batch-mode-batch-workflow
This crate provides a specialized approach to batch-based GPT expansions and reconciling partial or incomplete batch states. It defines traits and workflows for:
- Reconciling uncompleted batches from prior runs,
- Computing new requests for language model expansions,
- Processing these requests in batches (locally or via remote APIs),
- Handling responses and results systematically.
Key Traits
- FinishProcessingUncompletedBatches – Finalize partial data left from incomplete batch processing.
- ComputeLanguageModelRequests – Identify new items to process and generate requests to a language model API.
- ProcessBatchRequests – Handle chunked batch requests.
- LanguageModelBatchWorkflow – Integrates the above traits into a high-level, end-to-end batch-processing workflow.
Getting Started
Add this crate to your Cargo.toml:
[]
= "*"