17 6 269

Mel Massadian

melmass

https://melmassadian.com

AI & ML interests

Building tools on top of Generative AI & LLM models

Recent Activity

liked a model 3 days ago

hexgrad/Kokoro-82M

liked a model 3 days ago

stabilityai/stable-point-aware-3d

reacted to merve's post with 🔥 3 days ago

ByteDance just dropped SA2VA: a new family of vision LMs combining Qwen2VL/InternVL and SAM2 with MIT license 💗 https://huggingface.co/collections/ByteDance/sa2va-model-zoo-677e3084d71b5f108d00e093 > The models are capable of tasks involving vision-language understanding and visual referrals (referring segmentation) both for images and videos ⏯️ > The models come in 1B, 4B and 8B and are based on InternVL2.5 for base architecture and Qwen2, Qwen2.5 and InternLM2 for language model part (depending on the checkpoint) > The model is very interesting, it has different encoders for different modalities each (visual prompt, text prompt, image and video) then it concatenates these to feed into LLM 💬 the output segmentation tokens are passed to SAM2, to sort of match text (captions or semantic classes) to masks ⤵️ > Their annotation pipeline is also interesting, they seems to use two open large vision LMs to refine the annotations, and have different levels of descriptions to provide consistency.

View all activity

Organizations

melmass's activity

liked 2 models 3 days ago

hexgrad/Kokoro-82M

Text-to-Speech • Updated 7 days ago • 11.4k • 907

stabilityai/stable-point-aware-3d

Image-to-3D • Updated 4 days ago • 2.13k • 123

reacted to merve's post with 🔥 3 days ago

Post

1699

ByteDance just dropped SA2VA: a new family of vision LMs combining Qwen2VL/InternVL and SAM2 with MIT license 💗 ByteDance/sa2va-model-zoo-677e3084d71b5f108d00e093

> The models are capable of tasks involving vision-language understanding and visual referrals (referring segmentation) both for images and videos ⏯️

> The models come in 1B, 4B and 8B and are based on InternVL2.5 for base architecture and Qwen2, Qwen2.5 and InternLM2 for language model part (depending on the checkpoint)

> The model is very interesting, it has different encoders for different modalities each (visual prompt, text prompt, image and video) then it concatenates these to feed into LLM 💬

the output segmentation tokens are passed to SAM2, to sort of match text (captions or semantic classes) to masks ⤵️

> Their annotation pipeline is also interesting, they seems to use two open large vision LMs to refine the annotations, and have different levels of descriptions to provide consistency.