bharatgenai / Param-1

Param-1

BharatGen introduces Param-1 , a bilingual language model pretrained from scratch on English and Hindi. With 2.9 billion parameters, it serves as a powerful foundational model for text completion.

Param-1 outperforms leading models like LLaMA-3.2B , Gemma-2B , Granite-2B , and Granite-3B on various standard benchmarks.

This early release is equipped with inference support via NVIDIA NeMo .

🚀 Model Inference

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load tokenizer and model
model_name = "bharatgenai/Param-1"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=False)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    torch_dtype=torch.bfloat16 if torch.cuda.is_available() else torch.bfloat32,
    device_map="auto"
)

prompt = "Your promt here."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

# --- Generate output ---
with torch.no_grad():
    output = model.generate(
        **inputs,
        max_new_tokens=300,
        do_sample=True,
        top_k=50,
        top_p=0.95,
        temperature=0.6,
        eos_token_id=tokenizer.eos_token_id
    )

generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
print("Generated Text:\n", generated_text)

📊 Benchmarks

Notes:

• PT : Pre-Trained
• en-hi : English-Hindi
• Pre-trained on 5 Trillion tokens

🧠 Model Architecture

• Hidden size: 2048
• Intermediate size: 7168
• Number of attention heads: 16
• Number of hidden layers: 32
• Number of key-value heads: 8
• Maximum position embeddings: 2048
• Activation function: SiLU
• Positional embeddings: Rotary (RoPE) with rope_theta=10000.0
• Attention: Grouped-query attention
• Precision: bf16-mixed

🏗️ Training Details

• Training Infrastructure : Yotta’s Shakti Cloud
• Hardware : NVIDIA H100 – 512 GPUs
• Framework : NVIDIA NeMo

📜 License

This model is released under the BharatGen non-commercial license .
Please refer to the LICENSE for terms and conditions.