初始化项目，由ModelHub XC社区提供模型

Model: MathMindsAGI/Test_context_pretrain Source: Original Platform
2026-04-11 11:04:57 +08:00
commit 200675bd2d
23 changed files with 28409 additions and 0 deletions
--- a/artifacts/training/all_results.json
+++ b/artifacts/training/all_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "total_flos": 5.994863411375112e+18,
+    "train_loss": 0.031714059101823636,
+    "train_runtime": 3465.0177,
+    "train_samples_per_second": 1387.365,
+    "train_steps_per_second": 5.42
+}
--- a/artifacts/training/scripts/run.sh
+++ b/artifacts/training/scripts/run.sh
@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+#SBATCH --ntasks=1
+#SBATCH --nodes=1
+#SBATCH --partition=short-unkillable
+#SBATCH --gres=gpu:h100:4
+#SBATCH -c 24
+#SBATCH --mem=64G
+#SBATCH -t 2:59:0
+
+
+set -euo pipefail
+
+SCRATCH_ROOT="${SCRATCH_ROOT:-/network/scratch/k/kamran.chitsaz}"
+export HF_HOME="${HF_HOME:-${SCRATCH_ROOT}/.cache/huggingface}"
+export HF_DATASETS_CACHE="${HF_DATASETS_CACHE:-${HF_HOME}/datasets}"
+export TRANSFORMERS_CACHE="${TRANSFORMERS_CACHE:-${HF_HOME}/transformers}"
+export TMPDIR="${TMPDIR:-${SCRATCH_ROOT}/tmp}"
+
+mkdir -p "${HF_DATASETS_CACHE}" "${TRANSFORMERS_CACHE}" "${TMPDIR}"
+
+export PREPROCESSING_NUM_WORKERS=6
+export DATALOADER_NUM_WORKERS=6
+
+bash scripts/composition/op-difficulty-10B/script_pt/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh
--- a/artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh
+++ b/artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh
@@ -0,0 +1,62 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${SCRIPT_DIR}/../../../.." && pwd)"
+cd "${REPO_ROOT}"
+
+# llamafactory-cli launches distributed training via `torchrun`, so the venv
+# bin dir must be on PATH even when the CLI itself is invoked by absolute path.
+export PATH="${REPO_ROOT}/.venv/bin:${PATH}"
+
+if [[ -n "${SLURM_CPUS_PER_TASK:-}" ]]; then
+    CPU_WORKERS_DEFAULT="${SLURM_CPUS_PER_TASK}"
+elif command -v nproc >/dev/null 2>&1; then
+    CPU_WORKERS_DEFAULT="$(nproc)"
+else
+    CPU_WORKERS_DEFAULT=1
+fi
+
+DEFAULT_PREPROCESSING_WORKERS="${PREPROCESSING_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
+DEFAULT_DATALOADER_WORKERS="${DATALOADER_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
+
+if [[ -z "${CUDA_VISIBLE_DEVICES:-}" ]]; then
+    echo "CUDA_VISIBLE_DEVICES must be set before running this script" >&2
+    exit 1
+fi
+
+LLAMA_BIN_DEFAULT="${REPO_ROOT}/.venv/bin/llamafactory-cli"
+DATASET_DIR_ROOT="${DATASET_DIR_ROOT:-data}"
+if [[ ! -x "${LLAMA_BIN_DEFAULT}" ]]; then
+    echo "Missing ${LLAMA_BIN_DEFAULT}. Run scripts/setup/install_local_llamafactory.sh first." >&2
+    exit 1
+fi
+
+if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/train" ]]; then
+    echo "Missing ${DATASET_DIR_ROOT}/composition/train. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
+    exit 1
+fi
+
+if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/test" ]]; then
+    echo "Missing ${DATASET_DIR_ROOT}/composition/test. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
+    exit 1
+fi
+
+export WANDB_PROJECT="${WANDB_PROJECT:-Interplay-LM-Reasoning}"
+export WANDB_ENTITY="${WANDB_ENTITY:-kmchiti}"
+
+DEFAULT_LLAMA_ARGS=(
+    "preprocessing_num_workers=${DEFAULT_PREPROCESSING_WORKERS}"
+    "dataloader_num_workers=${DEFAULT_DATALOADER_WORKERS}"
+)
+
+if [[ -n "${LLAMA_EXTRA_ARGS:-}" ]]; then
+    export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]} ${LLAMA_EXTRA_ARGS}"
+else
+    export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]}"
+fi
+
+EVAL_DATA_ROOT="${EVAL_DATA_ROOT:-${DATASET_DIR_ROOT}/composition/test}" \
+LLAMA_BIN="${LLAMA_BIN:-${LLAMA_BIN_DEFAULT}}" \
+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" \
+    ./scripts/meta_run.sh --skip-rl "$@"
--- a/artifacts/training/train_results.json
+++ b/artifacts/training/train_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "total_flos": 5.994863411375112e+18,
+    "train_loss": 0.031714059101823636,
+    "train_runtime": 3465.0177,
+    "train_samples_per_second": 1387.365,
+    "train_steps_per_second": 5.42
+}
--- a/artifacts/training/trainer_log.jsonl
+++ b/artifacts/training/trainer_log.jsonl
--- a/artifacts/training/trainer_state.json
+++ b/artifacts/training/trainer_state.json
--- a/artifacts/training/training_args.bin
+++ b/artifacts/training/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:513e25aff4b26f6d4627ba5293f2d27ea5ae535068c635b6f731c100fe640f23
+size 6353
--- a/artifacts/training/training_loss.png
+++ b/artifacts/training/training_loss.png
--- a/artifacts/upload_manifest.json
+++ b/artifacts/upload_manifest.json
@@ -0,0 +1,114 @@
+{
+  "repo_id": "MathMindsAGI/Test_context_pretrain",
+  "requested_repo_id": "MathMindsAGI/Test_context_pretrain",
+  "model_export_dir": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt",
+  "training_run_dir": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt",
+  "results_dir": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty",
+  "uploaded_files": [
+    {
+      "path_in_repo": "README.md",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/README.md",
+      "size_bytes": 1318
+    },
+    {
+      "path_in_repo": "config.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/config.json",
+      "size_bytes": 691
+    },
+    {
+      "path_in_repo": "generation_config.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/generation_config.json",
+      "size_bytes": 133
+    },
+    {
+      "path_in_repo": "model.safetensors",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/model.safetensors",
+      "size_bytes": 412699256
+    },
+    {
+      "path_in_repo": "special_tokens_map.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/special_tokens_map.json",
+      "size_bytes": 699
+    },
+    {
+      "path_in_repo": "tokenizer.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/tokenizer.json",
+      "size_bytes": 132794
+    },
+    {
+      "path_in_repo": "tokenizer_config.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/tokenizer_config.json",
+      "size_bytes": 3143
+    },
+    {
+      "path_in_repo": "chat_template.jinja",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/chat_template.jinja",
+      "size_bytes": 443
+    },
+    {
+      "path_in_repo": "artifacts/training/all_results.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/all_results.json",
+      "size_bytes": 211
+    },
+    {
+      "path_in_repo": "artifacts/training/train_results.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/train_results.json",
+      "size_bytes": 211
+    },
+    {
+      "path_in_repo": "artifacts/training/trainer_log.jsonl",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/trainer_log.jsonl",
+      "size_bytes": 385767
+    },
+    {
+      "path_in_repo": "artifacts/training/trainer_state.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/trainer_state.json",
+      "size_bytes": 330574
+    },
+    {
+      "path_in_repo": "artifacts/training/training_args.bin",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/training_args.bin",
+      "size_bytes": 6353
+    },
+    {
+      "path_in_repo": "artifacts/training/training_loss.png",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/saves/composition-10B/op_level/id2-10_0.25easy_0.25medium_0.5hard/pt/training_loss.png",
+      "size_bytes": 28679
+    },
+    {
+      "path_in_repo": "artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/scripts/composition/op-difficulty-10B/script_pt/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh",
+      "size_bytes": 2320
+    },
+    {
+      "path_in_repo": "artifacts/training/scripts/run.sh",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/run.sh",
+      "size_bytes": 754
+    },
+    {
+      "path_in_repo": "results/eval_difficulty/checkpoint-18779_id_generations.jsonl",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty/checkpoint-18779_id_generations.jsonl",
+      "size_bytes": 1790663092
+    },
+    {
+      "path_in_repo": "results/eval_difficulty/checkpoint-18779_metrics.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty/checkpoint-18779_metrics.json",
+      "size_bytes": 66147
+    },
+    {
+      "path_in_repo": "results/eval_difficulty/checkpoint-18779_ood_generations.jsonl",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty/checkpoint-18779_ood_generations.jsonl",
+      "size_bytes": 2445405195
+    },
+    {
+      "path_in_repo": "results/eval_difficulty/summary.csv",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty/summary.csv",
+      "size_bytes": 322
+    },
+    {
+      "path_in_repo": "results/eval_difficulty/summary.json",
+      "source_path": "/network/scratch/k/kamran.chitsaz/Interplay-LM-Reasoning/results/eval_difficulty/summary.json",
+      "size_bytes": 22961
+    }
+  ]
+}