Many researchers only organise data after paper acceptance. Poor archival practice creates reproducibility risks. What is your step‑by‑step routine to archive raw and intermediate data while the project is ongoing?

Clean metadata is critical for reproducibility, yet manually organising metadata is extremely time‑consuming. Are there lesser‑known lightweight open utilities you rely on for dataset metadata standardisation?

Digital or physical lab notebooks are your core research asset. Record raw data, parameter settings and unexpected anomalies in real‑time instead of recalling information days later.

PEFT methods are everywhere these days. But I’m starting to question: for extremely niche small‑size domain data, are LoRA/QLoRA consistently better than full fine‑tune with careful regularization? Would love real‑world hands‑on observations, not just paper results.