ModelListedExternal

Paligemma Clevr Cogen Qlora

This repository contains the QLoRA adapters for the google/paligemma-3b-pt-224 model, fine-tuned for a Visual Question Answering (VQA) task on the leonardPKU/clevrcogenatrain dataset.

Sourcehuggingface.co/tahamajs/paligemma-clevr-cogen-qlora/tree/b48847caa189a367a80eef5b9100bb9ff6a7c527

KindOtherBase modelgoogle/paligemma-3b-pt-224VersionAdapter modelCgrade

Model source

Kind: Other
File size: 1.35 KB–4.64 GB
Base model: google/paligemma-3b-pt-224
Version: vAdapter model
Updated: Oct 23, 2026
Source: Hugging Face

Source description

QLoRA Fine-tuned PaliGemma-3B for Visual Reasoning on CLEVR-CoGen

This repository contains the QLoRA adapters for the google/paligemma-3b-pt-224 model, fine-tuned for a Visual Question Answering (VQA) task on the leonardPKU/clevr_cogen_a_train dataset.

This fine-tuned model demonstrates significantly improved performance on questions requiring spatial and logical reasoning about complex scenes with multiple objects compared to the base PaliGemma model. The use of QLoRA (4-bit quantization) makes it possible to run and train this powerful model on consumer-grade hardware.

Model Description

Base Model: google/paligemma-3b-pt-224
Fine-tuning Technique: QLoRA (Quantized Low-Rank Adaptation)
Task: Visual Question Answering (VQA)
Dataset: A subset of leonardPKU/clevr_cogen_a_train
Key Improvement: Enhanced ability to perform complex reasoning, counting, and attribute identification in visual scenes.

How to Use

To use this model, you must load the 4-bit quantized base model and then apply the PEFT adapters from this repository.

Installation

First, ensure you have the necessary libraries installed:

pip install -q transformers peft bitsandbytes accelerate Pillow requests

Sources

1 source

Hugging FacePrimaryVersion b48847caa189a367a80eef5b9100bb9ff6a7c527

huggingface.co/tahamajs/paligemma-clevr-cogen-qlora/tree/b48847caa189a367a80eef5b9100bb9ff6a7c527

Verified Oct 23

Download files

Model artifacts

19 artifacts

checkpoint-100/adapter_model.safetensors

safetensors · 345 MB · SHA-256 1777d60bbf64…2635 · Hugging Face

Download file

checkpoint-100/scaler.pt

pt · 1.35 KB · SHA-256 c67871b8673b…29c8 · Hugging Face

Download file

checkpoint-200/adapter_model.safetensors

safetensors · 345 MB · SHA-256 48719daa29ac…34f0 · Hugging Face

Download file

checkpoint-200/scaler.pt

pt · 1.35 KB · SHA-256 407deb9bd6b9…7bfa · Hugging Face

Download file

checkpoint-300/adapter_model.safetensors

safetensors · 345 MB · SHA-256 6ef21c14b362…1371 · Hugging Face

Download file

checkpoint-300/scaler.pt

pt · 1.35 KB · SHA-256 317b5a305b2a…138a · Hugging Face

Download file

checkpoint-400/adapter_model.safetensors

safetensors · 345 MB · SHA-256 40d57e915144…a287 · Hugging Face

Download file

checkpoint-400/scaler.pt

pt · 1.35 KB · SHA-256 9317116283c5…a005 · Hugging Face

Download file

checkpoint-500/adapter_model.safetensors

safetensors · 345 MB · SHA-256 daf594ae9d69…12e9 · Hugging Face

Download file

checkpoint-500/scaler.pt

pt · 1.35 KB · SHA-256 b02df20a5561…e77c · Hugging Face

Download file

checkpoint-600/adapter_model.safetensors

safetensors · 345 MB · SHA-256 5931ed25091c…3d3b · Hugging Face

Download file

checkpoint-600/scaler.pt

pt · 1.35 KB · SHA-256 1c99d07482e6…bf7d · Hugging Face

Download file

checkpoint-7/adapter_model.safetensors

safetensors · 86.3 MB · SHA-256 5ae5bb34bd9b…26a7 · Hugging Face

Download file

checkpoint-700/adapter_model.safetensors

safetensors · 345 MB · SHA-256 bacb3e29f4f3…5089 · Hugging Face

Download file

checkpoint-700/scaler.pt

pt · 1.35 KB · SHA-256 31b2fd2834f0…f41a · Hugging Face

Download file

checkpoint-788/adapter_model.safetensors

safetensors · 345 MB · SHA-256 2cc660aae397…5e6e · Hugging Face

Download file

checkpoint-788/scaler.pt

pt · 1.35 KB · SHA-256 8b31d4ee319c…84d0 · Hugging Face

Download file

pytorch_model-00001-of-00002.bin

bin · 4.64 GB · SHA-256 0d2c2dbb6403…a88c · Hugging Face

Download shard

pytorch_model-00002-of-00002.bin

bin · 822 MB · SHA-256 96517dca9737…1b33 · Hugging Face

Download shard

Source excerpts

2 excerpts

ReadmeAug 6huggingface.co/tahamajs/paligemma-clevr-cogen-qlora

--- language: en license: apache-2.0 library_name: peft tags: - paligemma - visual-question-answering - vqa - clevr - qlora - multimodal - peft base_model: google/paligemma-3b-pt-224 datasets: - leonardPKU/clevr_cogen_a_train pipeline_tag: visual-question-answering --- # QLoRA Fine-tuned PaliGemma-3B for Visual Reasoning on CLEVR-CoGen This repository contains the QLoRA adapters for the `google/paligemma-3b-pt-224` model, fine-tuned for a Visual Question Answering (VQA) task on the `leonardPKU/clevr_cogen_a_train` dataset. This fine-tuned model demonstrates significantly improved performance on questions requiring spatial and logical reasoning about complex scenes with multiple objects compared to the base PaliGemma model. The use of QLoRA (4-bit quantization) makes it possible to run and train this powerful model on consumer-grade hardware. ## Model Description - **Base Model:** `google/paligemma-3b-pt-224` - **Fine-tuning Technique:** QLoRA (Quantized Low-Rank Adaptation) - **Task:** Visual Question Answering (VQA) - **Dataset:** A subset of `leonardPKU/clevr_cogen_a_train` - **Key Improvement:** Enhanced ability to perform complex reasoning, counting, and attribute identification in visual scenes. ## How to Use To use this model, you must load the 4-bit quantized base model and then apply the PEFT adapters from this repository. ### Installation First, ensure you have the necessary libraries installed: ```bash pip install -q transformers peft bitsandbytes accelerate Pillow requests

NoteAug 6huggingface.co/tahamajs/paligemma-clevr-cogen-qlora

Source context: 0 downloads · 0 likes · Pipeline visual-question-answering · Library peft · Repo tahamajs/paligemma-clevr-cogen-qlora