Trending Research
Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models
osilly/vision-r1
•
•
However, direct training with RL struggles to activate complex reasoning capabilities such as questioning and reflection in MLLMs, due to the absence of substantial high-quality multimodal reasoning data.
LBM: Latent Bridge Matching for Fast Image-to-Image Translation
gojasper/lbm
•
•
In this paper, we introduce Latent Bridge Matching (LBM), a new, versatile and scalable method that relies on Bridge Matching in a latent space to achieve fast image-to-image translation.
Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond
The Light-R1 series of work validates training long-COT models from scratch, showcases the art in SFT data and releases SOTA models from RL.
Zero-shot Voice Conversion with Diffusion Transformers
Plachtaa/seed-vc
•
•
Zero-shot voice conversion aims to transform a source speech utterance to match the timbre of a reference speech from an unseen speaker.
GoT: Unleashing Reasoning Capability of Multimodal Large Language Model for Visual Generation and Editing
rongyaofang/got
•
•
We present Generation Chain-of-Thought (GoT), a novel paradigm that enables generation and editing through an explicit language reasoning process before outputting images.
Slim attention: cut your context memory in half without loss of accuracy -- K-cache is all you need for MHA
openmachine-ai/transformer-tricks
•
•
For encoder-decoder transformers, the context memory size can be reduced even further: For the Whisper models for example, slim attention reduces the context memory by 8x, which can speed up token generation by 5x for batch size 64 for example.
Comet: Fine-grained Computation-communication Overlapping for Mixture-of-Experts
bytedance/flux
•
•
The inter-device communication of a MoE layer can occupy 47% time of the entire model execution with popular models and frameworks.
LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL
tidedra/lmm-r1
•
•
Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment.
2 OLMo 2 Furious
Our modified model architecture and training recipe achieve both better training stability and improved per-token efficiency.
HealthGPT: A Medical Large Vision-Language Model for Unifying Comprehension and Generation via Heterogeneous Knowledge Adaptation
To effectively learn the HealthGPT, we devise a comprehensive medical domain-specific comprehension and generation dataset called VL-Health.
