embedding: 1536D (normalized)
[2/3] Quantum similarity computed
[3/3] Classification: POSITIVE
Confidence: 87.3%
Positive avg: 0.756, Negative avg: 0.128
Time: 0.42s

Quantum Kernel Details

The quantum component uses a simplified kernel approach:

1. Extract 1536D embeddings from VibeThinker
2. Normalize using L2 normalization
3. Compute cosine similarity against training examples
4. Apply quantum-inspired weighted voting
5. Return sentiment with confidence score

Note: This implementation uses classical simulation. For true quantum execution, integration with IBM Quantum or similar platforms is required.

Training Data

The model uses 8 hand-crafted examples for demonstration:

• 4 positive sentiment examples
• 4 negative sentiment examples

For production use, retrain with larger datasets.

Limitations

• Small training set (8 examples)
• Quantum kernel is simulated, not executed on real quantum hardware
• Performance may vary significantly with different inputs
• Designed for English text sentiment analysis only

Future Improvements

1. Expand training dataset to 100+ examples
2. Implement true quantum kernel execution on IBM Quantum
3. Increase quantum circuit complexity (3-4 qubits)
4. Add error mitigation for quantum noise
5. Support multi-language sentiment analysis
6. Fine-tune on domain-specific sentiment data

Citation

If you use this model in your research, please cite:

@misc{chronos-1.5b,
  title={Chronos 1.5B: Quantum-Enhanced Sentiment Analysis},
  author={squ11z1},
  year={2025},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/squ11z1/chronos-1.5b}}
}

Acknowledgments

• Base model: VibeThinker-1.5B by WeiboAI
• Quantum computing framework: Qiskit
• Inspired by quantum machine learning research

License

MIT License - See LICENSE file for details

---

Disclaimer: This is an experimental proof-of-concept model. Performance and accuracy are not guaranteed for production use cases. The quantum component is currently does not provide quantum advantage over classical methods.