training_report.json

{
  "model_parameters": {
    "vocab_size": 787,
    "N": 6,
    "d_model": 256,
    "d_ff": 2048,
    "H": 8,
    "dropout": 0.1
  },
  "control_ranges": {
    "Egc": [
      -86,
      86
    ],
    "Egb": [
      -87,
      87
    ],
    "Eea": [
      -33,
      33
    ],
    "Eat": [
      -15,
      15
    ],
    "Nc": [
      -20,
      20
    ],
    "EPS": [
      -51,
      51
    ],
    "Ei": [
      -45,
      45
    ],
    "Xc": [
      -41,
      41
    ]
  },
  "initialization": {
    "source_model_parameters": {
      "vocab_size": 102,
      "N": 6,
      "d_model": 256,
      "d_ff": 2048,
      "H": 8,
      "dropout": 0.1
    },
    "loaded_tensors": 258,
    "randomly_initialized_tensors": [
      "generator.proj.bias",
      "generator.proj.weight",
      "src_embed.0.lut.weight",
      "tgt_embed.0.lut.weight"
    ]
  },
  "best_epoch": 7,
  "best_valid_nll": 1.0740462712092977,
  "history": [
    {
      "epoch": 1,
      "train_nll": 0.4001046397001366,
      "valid_nll": 1.1691341886524127
    },
    {
      "epoch": 2,
      "train_nll": 0.20935184885564267,
      "valid_nll": 1.2250727920262847
    },
    {
      "epoch": 3,
      "train_nll": 0.1799837162392324,
      "valid_nll": 1.203550871665359
    },
    {
      "epoch": 4,
      "train_nll": 0.1625040992593177,
      "valid_nll": 1.1592993346042566
    },
    {
      "epoch": 5,
      "train_nll": 0.14981824498221571,
      "valid_nll": 1.1696538470969007
    },
    {
      "epoch": 6,
      "train_nll": 0.14050288035990963,
      "valid_nll": 1.1980752255786582
    },
    {
      "epoch": 7,
      "train_nll": 0.13263638268426964,
      "valid_nll": 1.0740462712092977
    },
    {
      "epoch": 8,
      "train_nll": 0.125997236210437,
      "valid_nll": 1.0892458209714595
    },
    {
      "epoch": 9,
      "train_nll": 0.12003581032395527,
      "valid_nll": 1.1086516005051001
    },
    {
      "epoch": 10,
      "train_nll": 0.11488247974711449,
      "valid_nll": 1.1837406366467242
    }
  ],
  "sampling_alpha": 0.0,
  "train_examples": 152036,
  "validation_examples": 362,
  "model_label": "Molecular Optimization Transformer polymer-adapted"
}