“The only reason neural networks don't have control flow is because they are not very good. They are incredibly inefficient and the only way to properly solve that is to introduce control flow, for example: https://arxiv.org/abs/2311.10770”
Exponentially Faster Language Modelling
Language models only really need to use an exponential fraction of their neurons for individual inferences. As proof, we present UltraFastBERT, a BERT variant that uses 0.3% of its neurons during inference while performing on par with similar BERT models. UltraFastBERT selectively engages just 12 out of 4095 neurons for each layer inference. This is achieved by replacing feedforward networks with fast feedforward networks (FFFs). While no truly efficient implementation currently exists to unlock the full acceleration potential of conditional neural execution, we provide high-level CPU code achieving 78x speedup over the optimized baseline feedforward implementation, and a PyTorch implementation delivering 40x speedup over the equivalent batched feedforward inference. We publish our training code, benchmarking setup, and model weights.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及- Exponentially Faster Language Modelling ↗born-jre · 2023-11-21 · 2 分
- Exponentially Faster Language Modelling ↗fgfm · 2023-11-21 · 1 分
- Ultra Fast Bert ↗gyre007 · 2023-11-22 · 2 分