Instructions to use ahmed-farhanur-rashid/bangla-gamba with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ahmed-farhanur-rashid/bangla-gamba with Transformers:

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

pipe = pipeline("text-generation", model="ahmed-farhanur-rashid/bangla-gamba", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("ahmed-farhanur-rashid/bangla-gamba", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ahmed-farhanur-rashid/bangla-gamba with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ahmed-farhanur-rashid/bangla-gamba

SGLang

How to use ahmed-farhanur-rashid/bangla-gamba 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 "ahmed-farhanur-rashid/bangla-gamba" \
    --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": "ahmed-farhanur-rashid/bangla-gamba",
		"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 "ahmed-farhanur-rashid/bangla-gamba" \
        --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": "ahmed-farhanur-rashid/bangla-gamba",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ahmed-farhanur-rashid/bangla-gamba with Docker Model Runner:
```
docker model run hf.co/ahmed-farhanur-rashid/bangla-gamba
```

BanglaGamba

Huggingface compatible version of Bangla-Gamba Base model.

Model Details

Property	Value
Parameters	~200M
Context Length	2048
Architecture	Hybrid Mamba-3 + GQA (1:1 interleaved)
Primary Language	Bengali
Secondary Language	English
Training Tokens	~9.62B
Training Dataset	`ahmed-farhanur-rashid/bn-foundational-pretrain-corpus`
License	CC BY-NC-SA 4.0

Resources

GitHub Repository: ahmed-farhanur-rashid/bangla-gamba
Training Dataset: datasets/ahmed-farhanur-rashid/bn-foundational-pretrain-corpus

Requirements

This model uses a custom architecture and tokenizer implementation. Loading requires enabling trust_remote_code=True.

Install the required dependencies:

pip install transformers torch mamba-ssm causal-conv1d bnunicodenormalizer

Note: The model was trained on text normalized using bnunicodenormalizer in data pipeline. Running without it may significantly degrade generation quality.

Usage

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_id = "ahmed-farhanur-rashid/bangla-gamba"
tokenizer = AutoTokenizer.from_pretrained(
    model_id,
    trust_remote_code=True,
)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    trust_remote_code=True,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
model.eval()
prompt = "বাংলাদেশের রাজধানী"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=50,
    do_sample=True,
    temperature=0.7,
    top_p=0.9,
)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Architecture

Unlike conventional Transformer-only language models, BanglaGamba combines state-space modeling with attention mechanisms:

Mamba-3 (State Space Model): Efficient linear-time sequential modeling and long-range state recurrence without quadratic attention overhead.
Grouped Query Attention (GQA): Improves inference efficiency through optimized key-value caching, enhanced with per-head QK-Norm and RoPE.
SwiGLU FFN: Interleaved in every block for non-linear representation capacity.

This hybrid design aims to balance computational efficiency with strong language modeling performance.

Related Models

BanglaGamba is part of a family of Bengali foundation language models.

Model	Architecture	Description
BanglaGamba	Mamba-3 + GQA	Hybrid state-space and GQA architecture optimized for efficient Bengali language modeling.
BanglaGSG	GDN + SWA + GQA	Sibling hybrid architecture trained on the same corpus.

BanglaGSG: tasmin-jahan/bangla-gsg

Limitations

The model expects text normalized using bnunicodenormalizer, consistent with the preprocessing pipeline used during training.
Loading requires execution of custom Python modules (modeling_banglagamba.py, configuration_banglagamba.py, tokenization_banglagamba.py) via trust_remote_code=True.
While primarily trained for Bengali, English support is intended mainly for multilingual understanding and translation-related capabilities.
As with other large language models, outputs may occasionally be inaccurate or reflect biases present in the training data.

Citation

If you use BanglaGamba in your research, please cite the model:

@misc{banglagamba2026,
  title        = {BanglaGamba},
  author       = {Ahmed Farhanur Rashid},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/ahmed-farhanur-rashid/bangla-gamba}}
}

Downloads last month: 807

Model tree for ahmed-farhanur-rashid/bangla-gamba

Base model

ahmed-farhanur-rashid/bangla-gamba-base

Finetuned

(1)

this model

ahmed-farhanur-rashid
/

bangla-gamba