deepopen
A Laya-based Intent Classification Framework

Historical BANKING77 training used the full training set, including a previously used development subset. That subset is not an independent validation set for that model. The existing test results have been observed; the new continuation results are retrospective test comparisons, not fresh independent confirmation. No further tuning follows these test results.

The headline is the best observed single-model result across three fixed SAM seeds, selected for display after test evaluation. All candidates started from the same historical full-train checkpoint.

Uniform evaluation: eager FP32, SDPA, batch 32, maximum 128 tokens. The historical checkpoint scores 93.9935% under this path. SAM mean gain: +0.0433 percentage points (shared-row 95% CI: -0.2273 to +0.3247); no significant improvement is established. The best-seed score is subject to test-selection bias.

Archived historical accuracy: 94.0909%. Current replay of the original BF16 script: 94.0584%. Exact historical prediction reproduction remains unresolved despite matching data and weight hashes. Robust improvement across execution settings is not established.

Execution audit / Training and evaluation configuration

A Laya-based Framework for Intent Classification

deepopen organizes intent classification around a common processing pipeline. Input construction, prediction heads, training objectives, and optional fusion are configured for each dataset.

The shared diagram describes the framework workflow. Dataset configurations differ in architecture and training strategy, not only numerical hyperparameters. Reported scores are tied to the configurations above, rather than one universal checkpoint.

Historical BANKING77 and CLINC150 GPU allocation, CUDA versions, and full training hyperparameters are not established by the release-loading record. They are marked unverified here rather than replaced with current-machine values. No throughput or latency claim is made.

Verified historical-package loading environment

This is the clean release-loading environment, not a claim about the original training environment or a reproduction of test accuracy. Additional recorded dependencies are in the downloadable JSON.

Historical prediction rows preserve artifact order, not official dataset IDs. Test utterances are not redistributed. Metrics are recounted from saved predictions; BANKING77 now uses the fixed-checkpoint retrospective test predictions of SAM seed 13.

Public model weights, inference code, dependency locks, licenses, training-source archives, and per-example evaluation evidence. Each downloaded model was checked against file hashes and its documented prediction reference. CLINC150 single reproduces the historical total score with two disclosed prediction differences; both historical and current-reference predictions are preserved.

BANKING77 is the best observed SAM seed after test evaluation; it is not a new independent validation-selected improvement. CLINC150 excludes out-of-scope inputs; single and ensemble results remain separate. Reproduction verifies the fixed historical artifacts and does not establish an official leaderboard rank.

For a Hugging Face cache snapshot, run from the downloaded repository directory using python -m infer_banking, python -m infer_clinc, or python -m infer_hybrid, followed by --checkpoint . --text "your text". Module execution preserves sibling imports when snapshot files are symbolic links.