#!/usr/bin/env bash
# Chain: GRPO from Options 1+3 final ckpt → TF + AR ablations → push to HF under grpo_opt13/
set -uo pipefail

REPO="LauraGG/blt-reasoner-pilot1"
OUT="/home/ubuntu/work/blt_grpo_opt13"
CFG="/home/ubuntu/experiments/blt_reasoner/configs/grpo_from_opt13.json"
LOG="/home/ubuntu/work/queue_grpo_opt13.log"

log() { echo "[$(date +%T)] $*" | tee -a "$LOG"; }

mkdir -p "$OUT"
cd /home/ubuntu
export TOKENIZERS_PARALLELISM=false TRANSFORMERS_NO_ADVISORY_WARNINGS=1 HF_HUB_DISABLE_PROGRESS_BARS=1
export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True

log "==========================================="
log "GRPO Phase C from BLT 7B Options 1+3 ckpt (51% AR baseline)"
log "==========================================="
python3 -u -m experiments.blt_reasoner.grpo_train --config "$CFG" \
    > "$OUT/train.log" 2>&1
log "grpo train exit=$?"

log "TF ablation on grpo final"
python3 -u -m experiments.blt_reasoner.scripts.ablate_teacher_forced \
    --ckpt "$OUT/final" --config "$CFG" --n 200 --K 16 \
    --out "$OUT/final/ablation_teacher_forced.json" \
    > "$OUT/tf_eval.log" 2>&1
log "TF ablate exit=$?"

log "AR ablation on grpo final"
python3 -u -m experiments.blt_reasoner.eval \
    --ckpt "$OUT/final" --config "$CFG" --n 200 --K 16 \
    --max_new_tokens 192 --temperature 0.0 \
    --out "$OUT/final/ablation_n200_K16.json" \
    > "$OUT/ar_eval.log" 2>&1
log "AR ablate exit=$?"

log "pushing grpo_opt13/ to HF"
python3 - <<PYEND
import os
from huggingface_hub import HfApi
token = os.environ.get("BLT_HF_TOKEN", "").strip()
assert token.startswith("hf_"), "BLT_HF_TOKEN missing"
api = HfApi(token=token)
api.upload_folder(folder_path="$OUT", path_in_repo="grpo_opt13",
                   repo_id="$REPO", repo_type="model",
                   commit_message="GRPO Phase C from Options 1+3 ckpt")
print("[push] done")
PYEND
log "queue_grpo_opt13.sh DONE"