AI Architecture & Systems
1
Microarchitecture
2
Kernel
3
Compiler
4
Architecture
5
System
6
Algorithms
Algorithms
/
§6.19
Parallel, Blockwise, and Diffusion Language Generation
6.19.1
Autoregressive, Semi-Autoregressive, and Non-Autoregressive Generation
#
6.19.2
Blockwise Parallel Decoding and Iterative Refinement
#
6.19.3
Set Block Decoding (SBD)
#
6.19.4
Masked Diffusion Language Modeling
#
6.19.5
LLaDA, LLaDA 2.0, Dream, and Block Diffusion
#
6.19.6
Token Update Order, Remasking, and Sampling Steps
#
6.19.7
Trade-offs Between KV Cache Compatibility and Generation Quality
#
← Previous
6.18 Speculative Decoding
→ Next
6.20 Quantization: Algorithms and Numerical Methods