docs/results/hf_training_status.json

{
  "status": "ok",
  "started_at": 1777180786.0648105,
  "finished_at": 1777188659.441074,
  "commands": [
    {
      "args": [
        "python",
        "scripts/bootstrap_data.py"
      ],
      "returncode": 0,
      "elapsed_seconds": 0.507
    },
    {
      "args": [
        "python",
        "scripts/build_training_corpus.py",
        "--profile",
        "massive",
        "--with-local",
        "--with-synthetic",
        "--with-hf"
      ],
      "returncode": 0,
      "elapsed_seconds": 3.695
    },
    {
      "args": [
        "python",
        "scripts/train_sft_trl.py",
        "--model-id",
        "Qwen/Qwen2.5-3B-Instruct",
        "--dataset-path",
        "data/processed/training_corpus_sft.json",
        "--output-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct",
        "--report-path",
        "outputs/reports/sweeps/qwen-qwen2-5-3b-instruct/sft_trl_run.json",
        "--epochs",
        "2",
        "--max-steps",
        "0",
        "--batch-size",
        "2",
        "--max-seq-len",
        "512",
        "--learning-rate",
        "2e-05",
        "--use-unsloth"
      ],
      "returncode": 0,
      "elapsed_seconds": 737.28
    },
    {
      "args": [
        "python",
        "scripts/train_grpo_trl.py",
        "--model-id",
        "Qwen/Qwen2.5-3B-Instruct",
        "--prompts-path",
        "data/processed/training_corpus_grpo_prompts.jsonl",
        "--output-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct",
        "--report-path",
        "outputs/reports/sweeps/qwen-qwen2-5-3b-instruct/grpo_trl_run.json",
        "--max-prompts",
        "0",
        "--max-steps",
        "0",
        "--epochs",
        "1.0",
        "--batch-size",
        "2",
        "--grad-accum",
        "1",
        "--num-generations",
        "2",
        "--max-prompt-length",
        "384",
        "--max-completion-length",
        "64",
        "--learning-rate",
        "1e-06",
        "--use-unsloth"
      ],
      "returncode": 0,
      "elapsed_seconds": 6885.399
    },
    {
      "args": [
        "python",
        "scripts/merge_adapters_safe.py",
        "--adapter-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/sft_adapter",
        "--output-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/merged"
      ],
      "returncode": 0,
      "elapsed_seconds": 15.74
    },
    {
      "args": [
        "python",
        "scripts/test_inference_postsave.py",
        "--samples",
        "5",
        "--base-model",
        "Qwen/Qwen2.5-3B-Instruct",
        "--merged-model",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/merged",
        "--adapter-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/sft_adapter",
        "--output",
        "outputs/reports/sweeps/qwen-qwen2-5-3b-instruct/postsave_inference_sft.json"
      ],
      "returncode": 0,
      "elapsed_seconds": 20.985
    },
    {
      "args": [
        "python",
        "scripts/test_inference_postsave.py",
        "--samples",
        "5",
        "--base-model",
        "Qwen/Qwen2.5-3B-Instruct",
        "--merged-model",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/missing_merged_grpo",
        "--adapter-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct/grpo_adapter",
        "--output",
        "outputs/reports/sweeps/qwen-qwen2-5-3b-instruct/postsave_inference_grpo.json"
      ],
      "returncode": 0,
      "elapsed_seconds": 26.691
    },
    {
      "args": [
        "python",
        "scripts/evaluate_policy_ablations.py",
        "--episodes",
        "8",
        "--checkpoint-dir",
        "checkpoints/sweeps/qwen-qwen2-5-3b-instruct",
        "--output",
        "outputs/reports/sweeps/qwen-qwen2-5-3b-instruct/grpo_ablation_report.json"
      ],
      "returncode": 0,
      "elapsed_seconds": 4.001
    },
    {
      "args": [
        "reuse_artifact",
        "grpo_adapter",
        "/app/checkpoints/grpo_adapter"
      ],
      "returncode": 0,
      "elapsed_seconds": 0.0
    },
    {
      "args": [
        "python",
        "scripts/evaluate_baselines.py"
      ],
      "returncode": 0,
      "elapsed_seconds": 4.163
    },
    {
      "args": [
        "python",
        "scripts/evaluate_all.py"
      ],
      "returncode": 0,
      "elapsed_seconds": 3.798
    },
    {
      "args": [
        "python",
        "scripts/evaluate_compare_runs.py",
        "--baseline",
        "outputs/reports/baselines.json",
        "--candidate",
        "outputs/reports/benchmark_report.json",
        "--output",
        "outputs/reports/improvement_report.json"
      ],
      "returncode": 0,
      "elapsed_seconds": 0.034
    },
    {
      "args": [
        "python",
        "scripts/benchmark_inference.py"
      ],
      "returncode": 0,
      "elapsed_seconds": 2.39
    },
    {
      "args": [
        "python",
        "scripts/run_robustness_suite.py"
      ],
      "returncode": 0,
      "elapsed_seconds": 2.692
    },
    {
      "args": [
        "python",
        "scripts/generate_hf_training_report.py",
        "--mode",
        "full"
      ],
      "returncode": 0,
      "elapsed_seconds": 2.078
    }
  ],
  "artifact_repo_id": "adithya9903/polyguard-openenv-training-3b-artifacts",
  "training_mode": "full",
  "model_sweep": [
    "Qwen/Qwen2.5-3B-Instruct"
  ],
  "improved": true,
  "anti_hacking_passed": false,
  "completed_run_ids": [
    "qwen-qwen2-5-3b-instruct"
  ]
}