Krishna Wisdom Assistant

Model Details

Model Description

Krishna Wisdom Assistant is a conversational language model fine-tuned to respond in a calm, reflective, spiritually inspired tone influenced by Krishna philosophy and dharmic wisdom.

The model is designed to answer everyday emotional, philosophical, and personal-reflection questions in a gentle and uplifting style. It is intended for supportive, inspirational, and wisdom-oriented dialogue rather than factual authority or professional advice.

This model was fine-tuned using supervised fine-tuning (SFT) with LoRA on a custom prompt-response dataset containing philosophical and analogy-based conversational examples.

Developed by: aaryanpethkar48
Funded by: Personal project
Shared by: Aaryan Pethkar
Model type: Causal Language Model
Language(s): English
License: Apache-2.0
Finetuned from model: TinyLlama/TinyLlama-1.1B-Chat-v1.0

Model Sources

Repository: https://huggingface.co/aaryanpethkar48/mindful-ai
Base Model: https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0
Demo: Not available

Uses

Direct Use

This model is intended for:

spiritual-style conversation
motivational and reflective chat
calm, wisdom-oriented responses
journaling prompts
gentle emotional support language
quote-style or dharmic-inspired assistant use cases

Example prompts:

Why do people change?
How do I stay calm when life feels unfair?
What does detachment really mean?
How can I stop overthinking?

Downstream Use

This model may be used in:

spiritual chatbot apps
journaling assistants
mindfulness or reflection tools
content generation for short inspirational responses
devotional or wisdom-themed conversational interfaces

Out-of-Scope Use

This model is not intended for:

medical advice
mental health diagnosis or crisis intervention
legal advice
financial advice
religious authority or scripture interpretation with scholarly accuracy
high-stakes decision-making
factual Q&A requiring high reliability

This model should not be used as a substitute for professional, clinical, legal, or emergency support.

Bias, Risks, and Limitations

This model reflects the style and tone of its fine-tuning data and may:

produce repetitive spiritual phrasing
overgeneralize life advice
respond poetically instead of analytically
provide emotionally comforting but incomplete answers
hallucinate if asked factual or technical questions outside its domain
reflect bias present in the source dataset

Because the fine-tuning dataset is relatively large but domain-specific, the model may also:

overfit to certain response patterns
produce stylistically similar outputs
struggle with highly technical or factual queries
underperform on general-purpose assistant tasks

Recommendations

Users should:

use this model primarily for inspiration, reflection, and gentle conversation
avoid relying on it for factual, legal, financial, or medical decisions
test outputs carefully before deployment in public-facing products
add moderation and safety layers for production use

How to Get Started with the Model

Transformers Inference Example

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "aaryanpethkar48/mindful-ai"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    device_map="auto"
)

messages = [
    {"role": "system", "content": "You are a wise AI inspired by Krishna philosophy."},
    {"role": "user", "content": "Why do people change?"},
]

text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

inputs = tokenizer(text, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=150,
    do_sample=True,
    temperature=0.7,
    top_p=0.9,
    repetition_penalty=1.1
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Downloads last month: 492

Safetensors

Model size

1B params

Tensor type

F16

aaryanpethkar48
/

mindful-ai