English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Large Language Models (LLMs) have shown remarkable results on various complexreasoning benchmarks. The reasoning capabilities of LLMs enable them to executefunction calls, using user-provided functions to overcome their inherentlimitations, such as knowledge cutoffs, poor arithmetic skills, or lack ofaccess to private data. This development has expanded LLMs' scope to includemulti-function calling, where LLMs are equipped with a variety of functions andselect the proper functions based on the context. Multi-function callingabilities of LLMs have catalyzed LLM-based software development, allowing themto tackle more complex problems. However, current methods for multi-functioncalling often require sequential reasoning and acting for each function whichcan result in high latency, cost, and sometimes inaccurate behavior. To addressthis, we introduce LLMCompiler, which executes functions in parallel toefficiently orchestrate multi-function calling. Drawing from the principles ofclassical compilers, LLMCompiler streamlines parallel function calling withthree components: (i) an LLM Planner, formulating execution strategies anddependencies; (ii) a Task Fetching Unit, dispatching function calling tasks;and (iii) an Executor, executing these tasks in parallel. LLMCompilerautomatically computes an optimized orchestration for the function calls andcan be used with open-source models such as LLaMA-2. We have benchmarkedLLMCompiler on a range of tasks including cases with non-trivialinter-dependency between function calls, as well as cases that require dynamicreplanning based on intermediate results. We observe consistent latency speedupof up to 3.7x, cost savings of up to 6.7x, and accuracy improvement of up to~9% as compared to ReAct. Additionally, LLMCompiler achieves up to 1.35xlatency gain over OpenAI's recent parallel function calling, while achievingsimilar accuracy.

An LLM Compiler for Parallel Function Calling