Expand description
Standardized benchmark suite for Engram
Provides implementations of standard AI memory benchmarks:
- LOCOMO: Multi-session conversation memory
- LongMemEval: 5-dimension memory evaluation
- MemBench: CRUD throughput and search quality
Modules§
- locomo
- LOCOMO benchmark — Multi-session conversation memory
- longmemeval
- LongMemEval benchmark — 5-dimension memory evaluation
- membench
- MemBench — CRUD throughput and search quality benchmark
Structs§
- Benchmark
Result - Result of running a single benchmark
- Benchmark
Suite - Suite that manages and runs multiple benchmarks
Traits§
- Benchmark
- A benchmark that can be run against an Engram database
Functions§
- default_
suite - Build the default suite with all benchmarks