11 9 38

Flo Schneider

floschne

https://www.inf.uni-hamburg.de/en/inst/ab/lt/people/florian-schneider.html

floschne

AI & ML interests

Multi Modal Information Retrieval and Representation Learning

Recent Activity

New activity 24 days ago

neulab/PangeaBench-xmmmu:Issues when downloading the dataset

upvoted a paper about 1 month ago

Aria: An Open Multimodal Native Mixture-of-Experts Model

View all activity

Organizations

floschne's activity

New activity in neulab/PangeaBench-xmmmu 24 days ago

Issues when downloading the dataset

#1 opened 24 days ago by

floschne

upvoted a paper about 1 month ago

Aria: An Open Multimodal Native Mixture-of-Experts Model

Paper • 2410.05993 • Published Oct 8 • 107

upvoted a collection 2 months ago

LLaVA-Onevision

Collection

LLaVa_Onevision models for single-image, multi-image, and video scenarios • 9 items • Updated Sep 18 • 12

liked a dataset 2 months ago

facebook/belebele

Viewer • Updated Aug 12 • 110k • 7.73k • 98

liked a model 2 months ago

Qwen/Qwen2-VL-7B-Instruct

Image-Text-to-Text • Updated Sep 21 • 1.67M • • 839

upvoted an article 2 months ago

Article

Introducing IDEFICS: An Open Reproduction of State-of-the-art Visual Language Model

Aug 22, 2023

• 27

upvoted a paper 2 months ago

M5 -- A Diverse Benchmark to Assess the Performance of Large Multimodal Models Across Multilingual and Multicultural Vision-Language Tasks

Paper • 2407.03791 • Published Jul 4 • 1

liked a model 2 months ago

royokong/e5-v

Image-Text-to-Text • Updated 24 days ago • 13.7k • 18

liked a dataset 3 months ago

Rocktim/EXAMS-V

Viewer • Updated May 7 • 21.3k • 337 • 7

upvoted a paper 3 months ago

LLaVA-OneVision: Easy Visual Task Transfer

Paper • 2408.03326 • Published Aug 6 • 59

liked a dataset 5 months ago

afaji/cvqa

Viewer • Updated 20 days ago • 10.4k • 631 • 21

Reacted to mrm8488's post with ❤️ 5 months ago

Post

4374

🚨Exciting news for the Multilingual Synthetic Data Community!🚨

I’ve taken inspiration from the MAGPIE paper on Llama-3-8B-instruct and extended its capabilities. Here’s what’s new!

🗞 The MAGPIE paper showcased that if you use the instruction-tuned version (Llama-3-8B-instruct) to generate synthetic instructions and then fine-tune the base version (Llama-3-8B) on this dataset, you can improve even the it-tuned version

🤔 While reading a script by Sebastian Raschka, PhD, I wondered: Could these advancements be replicated in other languages? Specifically, could they benefit non-English datasets?

🎉 And the answer is YES! At least for Spanish. I've successfully adapted the techniques for Spanish, proving the model's flexibility and multilingual capabilities.

👩‍💻 To make this accessible, I created a basic script (heavily inspired by the Sebastian Raschka one) that allows you to generate similar datasets using ollama models (initially phi and llama3) automatically and upload it to the Hugging Face Hub!
[Script](https://gist.github.com/mrm8488/4650a5e3cc45523798a527a3446eb312)

🔍 Explore the datasets 📚 generated using our new script!

- [Llama-3-8B](https://huggingface.co/datasets/mrm8488/dataset_llama3_5000_samples_es_4231_filtered)
- [Phi-3-medium](https://huggingface.co/datasets/mrm8488/dataset_phi3-medium_5000_samples_es_3906_filtered)
- [Phi-3-mini](https://huggingface.co/datasets/mrm8488/dataset_phi3_5000_samples_es_3282_filtered)

Note: These datasets have basic filtering. Apply additional quality filters before using them to fine-tune large language models.

Inspiration and base script:
https://github.com/rasbt/LLMs-from-scratch/blob/main/ch07/05_dataset-generation/llama3-ollama.ipynb
https://www.linkedin.com/feed/update/urn:li:activity:7210982019751661568/