Nagori's picture

1 7 7

Nagori

MohammedNaeem

·

Naeem_1144

AI & ML interests

None yet

Recent Activity

reacted to chansung's post with ❤️ 2 days ago

Simple summary on DeepSeek AI's Janus-Pro: A fresh take on multimodal AI! It builds on its predecessor, Janus, by tweaking the training methodology rather than the model architecture. The result? Improved performance in understanding and generating multimodal data. Janus-Pro uses a three-stage training strategy, similar to Janus, but with key modifications: ✦ Stage 1 & 2: Focus on separate training for specific objectives, rather than mixing data. ✦ Stage 3: Fine-tuning with a careful balance of multimodal data. Benchmarks show Janus-Pro holds its own against specialized models like TokenFlow XL and MetaMorph, and other multimodal models like SD3 Medium and DALL-E 3. The main limitation? Low image resolution (384x384). However, this seems like a strategic choice to focus on establishing a solid "recipe" for multimodal models. Future work will likely leverage this recipe and increased computing power to achieve higher resolutions.

liked a model 4 days ago

deepseek-ai/DeepSeek-R1

upvoted a collection 4 days ago

View all activity

Organizations

None yet

MohammedNaeem's activity

upvoted a collection 4 days ago

Qwen2.5-VL

Vision-language model series based on Qwen2.5 • 3 items • Updated 4 days ago • 290

upvoted a collection about 1 month ago

DeepSeek-V3

3 items • Updated 25 days ago • 160

upvoted a paper about 1 month ago

Ensembling Large Language Models with Process Reward-Guided Tree Search for Better Complex Reasoning

Paper • 2412.15797 • Published Dec 20, 2024 • 17

upvoted 3 collections about 1 month ago

QVQ

QVQ: Qwen models for visual reasoning • 7 items • Updated 30 days ago • 43

TimesFM Release

TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting. • 4 items • Updated Dec 24, 2024 • 13

Eagle 2

Eagle 2 is a family of frontier vision-language models with vision-centric design. The model supports 4K HD input, long-context video, and grounding. • 9 items • Updated 8 days ago • 27

upvoted a collection 3 months ago

Llama-3.1-Nemotron-70B

SOTA models on Arena Hard and RewardBench as of 1 Oct 2024. • 6 items • Updated 14 days ago • 152