Efficiently Scaling Transformer Inference
Chop a 540B model across a TPU pod: 29ms/token, 76% MFU, 32x longer context
Pope et al. · MLSys 2023 · Serving. Read the paper ↗
A free, interactive, animated visual explainer of Efficiently Scaling Transformer Inference — every exhibit computed from the real formulas, with verbatim quotes from the source.
Questions
- What is Efficiently Scaling Transformer Inference?
- Chop a 540B model across a TPU pod: 29ms/token, 76% MFU, 32x longer context
- Who published Efficiently Scaling Transformer Inference, and where?
- Pope et al. — MLSys 2023 (arXiv:2211.05102).
- Where can I find a visual explainer of Efficiently Scaling Transformer Inference?
- Right here — a free, interactive, animated walkthrough of the whole paper, with exhibits computed from the real formulas and verbatim quotes from the source.
Related explainers
- PagedAttention (vLLM)
- Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving
- Fast Inference from Transformers via Speculative Decoding
- EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
- AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration
- DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving
- CacheBlend: Fast Large Language Model Serving for RAG with Cached Knowledge Fusion
- Efficient Streaming Language Models with Attention Sinks