Upload artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh
Browse files
artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env bash
|
| 2 |
+
set -euo pipefail
|
| 3 |
+
|
| 4 |
+
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
| 5 |
+
REPO_ROOT="$(cd "${SCRIPT_DIR}/../../../.." && pwd)"
|
| 6 |
+
cd "${REPO_ROOT}"
|
| 7 |
+
|
| 8 |
+
# llamafactory-cli launches distributed training via `torchrun`, so the venv
|
| 9 |
+
# bin dir must be on PATH even when the CLI itself is invoked by absolute path.
|
| 10 |
+
export PATH="${REPO_ROOT}/.venv/bin:${PATH}"
|
| 11 |
+
|
| 12 |
+
if [[ -n "${SLURM_CPUS_PER_TASK:-}" ]]; then
|
| 13 |
+
CPU_WORKERS_DEFAULT="${SLURM_CPUS_PER_TASK}"
|
| 14 |
+
elif command -v nproc >/dev/null 2>&1; then
|
| 15 |
+
CPU_WORKERS_DEFAULT="$(nproc)"
|
| 16 |
+
else
|
| 17 |
+
CPU_WORKERS_DEFAULT=1
|
| 18 |
+
fi
|
| 19 |
+
|
| 20 |
+
DEFAULT_PREPROCESSING_WORKERS="${PREPROCESSING_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
|
| 21 |
+
DEFAULT_DATALOADER_WORKERS="${DATALOADER_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
|
| 22 |
+
|
| 23 |
+
if [[ -z "${CUDA_VISIBLE_DEVICES:-}" ]]; then
|
| 24 |
+
echo "CUDA_VISIBLE_DEVICES must be set before running this script" >&2
|
| 25 |
+
exit 1
|
| 26 |
+
fi
|
| 27 |
+
|
| 28 |
+
LLAMA_BIN_DEFAULT="${REPO_ROOT}/.venv/bin/llamafactory-cli"
|
| 29 |
+
DATASET_DIR_ROOT="${DATASET_DIR_ROOT:-data}"
|
| 30 |
+
if [[ ! -x "${LLAMA_BIN_DEFAULT}" ]]; then
|
| 31 |
+
echo "Missing ${LLAMA_BIN_DEFAULT}. Run scripts/setup/install_local_llamafactory.sh first." >&2
|
| 32 |
+
exit 1
|
| 33 |
+
fi
|
| 34 |
+
|
| 35 |
+
if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/train" ]]; then
|
| 36 |
+
echo "Missing ${DATASET_DIR_ROOT}/composition/train. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
|
| 37 |
+
exit 1
|
| 38 |
+
fi
|
| 39 |
+
|
| 40 |
+
if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/test" ]]; then
|
| 41 |
+
echo "Missing ${DATASET_DIR_ROOT}/composition/test. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
|
| 42 |
+
exit 1
|
| 43 |
+
fi
|
| 44 |
+
|
| 45 |
+
export WANDB_PROJECT="${WANDB_PROJECT:-Interplay-LM-Reasoning}"
|
| 46 |
+
export WANDB_ENTITY="${WANDB_ENTITY:-kmchiti}"
|
| 47 |
+
|
| 48 |
+
DEFAULT_LLAMA_ARGS=(
|
| 49 |
+
"preprocessing_num_workers=${DEFAULT_PREPROCESSING_WORKERS}"
|
| 50 |
+
"dataloader_num_workers=${DEFAULT_DATALOADER_WORKERS}"
|
| 51 |
+
)
|
| 52 |
+
|
| 53 |
+
if [[ -n "${LLAMA_EXTRA_ARGS:-}" ]]; then
|
| 54 |
+
export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]} ${LLAMA_EXTRA_ARGS}"
|
| 55 |
+
else
|
| 56 |
+
export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]}"
|
| 57 |
+
fi
|
| 58 |
+
|
| 59 |
+
EVAL_DATA_ROOT="${EVAL_DATA_ROOT:-${DATASET_DIR_ROOT}/composition/test}" \
|
| 60 |
+
LLAMA_BIN="${LLAMA_BIN:-${LLAMA_BIN_DEFAULT}}" \
|
| 61 |
+
LLAMA_CONFIG="scripts/composition/op-difficulty-10B/pt-diff2_10-tok10B-lr1e-4-bs512k-schedcos-minlr3e-5/id2-10_0.25easy_0.25medium_0.5hard.yaml" \
|
| 62 |
+
./scripts/meta_run.sh --skip-rl "$@"
|