---
library_name: peft
base_model: google/gemma-3-4b-it
language:
- nso
- en
tags:
- translation
- african-languages
- scientific-translation
- afriscience-mt
- lora
- peft
- gemma
license: apache-2.0
pipeline_tag: translation
model-index:
- name: gemma_3_4b_it-lora-r64-nso-eng
  results:
  - task:
      type: translation
    metrics:
    - name: BLEU (test)
      type: bleu
      value: 39.14
    - name: chrF (test)
      type: chrf
      value: 59.10
    - name: SSA-COMET (test)
      type: comet
      value: 65.56
---

# gemma_3_4b_it-lora-r64-nso-eng

[![Model on HF](https://huggingface.co/datasets/huggingface/badges/raw/main/model-on-hf-sm.svg)](https://huggingface.co/dsfsi/gemma_3_4b_it-lora-r64-nso-eng)

This is a **LoRA adapter** for the AfriScience-MT project, enabling efficient scientific machine translation for African languages.

## Adapter Description

| Property | Value |
|----------|-------|
| **Base Model** | [google/gemma-3-4b-it](https://huggingface.co/google/gemma-3-4b-it) |
| **Translation Direction** | Northern Sotho → English |
| **LoRA Rank (r)** | 64 |
| **LoRA Alpha** | 128 |
| **Training Method** | QLoRA (4-bit quantization) |
| **Domain** | Scientific/Academic texts |

### Why LoRA?

LoRA (Low-Rank Adaptation) enables efficient fine-tuning by training only a small number of additional parameters. This adapter adds only **~32.0M parameters** to the base model while achieving strong translation performance.

## Evaluation Results

Performance on the AfriScience-MT test set:

| Split | BLEU | chrF | SSA-COMET |
|-------|------|------|-----------|
| Validation | 43.93 | 63.31 | 66.87 |
| **Test** | **39.14** | **59.10** | **65.56** |

**Metrics explanation:**
- **BLEU**: Measures n-gram overlap with reference translations (0-100, higher is better)
- **chrF**: Character-level F-score, robust for morphologically rich languages (0-100, higher is better)
- **SSA-COMET**: Neural metric trained for Sub-Saharan African languages, shown as percentage (0-100, higher is better) ([McGill-NLP/ssa-comet-stl](https://huggingface.co/McGill-NLP/ssa-comet-stl))

## Usage

### Quick Start

```python
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from peft import PeftModel
import torch

# Configure 4-bit quantization (recommended for memory efficiency)
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_compute_dtype=torch.bfloat16,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_use_double_quant=True,
)

# Load base model
base_model = AutoModelForCausalLM.from_pretrained(
    "google/gemma-3-4b-it",
    quantization_config=bnb_config,
    device_map="auto",
    torch_dtype=torch.bfloat16,
)
tokenizer = AutoTokenizer.from_pretrained("google/gemma-3-4b-it")

# Load LoRA adapter
adapter_name = "dsfsi/gemma_3_4b_it-lora-r64-nso-eng"
model = PeftModel.from_pretrained(base_model, adapter_name)
model.eval()

# Prepare translation prompt
source_text = "Climate change significantly impacts agricultural productivity in sub-Saharan Africa."
instruction = "Translate the following Northern Sotho scientific text to English."

# Format for Gemma chat template
messages = [{"role": "user", "content": f"{instruction}\n\n{source_text}"}]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

# Generate translation
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=256,
        num_beams=5,
        early_stopping=True,
        pad_token_id=tokenizer.pad_token_id,
    )

# Decode only the generated part
generated = outputs[0][inputs["input_ids"].shape[1]:]
translation = tokenizer.decode(generated, skip_special_tokens=True)
print(translation)
```

### Without Quantization (Full Precision)

```python
# For GPUs with sufficient memory (>24GB for larger models)
base_model = AutoModelForCausalLM.from_pretrained(
    "google/gemma-3-4b-it",
    device_map="auto",
    torch_dtype=torch.bfloat16,
)
model = PeftModel.from_pretrained(base_model, "dsfsi/gemma_3_4b_it-lora-r64-nso-eng")
```

## Training Details

### Hyperparameters

| Parameter | Value |
|-----------|-------|
| LoRA Rank (r) | 64 |
| LoRA Alpha | 128 |
| LoRA Dropout | 0.05 |
| Target Modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
| Epochs | 3 |
| Batch Size | 2 |
| Learning Rate | 2e-04 |
| Max Sequence Length | 512 |
| Gradient Accumulation | 4 |

### Hardware Requirements

| Configuration | VRAM Required |
|---------------|---------------|
| 4-bit (QLoRA) | ~8-12 GB |
| 8-bit | ~16-20 GB |
| Full precision | ~24-40 GB |


## Reproducibility

To reproduce this adapter:

```bash
# Clone the AfriScience-MT repository
git clone https://github.com/afriscience-mt/afriscience-mt.git
cd afriscience-mt

# Install dependencies
pip install -r requirements.txt

# Run LoRA training
python -m afriscience_mt.scripts.run_lora_training \
    --data_dir ./data \
    --source_lang nso \
    --target_lang eng \
    --model_name google/gemma-3-4b-it \
    --model_type gemma \
    --lora_rank 64 \
    --output_dir ./output \
    --num_epochs 3 \
    --batch_size 4 \
    --load_in_4bit
```

## Limitations

- **Domain Specificity**: Optimized for scientific/academic texts; may underperform on casual or colloquial language.
- **Language Direction**: Only supports Northern Sotho → English translation.
- **Base Model Required**: Must be used with the [google/gemma-3-4b-it](https://huggingface.co/google/gemma-3-4b-it) base model.
- **Context Length**: Maximum context is model-dependent; longer texts should be chunked.

## Citation

If you use this model, please cite the AfriScience-MT paper ([arXiv:2605.29741](https://arxiv.org/abs/2605.29741)):

```bibtex
@article{abdulmumin2026afriscience,
  title   = {AfriScience-MT: Towards Decolonizing Science in Africa through Text Translation},
  author  = {Abdulmumin, Idris and Gwadabe, Tajuddeen and Muhammad, Shamsuddeen Hassan and Adelani, David Ifeoluwa and Khalo, Nomonde and Ahmad, Ibrahim Said and Modupe, Abiodun and Mumm, Anina and Biyela, Sibusiso and Rabie, Michelle and Havemann, Johanna and Rei, Marek and Abbott, Jade and Marivate, Vukosi},
  journal = {arXiv preprint arXiv:2605.29741},
  year    = {2026},
  url     = {https://arxiv.org/abs/2605.29741}
}
```

## License

This adapter is released under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0).

## Acknowledgments

- Base model: [google/gemma-3-4b-it](https://huggingface.co/google/gemma-3-4b-it)
- LoRA implementation: [PEFT](https://github.com/huggingface/peft)
- Evaluation: [SSA-COMET](https://huggingface.co/McGill-NLP/ssa-comet-stl) for African language assessment