AudioDeliveryPairwiseMLP for Phase 1 EmpathyEval
This repository distributes the final trained checkpoint for the
provided_text_conditioned_audio_delivery_ranking system.
Code and Technical Report
The final-only code, setup instructions, tests, and technical report are at:
This checkpoint is a custom PyTorch state-dictionary package. It is loaded by
the AudioDeliveryPairwiseMLP implementation in that repository; it is not a
standard transformers.AutoModel checkpoint.
Files
model.pt: final AudioDeliveryPairwiseMLP checkpoint.config.yaml: model and training configuration.feature_group_importance.json: final feature-group analysis.
Integrity
model.pt SHA-256: 24303cb353b767c6f8c964876beec85cbf868633aa88e3f291cf8004a5e23083config.yaml SHA-256: d26c614e9eb27207bd6c9acdc00419e64e998ec32899410e159564af1d9d5038
The expected submitted JSONL has SHA-256
f84f1a12568f2556db273fc3b70e7d87b6319e83cf086c7772405e3abd185222.
Method
The scorer uses prosody, Emotion2Vec Plus Large, WavLM Base Plus, task
metadata, and text-length metadata from official JSON-provided text. It does
not use ASR, teacher models, fusion, fallback, speech-quality features, or
test-time training.
See the paired GitHub code repository and TECHNICAL_REPORT.md for setup,
feature construction, exact runtime versions, and reproduction instructions.
