Multimodal Co-Generation With Continuous Bitstream Diffusion
cs.CL
Submitted: 2026-10-09
Updated: 2026-10-09
Code: https://github.com/GBATZOLIS/BitstreamDiffusion
Project page: https://demo-2546.github.io/cobit-text-audio-samples
Sources
- SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model
- Structured Denoising Diffusion Models in Discrete State-Spaces
- CoBit: Language Modeling with Bitstream Diffusion
- Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design
- LangFlow: Continuous Diffusion Rivals Discrete in Language Modeling
- Flow-SLM: Joint Learning of Linguistic and Acoustic Information for Spoken Language Modeling
- Moshi: a speech-text foundation model for real-time dialogue
- ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification
- Continuous diffusion for categorical data
- SEED-X: Multimodal Models with Unified Multi-granularity Comprehension and Generation
- ELF: Embedded Language Flows
- Diffusion-LM Improves Controllable Text Generation
- World Model on Million-Length Video And Language With Blockwise RingAttention
- SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign
- MLS: A Large-Scale Multilingual Dataset for Speech Research
- Noise Scheduling as Information-Guided Allocation in Diffusion Training
- Categorical Flow Maps
- Score-Based Generative Modeling through Stochastic Differential Equations
- Llama-Mimi: Exploring the Limits of Flattened Speech Language Modeling
- Entropy Across the Bridge: Conditional-Marginal Discretization for Flow and Schr\"odinger Samplers
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering