HPD-Parsing: Hierarchical Parallel Document Parsing
For document parsing tasks, this work introduces a new decoding paradigm that significantly improves throughput without sacrificing accuracy, addressing the sequential bottleneck in existing VLM-based parsers.
HPD-Parsing replaces full-page autoregressive generation with hierarchical parallel decoding, achieving 4,752 tokens per second—2.62× throughput of the fastest existing model and 3.06× of the vanilla baseline—while maintaining competitive parsing accuracy.
Efficient teamwork typically combines global coordination with parallel execution, a principle not yet fully reflected in unified Vision-Language Model (VLM)-based document parsers. Existing unified parsers process an entire page jointly but generate its output through a single token-by-token autoregressive trajectory, creating a sequential bottleneck that grows with document length. Such full-page sequential generation overlooks a key property of document parsing: layout must be analyzed globally, whereas block content can be parsed in parallel. Based on this observation, we introduce HPD-Parsing, which replaces full-page autoregressive generation with a Hierarchical Parallel Decoding paradigm. A main layout branch organizes the overall document structure and dynamically assigns block-level content decoding to concurrent branches, while progressive multi-token prediction (P-MTP) further reduces the decoding steps within each branch. Experiments on public benchmarks show that HPD-Parsing achieves 4,752 tokens per second, delivering $2.62\times$ the throughput of the fastest existing document parsing model and $3.06\times$ that of the vanilla autoregressive baseline, while maintaining competitive parsing accuracy. These results establish hierarchical parallel decoding as an effective alternative to full-page autoregressive generation, opening a new direction for efficient unified document parsing.