scripts/02_predictor/train_predictor.sh

#!/usr/bin/env bash
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
source "${SCRIPT_DIR}/../00_setup/env.sh"

cd "${GENERATOR_DIR}"
python3 src/tasks/downstream/sequence_understanding.py \
  --model_name "${GENERATOR_BASE_MODEL}" \
  --dataset_name "${DEEPSTARR_DIR}" \
  --problem_type regression \
  --main_metrics pearson \
  --output_dir "${PREDICTOR_DIR}" \
  --batch_size "${PREDICTOR_BATCH_SIZE:-1024}" \
  --gradient_accumulation_steps "${PREDICTOR_GRAD_ACCUM:-1}" \
  --learning_rate "${PREDICTOR_LR:-1e-5}" \
  --num_train_epochs "${PREDICTOR_EPOCHS:-3}" \
  --max_length "${PREDICTOR_MAX_LENGTH:-256}" \
  --attn_implementation "${ATTN_IMPLEMENTATION:-sdpa}" \
  --enable_gradient_checkpointing