Hugging Face's logo
Fatih C. Akyon's picture
75 3 117

Fatih C. Akyon

fcakyon
GershonJonish's profile picture asmaa1's profile picture Ak4215's profile picture
Ā·
https://www.linkedin.com/in/fcakyon/

AI & ML interests

multi-modal learning, video understanding

Recent Activity

reacted to merve's post with šŸ¤— 4 days ago
Apollo is a new family of open-source video language models by Meta, where 3B model outperforms most 7B models and 7B outperforms most 30B models šŸ§¶ āœØ the models come in 1.5B https://huggingface.co/Apollo-LMMs/Apollo-1_5B-t32, 3B https://huggingface.co/Apollo-LMMs/Apollo-3B-t32 and 7B https://huggingface.co/Apollo-LMMs/Apollo-7B-t32 with A2.0 license, based on Qwen1.5 & Qwen2 āœØ the authors also release a benchmark dataset https://huggingface.co/spaces/Apollo-LMMs/ApolloBench The paper has a lot of experiments (they trained 84 models!) about what makes the video LMs work āÆļø Try the demo for best setup here https://huggingface.co/spaces/Apollo-LMMs/Apollo-3B they evaluate sampling strategies, scaling laws for models and datasets, video representation and more! > The authors find out that whatever design decision was applied to small models also scale properly when the model and dataset are scaled šŸ“ˆ scaling dataset has diminishing returns for smaller models > They evaluate frame sampling strategies, and find that FPS sampling is better than uniform sampling, and they find 8-32 tokens per frame optimal > They also compare image encoders, they try a variation of models from shape optimized SigLIP to DINOv2 they find https://huggingface.co/google/siglip-so400m-patch14-384 to be most powerful šŸ”„ > they also compare freezing different parts of models, training all stages with some frozen parts give the best yield They eventually release three models, where Apollo-3B outperforms most 7B models and Apollo 7B outperforms 30B models šŸ”„
reacted to merve's post with šŸ‘€ 4 days ago
Apollo is a new family of open-source video language models by Meta, where 3B model outperforms most 7B models and 7B outperforms most 30B models šŸ§¶ āœØ the models come in 1.5B https://huggingface.co/Apollo-LMMs/Apollo-1_5B-t32, 3B https://huggingface.co/Apollo-LMMs/Apollo-3B-t32 and 7B https://huggingface.co/Apollo-LMMs/Apollo-7B-t32 with A2.0 license, based on Qwen1.5 & Qwen2 āœØ the authors also release a benchmark dataset https://huggingface.co/spaces/Apollo-LMMs/ApolloBench The paper has a lot of experiments (they trained 84 models!) about what makes the video LMs work āÆļø Try the demo for best setup here https://huggingface.co/spaces/Apollo-LMMs/Apollo-3B they evaluate sampling strategies, scaling laws for models and datasets, video representation and more! > The authors find out that whatever design decision was applied to small models also scale properly when the model and dataset are scaled šŸ“ˆ scaling dataset has diminishing returns for smaller models > They evaluate frame sampling strategies, and find that FPS sampling is better than uniform sampling, and they find 8-32 tokens per frame optimal > They also compare image encoders, they try a variation of models from shape optimized SigLIP to DINOv2 they find https://huggingface.co/google/siglip-so400m-patch14-384 to be most powerful šŸ”„ > they also compare freezing different parts of models, training all stages with some frozen parts give the best yield They eventually release three models, where Apollo-3B outperforms most 7B models and Apollo 7B outperforms 30B models šŸ”„
reacted to merve's post with šŸš€ 4 days ago
Apollo is a new family of open-source video language models by Meta, where 3B model outperforms most 7B models and 7B outperforms most 30B models šŸ§¶ āœØ the models come in 1.5B https://huggingface.co/Apollo-LMMs/Apollo-1_5B-t32, 3B https://huggingface.co/Apollo-LMMs/Apollo-3B-t32 and 7B https://huggingface.co/Apollo-LMMs/Apollo-7B-t32 with A2.0 license, based on Qwen1.5 & Qwen2 āœØ the authors also release a benchmark dataset https://huggingface.co/spaces/Apollo-LMMs/ApolloBench The paper has a lot of experiments (they trained 84 models!) about what makes the video LMs work āÆļø Try the demo for best setup here https://huggingface.co/spaces/Apollo-LMMs/Apollo-3B they evaluate sampling strategies, scaling laws for models and datasets, video representation and more! > The authors find out that whatever design decision was applied to small models also scale properly when the model and dataset are scaled šŸ“ˆ scaling dataset has diminishing returns for smaller models > They evaluate frame sampling strategies, and find that FPS sampling is better than uniform sampling, and they find 8-32 tokens per frame optimal > They also compare image encoders, they try a variation of models from shape optimized SigLIP to DINOv2 they find https://huggingface.co/google/siglip-so400m-patch14-384 to be most powerful šŸ”„ > they also compare freezing different parts of models, training all stages with some frozen parts give the best yield They eventually release three models, where Apollo-3B outperforms most 7B models and Apollo 7B outperforms 30B models šŸ”„
View all activity

Organizations

Deprem Yapay Zeka's profile picture Yapay ZekĆ¢ Araştırma Ä°nisiyatifi's profile picture Radiology-ai's profile picture OBSS's profile picture fixit's profile picture Gradio-Blocks-Party's profile picture ultralytics+'s profile picture Video Transformers's profile picture Viddexa AI's profile picture Ultralytics's profile picture

Papers 8