Fine tunned PaliGemma vision-language models using the ScienceQA dataset for visual question answering.
-
Updated
Oct 23, 2024 - Jupyter Notebook
#
scienceqa
Here are 6 public repositories matching this topic...
Mitigating positional bias in LLaVA 1.5 (7B) on ScienceQA via per-head activation steering & PCA.
-
Updated
May 17, 2025 - Jupyter Notebook
Vision-Language Model fine-tuning and evaluation on ScienceQA using Qwen3.5, LoRA, and PEFT.
machine-learning computer-vision deep-learning transformers pytorch artificial-intelligence lora vlm fine-tuning multimodal huggingface large-language-models vision-language-model qwen qwen-vl multimodal-ai scienceqa
-
Updated
Jul 15, 2026 - Jupyter Notebook
QI-Budget: Quality-aware inference budgeting via dynamic visual-token routing for efficient VLM reasoning.
mlx efficient-inference vision-language-model multimodal-reasoning scienceqa qwen2-5-vl visual-token
-
Updated
Aug 23, 2026 - Python
Efficient Vision-Language Model inference and LoRA fine-tuning on ScienceQA, with dynamic resolution, visual-token compression, latency analysis, and accuracy–efficiency trade-offs using Qwen3.5-0.8B.
transformers pytorch lora dynamic-resolution vlm fine-tuning multimodal-learning peft multimodal huggingface model-efficiency vision-language-model qwen scienceqa qwen3 visual-tokens
-
Updated
Aug 19, 2026 - Jupyter Notebook
Add this topic to your repo
To associate your repository with the scienceqa topic, visit your repo's landing page and select "manage topics."