Instructions to use adityabhushannagar/Dhir-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use adityabhushannagar/Dhir-2B with Transformers:

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

pipe = pipeline("text-generation", model="adityabhushannagar/Dhir-2B", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("adityabhushannagar/Dhir-2B", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use adityabhushannagar/Dhir-2B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "adityabhushannagar/Dhir-2B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "adityabhushannagar/Dhir-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/adityabhushannagar/Dhir-2B

SGLang

How to use adityabhushannagar/Dhir-2B 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 "adityabhushannagar/Dhir-2B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "adityabhushannagar/Dhir-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "adityabhushannagar/Dhir-2B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "adityabhushannagar/Dhir-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use adityabhushannagar/Dhir-2B with Docker Model Runner:
```
docker model run hf.co/adityabhushannagar/Dhir-2B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

These checkpoints are the unofficial one to one mirror of the original checkpoints uploaded on AIKosh by the original author ZenteiQ on AIKosh

Please contact the original author for more information on the model. It was uploaded to HF by me for more easy access to the model

Dhīr 2B

A compact foundation model, designed and pretrained from scratch in India, with an in-house tokenizer built for Indian languages.

ZenteiQ · BrahmAI family · Apache 2.0

Overview

Dhīr 2B is the lightweight model in ZenteiQ's BrahmAI family. It was designed and pretrained from scratch by ZenteiQ's team in India, rather than adapted from an existing foundation model. It understands English and ten Indian languages, and is small enough to run close to the user: on a laptop, a workstation, an edge device, or inside an enterprise boundary where data must stay private.


Developer	ZenteiQ AiTech Innovations Pvt. Ltd., Bengaluru, India
Model type	Dense decoder-only Transformer, instruction-tuned (SFT followed by GRPO reinforcement learning)
Parameters	2.03B
Languages	English, Hindi, Gujarati, Tamil, Bengali, Kannada, Malayalam, Marathi, Odia, Punjabi, Telugu
Context length	16,384 tokens
Licence	Apache 2.0
Support	IndiaAI Mission, MeitY, under the Foundation Model initiative
Companion model	Dhīr 5B MoE (more capacity for complex enterprise workloads)

Why Dhīr 2B

Built in India, end to end. Model design, tokenizer, data pipeline, pretraining and post-training were carried out by ZenteiQ's team in India.
Strong Indian-language fit. The lowest Indic bits-per-byte of the 31 open models in our comparison, and the best Indic ↔ English translation among the compared open models.
Answers from your documents. Leads the compared open models on question answering over supplied Indian-language passages, the core skill of RAG applications.
Native tool and function calling. Calls tools and APIs in a native tool format (BFCL v3 accuracy 73.4), so it can power assistants that act on documents, apps and services.
Private by design. 3.8 GB in BF16 and 1.1 GB at 4-bit, so it can run on-device or on-premises without a large cloud model.

Architecture

Component	Configuration
Layers	28
Hidden dimension	2048
Attention	Grouped-Query Attention: 16 query heads, 8 key/value heads, head dim 128
Feed-forward	Gated FFN, intermediate dimension 8192
Positional encoding	Rotary Position Embeddings (RoPE)
Normalisation	RMSNorm
Attention kernel	Flash Attention
Vocabulary	131,072 (in-house tokenizer for Indian languages)
Context	16k tokens
Precision	bfloat16

Training stack: MaxText (JAX) on Google TPU v6e. Checkpoints are trained in Orbax format and converted to Hugging Face format for release.

Quick start

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "zenteiq/Dhir-2B"  # confirm final repository name
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True
)

messages = [{"role": "user", "content": "भारत की राजधानी के बारे में दो वाक्य लिखिए।"}]
inputs = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)

output = model.generate(inputs, max_new_tokens=256, do_sample=False)
print(tokenizer.decode(output[0][inputs.shape[-1]:], skip_special_tokens=True))

trust_remote_code=True is required. Dhīr ships its architecture as custom code (modeling_dhir.py and configuration_dhir.py, included in this repository), so Transformers must be allowed to run it when loading the model. The same applies to other tools that load the model through Transformers: for example, pass --trust-remote-code to vllm serve.

Intended use

Use	Notes
Indic ↔ English translation	Drafting and everyday communication. Human review is recommended for published or legal content.
Question answering over supplied documents (RAG reader)	Answers from a passage you provide, in English and Indian languages.
Indian-language text understanding	Reading comprehension and summarisation.
Base model for Indic fine-tuning	A strong starting point for organisations building their own Indian-language model.
Basic Python assistance	Short, self-contained functions.
Tool and function calling	Native function calling for assistants that call tools and APIs.
Private, offline and edge AI	Personal assistants and privacy-sensitive applications on local hardware.

Evaluation

All figures are for the Dhīr 2B release checkpoint, from ZenteiQ's internal benchmark suite (14 October 2026). Ranks are against open models run on the same benchmark and harness: 17 instruction-tuned models for generative tasks, 31 models (base and instruct) for log-likelihood tasks.

Capability	Benchmark	Dhīr 2B	Next best	Rank
Translation, 24 Indic ↔ English directions (chrF)	IndicGenBench FLORES	51.4	46.0	#1 of 17
Document QA, 12 languages (token F1)	XQuAD-IN	63.5	61.3	#1 of 17
Document QA, 11 languages (token F1)	IndicQA	63.4	60.0	#1 of 17
Cross-lingual QA, 12 languages (token F1)	XorQA-IN	33.7	26.1	#1 of 17
Indic language modelling, 14 sets (bits-per-byte, lower is better)	Held-out Indic text	0.507	0.625	#1 of 31
Reading comprehension, 13 languages (accuracy)	Belebele	58.8	54.2 (best ≤3B peer)	Leads models up to 3B
Instruction following, English (strict)	IFEval	66.5		Ahead of Gemma-2-2B-it (55.5)
Python (pass@1)	HumanEval / MBPP	45.1 / 40.2		HumanEval ahead of Gemma-3-4B-it (26.2)
Grade-school maths, English	GSM8K	58.0		Basic maths reasoning
Function calling, native tool format (accuracy)	BFCL v3, AST checker	73.4		Simple calls 88.3, non-live 83.3

Translation, QA and language-modelling leads are outside the 95% confidence interval of the difference. Document QA is measured when the answer is in a supplied passage. Larger 4B-class models score higher than Dhīr 2B on Belebele; the lead is within the ≤3B class. Function calling is scored in the model's native tool format (BFCL v3, AST checker); native irrelevance, the rate of correctly not calling a tool when none is needed, is 53.8.

Deployment. Weights take 3.78 GB (BF16), 2.01 GB (8-bit) and 1.13 GB (4-bit). Deployment checks passed on NVIDIA L4 and RTX Pro 6000 GPUs, on CPU with 8-bit and 4-bit quantisation, and at 4K, 8K and 16K context. At 128 concurrent users the model sustains about 2,945 output tokens/s in aggregate, around 27 tokens/s per user.

Evaluation notes

Results come from ZenteiQ's internal evaluation suite on the release checkpoint. Independent reproduction is part of the release process, and the full methodology will appear in the Dhīr Technical Report.
Every model in a comparison was scored with the same harness, prompt and metric.
Each claim is tied to the benchmark that supports it. We do not extend a result beyond what the benchmark measures.

Precautions

Dhīr 2B is a compact model, and we recommend the following for responsible deployment:

Expert validation. Legal, medical and financial uses should be reviewed by a qualified professional. The model is not intended to give such advice on its own.
Grounding. Answers are most reliable when the relevant text is supplied in the prompt, as in RAG applications.
Language coverage. Supported languages are English and the ten Indian languages listed above.
Reasoning mode. Thinking mode is not supported in this release.
Human oversight. Like all language models, Dhīr 2B can produce incorrect or biased output. Keep a human in the loop for consequential decisions.

Safety

Dhīr 2B has been evaluated for safety behaviour, including probing with the NVIDIA Garak vulnerability scanner across jailbreak, prompt-injection and harmful-request categories. Detailed results will be published with the technical report. Deployers should add application-level safeguards suited to their use case.

Training

Pretraining: from scratch on Google TPU v6e using MaxText (JAX).
Tokenizer: in-house, designed for Indian languages (vocabulary 131,072).
Post-training: supervised fine-tuning followed by GRPO reinforcement learning.

Data composition, training compute and methodology will be provided in the Dhīr Technical Report.

Availability

Released under the Apache 2.0 licence through AIKosh, Hugging Face and ZenteiQ.

Acknowledgement

ZenteiQ acknowledges the support of the IndiaAI Mission, Ministry of Electronics and Information Technology (MeitY), under the Foundation Model initiative.

Citation

@misc{zenteiq2026dhir,
  title        = {Dh\={\i}r : Sovereign Compact Foundation Models },
  author       = {{ZenteiQ AiTech Innovations Pvt. Ltd.}},
  year         = {2026},
  howpublished = {Technical report forthcoming}
}

Contact

ZenteiQ AiTech Innovations Pvt. Ltd. · Vista Pixel, Jakkuru, Bangalore 560092, India research@zenteiq.ai · www.zenteiq.ai

Downloads last month: 213

Safetensors

Model size

2B params

Tensor type

BF16

Model tree for adityabhushannagar/Dhir-2B

Quantizations

1 model