Instructions to use Tamil-ai/tamil-qwen25-7b-instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Tamil-ai/tamil-qwen25-7b-instruct with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Tamil-ai/tamil-qwen25-7b-instruct")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Tamil-ai/tamil-qwen25-7b-instruct")
model = AutoModelForCausalLM.from_pretrained("Tamil-ai/tamil-qwen25-7b-instruct")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Tamil-ai/tamil-qwen25-7b-instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Tamil-ai/tamil-qwen25-7b-instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Tamil-ai/tamil-qwen25-7b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Tamil-ai/tamil-qwen25-7b-instruct

SGLang

How to use Tamil-ai/tamil-qwen25-7b-instruct with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Tamil-ai/tamil-qwen25-7b-instruct" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Tamil-ai/tamil-qwen25-7b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Tamil-ai/tamil-qwen25-7b-instruct" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Tamil-ai/tamil-qwen25-7b-instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Tamil-ai/tamil-qwen25-7b-instruct with Docker Model Runner:
```
docker model run hf.co/Tamil-ai/tamil-qwen25-7b-instruct
```

Tamil-Qwen2.5-7B-Instruct

A Tamil-specialized instruction-tuned LLM built on Qwen2.5-7B-Instruct using QLoRA fine-tuning on 150K deduplicated Tamil instruction pairs.

Paper: "A Thousand Language Problem: Morphological Understanding in Linguistic AI"

Model Details

Property	Value
Base model	Qwen/Qwen2.5-7B-Instruct
Parameters	7.6B
Method	QLoRA (r=64, alpha=128, dropout=0.05)
Training data	150K deduplicated Tamil instruction-response pairs
Tokenizer efficiency	4.62x ratio (best among tested models for Tamil)
Compute	RunPod RTX 5090, ~$5 total cost
Sequence length	1024
Batch size	32 (effective)
Epochs	1

Training Data

150,000 deduplicated instruction-response pairs from 5 Tamil datasets:

Tamil Alpaca
Tamil Orca
Tamil Dolly
Tamil-ai/samacheer-kalvi-tamil (morphological drills + grammar QA)
Additional Tamil instruction sets

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Tamil-ai/tamil-qwen25-7b-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are a helpful Tamil language assistant."},
    {"role": "user", "content": "வீடு என்ற சொல்லின் வேற்றுமை வடிவங்களைக் கூறுக."},
]

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=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

4-bit Quantized (for limited VRAM)

from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig

model = AutoModelForCausalLM.from_pretrained(
    "Tamil-ai/tamil-qwen25-7b-instruct",
    quantization_config=BitsAndBytesConfig(load_in_4bit=True),
    device_map="auto",
)

Why Qwen2.5?

Tokenizer analysis across 6 base models showed Qwen2.5 has the best Tamil tokenization efficiency:

Model	Tamil Token Ratio	Verdict
Qwen2.5	4.62x	Best for Tamil
Llama 3.1	5.8x
Gemma 2	6.1x
Mistral	7.2x
Falcon	10.5x	Worst

Lower ratio = fewer tokens per Tamil word = more efficient training and inference.

Intended Use

Tamil question answering and instruction following
Tamil morphological analysis
Tamil grammar and linguistics tasks
Research on low-resource language LLMs

Limitations

Trained primarily on instructional Tamil; may underperform on colloquial/slang
Morphological accuracy varies by category (see benchmark results)
English capabilities may degrade compared to base Qwen2.5

Citation

@misc{tamilai2026,
  title={A Thousand Language Problem: Morphological Understanding in Linguistic AI},
  author={Tamil-AI},
  year={2026},
  publisher={HuggingFace},
  url={https://huggingface.co/Tamil-ai/tamil-qwen25-7b-instruct}
}

Downloads last month: 7

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for Tamil-ai/tamil-qwen25-7b-instruct

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Finetuned

(2627)

this model