aosm/qwen2-vl-7b-medical-vqa-tr

🔬 Fine-tuned Qwen2-VL-7B model on Turkish medical visual question answering (VQA) tasks using Unsloth.

👨‍💻 Developed by: aosm
🧠 Fine-tuned from: unsloth/qwen2-vl-7b-instruct-unsloth-bnb-4bit
📄 License: apache-2.0
📊 Domain: Radiology, medical imaging, Turkish NLP
🖼️ Format: OpenChat-style JSON with image & text pairs

🧪 Training Details

Attribute	Value
Dataset	`aosm/turkish-medical-vqa-evaluated`
Examples	2,812
Epochs	5
Total steps	~140
Batch size	50 × 2 (accumulation)
Optimizer	AdamW 8-bit
LR scheduler	Cosine
Max sequence length	384
Vision modality	Enabled
Text language	Turkish

📚 Example Format

[
  {
    "role": "user",
    "content": [
      {"type": "text", "text": "Kitle hangi bölgede bulunur?"},
      {"type": "image", "image": "synpic53867.jpg"}
    ]
  },
  {
    "role": "assistant",
    "content": [{"type": "text", "text": "suprasellar"}]
  }
]


base_model: unsloth/qwen2-vl-7b-instruct-unsloth-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2_vl
- trl
license: apache-2.0
language:
- en
---

# Uploaded  model

- **Developed by:** aosm
- **License:** apache-2.0
- **Finetuned from model :** unsloth/qwen2-vl-7b-instruct-unsloth-bnb-4bit

This qwen2_vl model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.

[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)