You need to agree to share your contact information to access this model

The information you provide will be collected, stored, processed and shared in accordance with the Ekoaham Dutivnasti Privacy Policy.

VRINDA PROPRIETARY LICENSE AGREEMENT

VRINDA-8B Version Release Date: December 18, 2024

"Agreement" means the terms and conditions for use, reproduction, distribution and modification of the VRINDA Materials set forth herein.

"Documentation" means the specifications, manuals and documentation accompanying VRINDA-8B distributed by Ekoaham Dutivnasti at https://vrinda.ekoahamdutivnasti.com.

"Licensee" or "you" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity's behalf), of the applicable age required to consent to the terms of this Agreement.

"VRINDA Materials" means, collectively, Ekoaham Dutivnasti's proprietary VRINDA-8B model and any Documentation made available under this Agreement.

1. License Rights and Redistribution

a. Grant of Rights. Subject to the terms of this Agreement, Ekoaham Dutivnasti hereby grants you a non-exclusive, non-transferable, revocable license to:

Use the VRINDA Materials for personal, non-commercial purposes
Use the VRINDA Materials for academic research and evaluation
Use the VRINDA Materials for educational purposes

b. Restrictions. You shall NOT:

Use the VRINDA Materials for any commercial purpose without written permission
Modify, fine-tune, or create derivative works from the VRINDA Materials
Redistribute, share, or transfer the VRINDA Materials to any third party
Sell, sublicense, or commercially exploit the VRINDA Materials
Use the VRINDA Materials to train other AI models
Reverse engineer or extract components from the VRINDA Materials
Use the VRINDA Materials for any illegal, harmful, or malicious purposes

2. Additional Commercial Terms

If you wish to use VRINDA Materials for commercial purposes, you must obtain a separate commercial license from Ekoaham Dutivnasti. Contact: [email protected]

3. Disclaimer of Warranty

THE VRINDA MATERIALS ARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED. EKOAHAM DUTIVNASTI DISCLAIMS ALL WARRANTIES INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

4. Limitation of Liability

IN NO EVENT SHALL EKOAHAM DUTIVNASTI BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES ARISING FROM USE OF THE VRINDA MATERIALS.

5. Termination

This Agreement terminates automatically upon any violation of its terms. Upon termination, you must delete all copies of the VRINDA Materials.

6. Governing Law

This Agreement shall be governed by the laws of India.

By clicking "Submit" below, you accept the terms of this Agreement and acknowledge that you have read our Privacy Policy.

VRINDA-8B

A State-of-the-Art Bilingual Large Language Model

Developed by Ekoaham Dutivnasti

Overview

VRINDA-8B is a powerful 8 billion parameter large language model specifically designed for seamless bilingual conversations in English and Hindi. Built with a focus on natural dialogue, emotional intelligence, and cultural understanding, VRINDA represents the next generation of AI assistants tailored for the Indian subcontinent.

Highlights

Feature	Description
Architecture	Transformer-based decoder-only model
Parameters	8 Billion
Context Window	32,768 tokens
Languages	English, Hindi, Hinglish
Precision	BFloat16

Model Capabilities

🗣️ Bilingual Fluency

Native-level understanding and generation in both English and Hindi, with natural code-switching (Hinglish) support.

🧠 Advanced Reasoning

Strong performance on mathematical reasoning, logical deduction, and multi-step problem solving.

💬 Conversational Excellence

Designed for natural, engaging dialogue with personality and emotional awareness.

🇮🇳 Cultural Intelligence

Deep understanding of Indian context, expressions, idioms, and cultural nuances.

Quick Start

Installation

pip install transformers torch accelerate

Basic Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "ekoahamdutivnasti/VRINDA"

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True
)

# Create conversation
messages = [
    {"role": "system", "content": "You are VRINDA, a helpful AI assistant."},
    {"role": "user", "content": "Hello VRINDA, how are you?"}
]

# Generate response
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=512,
    temperature=0.7,
    top_p=0.9,
    do_sample=True
)

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

Quantized Inference (Low VRAM)

from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
import torch

quantization_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_compute_dtype=torch.bfloat16,
    bnb_4bit_use_double_quant=True,
    bnb_4bit_quant_type="nf4"
)

model = AutoModelForCausalLM.from_pretrained(
    "ekoahamdutivnasti/VRINDA",
    quantization_config=quantization_config,
    device_map="auto",
    trust_remote_code=True
)

Technical Specifications

Property	Value
Model Type	Causal Language Model (Decoder-only)
Total Parameters	8,000,000,000
Hidden Size	4,096
Intermediate Size	12,288
Number of Layers	36
Attention Heads	32
Key-Value Heads	8 (GQA)
Head Dimension	128
Vocabulary Size	151,936
Max Position Embeddings	40,960
RoPE Theta	1,000,000
Activation Function	SiLU
Normalization	RMSNorm (eps=1e-6)

Training

VRINDA-8B was trained on carefully curated datasets encompassing:

Conversational Data: Natural dialogue patterns in English and Hindi
Emotional Intelligence: Empathetic response generation
Cultural Content: Indian context, expressions, and cultural knowledge
Reasoning Tasks: Mathematical and logical problem-solving
Bilingual Corpus: Code-switching and translation capabilities

Intended Use

Recommended Applications

Personal AI assistants
Customer support chatbots (Hindi/English)
Educational tools
Content generation
Research and development

Out-of-Scope Use

Medical, legal, or financial advice
Critical decision-making systems
Generating harmful or misleading content
Any application violating the license terms

Limitations

May occasionally generate factually incorrect information
Performance may vary on highly specialized technical domains
Should not be used as a sole source for critical decisions
May reflect biases present in training data

License

This model is released under the VRINDA Proprietary License.

Permission	Status
Personal Use	✅ Allowed
Research Use	✅ Allowed
Commercial Use	❌ Requires License
Modification	❌ Prohibited
Redistribution	❌ Prohibited

See LICENSE for complete terms and conditions.

Try VRINDA

Contact

Organization	Ekoaham Dutivnasti
Business Inquiries	[email protected]
Website	ekoahamdutivnasti.com
VRINDA Web App	vrinda.ekoahamdutivnasti.com
Android App	Google Play Store
Privacy Policy	Privacy Policy

Citation

@misc{vrinda8b2024,
    title={VRINDA-8B: A Bilingual Large Language Model for English and Hindi},
    author={Ekoaham Dutivnasti},
    year={2024},
    publisher={HuggingFace},
    url={https://huggingface.co/ekoahamdutivnasti/VRINDA}
}

VRINDA-8B — Bridging Languages, Understanding Cultures

_{^{Built upon open-source foundation model architecture. We acknowledge the contributions of the open-source AI community.}}

Downloads last month: -

Safetensors

Model size

8B params

Tensor type

BF16