File size: 1,015 Bytes
52c8187 e9ad461 52c8187 e9ad461 52c8187 e9ad461 52c8187 e9ad461 52c8187 e9ad461 52c8187 e9ad461 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 | ---
base_model: meta-llama/Llama-3.2-3B-Instruct
language:
- ko
- en
license: cc-by-nc-4.0
tags:
- speech-to-sft
- emnlp-2026-industry
---
# expertQA_llama3.2-3b_exp2_medical_seed1
SFT checkpoint from the EMNLP 2026 Industry Track submission *A Factorial Ablation of a Speech-to-SFT Pipeline: Differential Effects on Data Quality and Downstream Transfer*.
| Field | Value |
|---|---|
| Pipeline condition | **Exp 2** (full pipeline (Phase 0 + Phase 2)) |
| Domain | medical |
| Seed | 1 |
| Base model | [meta-llama/Llama-3.2-3B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct) |
| Training | LoRA (rank 16, α=32, QLoRA 4-bit, lr 2e-4, 3 epochs) |
| Upstream STT | In-house STT (paper main pipeline) |
| License | CC BY-NC 4.0 (research and non-commercial use only) |
**Intended use**: research and non-commercial use only, matching the consent scope of the source audio.
**Companion repository** (code, configs, prompts, sample QA): <https://github.com/flitto/speech-to-sft-ablation-paper>
|