---
license: gemma
library_name: transformers
pipeline_tag: image-text-to-text
base_model: google/gemma-3-4B-it
tags:
  - gemma3
  - bg
  - bulgarian
language:
  - bg
  - en
---

# BgGPT-Gemma-3-4B-IT

BgGPT 3.0 is a series of Bulgarian-adapted LLMs based on Gemma 3, developed by [INSAIT](https://insait.ai). Available in 4B, 12B and 27B sizes.

**Blog post**: [BgGPT-3 Release](https://models.bggpt.ai/blog/bggpt-3-release-en)

### Key improvements over BgGPT 2.0

1. **Vision-language understanding** — The models understand both text and images within the same context.
2. **Instruction-following** — Trained on a broader range of tasks, multi-turn conversations, complex instructions, and system prompts.
3. **Longer context** — Effective context of 131k tokens for longer conversations and complex instructions.
4. **Updated knowledge cut-off** — Pretraining data up to May 2025, instruction fine-tuning up to October 2025.

<p align="center">
  <img src="performance_generative_tasks.png" alt="Performance on Generative Tasks" width="600"/>
</p>

*Figure 1: Performance on Generative Tasks (TriviaQA, GSM8k, IFEval, BigBenchHard)*

## Usage

### Transformers

```python
from transformers import AutoProcessor, Gemma3ForConditionalGeneration
import torch

model_id = "INSAIT-Institute/BgGPT-Gemma-3-4B-IT"

processor = AutoProcessor.from_pretrained(model_id)
model = Gemma3ForConditionalGeneration.from_pretrained(
    model_id, device_map="auto"
).eval()

messages = [
    {
        "role": "user",
        "content": [{"type": "text", "text": "Кога е основан Софийският университет?"}],
    },
]

inputs = processor.apply_chat_template(
    messages, add_generation_prompt=True, tokenize=True,
    return_dict=True, return_tensors="pt"
).to(model.device, dtype=torch.bfloat16)

input_len = inputs["input_ids"].shape[-1]

with torch.inference_mode():
    generation = model.generate(**inputs, max_new_tokens=512, do_sample=True, temperature=0.2)
    generation = generation[0][input_len:]

print(processor.decode(generation, skip_special_tokens=True))
```

#### With an image

```python
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
            {"type": "text", "text": "Опиши какво виждаш на изображението."},
        ],
    },
]

inputs = processor.apply_chat_template(
    messages, add_generation_prompt=True, tokenize=True,
    return_dict=True, return_tensors="pt"
).to(model.device, dtype=torch.bfloat16)

input_len = inputs["input_ids"].shape[-1]

with torch.inference_mode():
    generation = model.generate(**inputs, max_new_tokens=512, do_sample=True, temperature=0.2)
    generation = generation[0][input_len:]

print(processor.decode(generation, skip_special_tokens=True))
```

### vLLM

```python
from vllm import LLM, SamplingParams

llm = LLM(model="INSAIT-Institute/BgGPT-Gemma-3-4B-IT")
params = SamplingParams(max_tokens=512, temperature=0.2)

messages = [
    {
        "role": "user",
        "content": [{"type": "text", "text": "Кога е основан Софийският университет?"}],
    },
]

outputs = llm.chat(messages, sampling_params=params)
print(outputs[0].outputs[0].text)
```

Serving with the OpenAI-compatible API:

```bash
vllm serve INSAIT-Institute/BgGPT-Gemma-3-4B-IT
```

### vLLM with FP8 dynamic quantization

Load the model in FP8 at runtime for ~2x memory reduction with minimal quality loss — no separate quantized checkpoint needed:

```python
from vllm import LLM, SamplingParams

llm = LLM(
    model="INSAIT-Institute/BgGPT-Gemma-3-4B-IT",
    quantization="fp8",
)
params = SamplingParams(max_tokens=512, temperature=0.2)

messages = [
    {
        "role": "user",
        "content": [{"type": "text", "text": "Кога е основан Софийският университет?"}],
    },
]

outputs = llm.chat(messages, sampling_params=params)
print(outputs[0].outputs[0].text)
```

```bash
vllm serve INSAIT-Institute/BgGPT-Gemma-3-4B-IT --quantization fp8
```

> Requires a GPU with compute capability >= 8.9 (H100, H200, RTX 4090).

<!-- TODO: Add benchmarks, evaluation, citation -->

## License

BgGPT-Gemma-3-4B-IT is distributed under the [Gemma Terms of Use](https://ai.google.dev/gemma/terms).