Instructions to use retro56/gemma3-4b-bengali-multimodal-persona with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use retro56/gemma3-4b-bengali-multimodal-persona with Transformers:

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

pipe = pipeline("text-generation", model="retro56/gemma3-4b-bengali-multimodal-persona")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("retro56/gemma3-4b-bengali-multimodal-persona", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use retro56/gemma3-4b-bengali-multimodal-persona with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "retro56/gemma3-4b-bengali-multimodal-persona"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "retro56/gemma3-4b-bengali-multimodal-persona",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/retro56/gemma3-4b-bengali-multimodal-persona

SGLang

How to use retro56/gemma3-4b-bengali-multimodal-persona 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 "retro56/gemma3-4b-bengali-multimodal-persona" \
    --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": "retro56/gemma3-4b-bengali-multimodal-persona",
		"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 "retro56/gemma3-4b-bengali-multimodal-persona" \
        --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": "retro56/gemma3-4b-bengali-multimodal-persona",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use retro56/gemma3-4b-bengali-multimodal-persona with Docker Model Runner:
```
docker model run hf.co/retro56/gemma3-4b-bengali-multimodal-persona
```

retro56 commited on May 31, 2025

Commit

68ac7c8

verified ·

1 Parent(s): 21350ea

Update README.md

Browse files

Files changed (1) hide show

README.md +10 -11

README.md CHANGED Viewed

@@ -17,8 +17,7 @@ tags:
 - 4bit-quantization
 datasets:
 - iamshnoo/alpaca-cleaned-bengali
-- cfilt/iitb-english-bengali
-base_model: google/gemma-2-27b-it
 model_type: gemma2
 ---
@@ -28,7 +27,7 @@ model_type: gemma2
 ## Model Description
-This model is a fine-tuned version of [google/gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it) specifically optimized for Bengali language conversations and multimodal AI persona applications. The model has been trained to provide natural, helpful responses in Bengali and can be integrated with voice synthesis for complete multimodal AI experiences.
 ### Key Features
@@ -48,7 +47,7 @@ This model is a fine-tuned version of [google/gemma-2-27b-it](https://huggingfac
 - **Total Examples**: ~8,000 high-quality Bengali examples
 ### Training Configuration
-- **Base Model**: google/gemma-2-27b-it
 - **Fine-tuning Method**: LoRA (Low-Rank Adaptation)
 - **Quantization**: 4-bit using BitsAndBytesConfig
 - **LoRA Rank**: 16
@@ -76,7 +75,7 @@ import torch
 # Load the model and tokenizer
 base_model = AutoModelForCausalLM.from_pretrained(
-    "google/gemma-2-27b-it",
     torch_dtype=torch.float16,
     device_map="auto"
 )
@@ -177,23 +176,23 @@ See the [complete notebook](https://github.com/your-repo/gemma3-bengali-multimod
 ## Model Card Authors
-Created by the Bengali AI research team for advancing Bengali language AI capabilities.
 ## Citation
 ```bibtex
 @misc{gemma2-bengali-multimodal,
-  title={Gemma 2 27B Bengali Multimodal Persona},
-  author={Bengali AI Research Team},
-  year={2024},
   url={https://huggingface.co/retro56/gemma3-4b-bengali-multimodal-persona}
 }
 ```
 ## License
-This model is licensed under the Gemma License. See the [original model](https://huggingface.co/google/gemma-2-27b-it) for complete license terms.
 ---
-**Built with ❤️ for the Bengali AI community**

 - 4bit-quantization
 datasets:
 - iamshnoo/alpaca-cleaned-bengali
+base_model: google/gemma-2-4b-it
 model_type: gemma2
 ---
 ## Model Description
+This model is a fine-tuned version of [google/gemma-2-4b-it](https://huggingface.co/google/gemma-2-4b-it) specifically optimized for Bengali language conversations and multimodal AI persona applications. The model has been trained to provide natural, helpful responses in Bengali and can be integrated with voice synthesis for complete multimodal AI experiences.
 ### Key Features
 - **Total Examples**: ~8,000 high-quality Bengali examples
 ### Training Configuration
+- **Base Model**: google/gemma-2-4b-it
 - **Fine-tuning Method**: LoRA (Low-Rank Adaptation)
 - **Quantization**: 4-bit using BitsAndBytesConfig
 - **LoRA Rank**: 16
 # Load the model and tokenizer
 base_model = AutoModelForCausalLM.from_pretrained(
+    "google/gemma-2-4b-it",
     torch_dtype=torch.float16,
     device_map="auto"
 )
 ## Model Card Authors
+Created by the Personify research team for advancing Bengali language AI capabilities.
 ## Citation
 ```bibtex
 @misc{gemma2-bengali-multimodal,
+  title={Gemma 2 4B Bengali Multimodal Persona},
+  author={Personify Research Team},
+  year={2025},
   url={https://huggingface.co/retro56/gemma3-4b-bengali-multimodal-persona}
 }
 ```
 ## License
+This model is licensed under the Gemma License. See the [original model](https://huggingface.co/google/gemma-2-4b-it) for complete license terms.
 ---
+**Built with ❤️ for the Bengali AI community**