Tamil AI DJ Radio - Merged Model

Format: Merged (base + adapter fused)

Fine-tuned Qwen 2.5-0.5B model for generating energetic Tanglish (Tamil-English mix) radio DJ commentary. This is the merged model with LoRA adapter fused into the base model for easy deployment.

Model Details

  • Base Model: Qwen/Qwen2.5-0.5B-Instruct
  • Parameters: 494M (0.5B)
  • Precision: FP16
  • Training: Fine-tuned with LoRA on 5,027 DJ commentary examples
  • Language: Tanglish (Tamil-English code-mixed)
  • Use Case: Radio DJ commentary generation

Other Formats

This model is available in three formats:

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_name = "felixmanojh/DJ-AI-Radio"

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    torch_dtype=torch.float16
)

messages = [
    {"role": "system", "content": "You are a Tamil AI radio DJ who speaks energetic Tanglish. Create engaging commentary."},
    {"role": "user", "content": "Hype up a high-energy dance track for weekend party"}
]

text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")

outputs = model.generate(
    **inputs,
    max_new_tokens=150,
    temperature=0.8,
    top_p=0.9,
    do_sample=True
)

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

Demo

Try the live demo: Tamil AI DJ Radio Space

Training Data

Fine-tuned on 5,027 diverse DJ commentary examples covering:

  • Party & dance vibes
  • Chill & relaxation
  • Workout motivation
  • Study focus music
  • Late night romance
  • Morning energy

License

Apache 2.0

Citation

@misc{tamil-ai-dj-radio,
  author = {Felix Manojh},
  title = {Tamil AI DJ Radio - Tanglish Radio DJ Commentary Generator},
  year = {2026},
  publisher = {HuggingFace},
  url = {https://huggingface.co/felixmanojh/DJ-AI-Radio}
}

Built with ❤️ for the Tamil-speaking community

Downloads last month
39
Safetensors
Model size
0.5B params
Tensor type
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Space using felixmanojh/DJ-AI-Radio 1