research-ideas
- 2025 12 03 Matryoshka Transformers for Diffusion 0
- 2025 12 03 LORAs for diffusion steps 92
- 2025 01 27 Multi Label future token prediction head 54
- 2025 01 26 Super Fast Decoder Inference 0
- 2025 01 25 Take all branches in parallel 104
- 2025 01 25 Latent Generative visual reasoning 30
- 2025 01 25 Soft Verifiers 22
- 2025 01 25 GAN + Active Learning on top of Reasoning 106
- 2025 01 25 User Embedding Conditioned Generative Models 142
- 2025 01 25 Codebook KV Cache 476
- 2025 01 25 CLIP in GPT 597
- 2025 01 15 Commander - Super Fast Local Function Calling 413
- 2024 12 13 Pretrain on synthetic conversation data 235
- 2024 12 13 Predict token from positional embedding 0
- 2024 12 10 Neural Architecture Search for SSM Hybrids 185
- 2024 12 06 Teach VLM to Zoom and Pan 121
- 2024 11 27 Mixture of Modules 300
- 2024 11 08 Sapiens for Robotics 0
- 2024 11 08 Bad apples for label noise early stopping 0
- 2024 11 08 Small Proxy model to predict loss for given sample 40
- 2024 10 26 White space separated conv text encoder 0
- 2024 10 26 Early Fusion Multimodal Encoder Models 338
- 2024 10 25 Learning Skip Layers 0
- 2024 10 22 Learn to Initialize from OS Models 62
- 2024 10 19 Two Stream SSMs 0
- 2024 10 16 SSMs 4 Rec 0
- 2024 10 11 Universal embedding space for popular foundational models (or adapters) 532
- 2024 10 10 Tiny LLMs with rag in the middle 328
- 2024 10 09 Tiny Foundational model by distilling from a lot of SOTA models 0
- 2024 10 09 Remove all the things 609
- 2024 10 09 Multi Modal Learning to Rank as a replacement for CLIP 209
- 2024 10 09 Latent Transformers with small vocabularies 406
- 2024 10 09 Recurrent Computation with Transformers by repeating layers 269
- 2024 10 09 Task Routing for Multimodal LLMs 72
- 2024 10 09 VLMs for better Vision Backbones 578