Capability
Batch Processing With Map Reduce Pattern
3 artifacts provide this capability.
Want a personalized recommendation?
Find the best match →Top Matches
1
PocketFlowAgent44/100
via “batch processing with map-reduce pattern”
Pocket Flow: 100-line LLM framework. Let Agents build Agents!
Unique: Implements map-reduce as a first-class Flow type within the Graph + Shared Store model, enabling batch processing to be composed with agent and RAG nodes without external distributed computing frameworks
vs others: Simpler than Ray/Dask (no cluster management) but less scalable (single-machine only); more integrated than Celery (no separate worker processes required)