{
  "format_version": 1,
  "architecture": "smollm2-memory-fusion-sequential-accepted-prefix",
  "base_model": "HuggingFaceTB/SmolLM2-135M",
  "source_repository": "https://github.com/vtavakkoli/TinyCeNN-LM",
  "accepted_layers": [
    0
  ],
  "num_hidden_layers": 30,
  "memory_fusion": {
    "feature_dim": 32,
    "memory_rank": 64,
    "dilations": [
      1,
      2,
      4,
      8,
      16,
      32,
      64,
      128
    ],
    "shifted_window": 8,
    "train_output_projection": true
  },
  "last_accepted_report": {
    "layer": 0,
    "accepted": true,
    "steps": 125,
    "nmse": 0.018665021285414696,
    "cosine": 0.9919254779815674,
    "probe_nll": 2.9270507097244263,
    "incremental_delta_nll": 0.01395869255065918,
    "cumulative_delta_nll": 0.01395869255065918,
    "round": 4
  },
  "trainer_status_at_export": {
    "status": "current_layer_needs_more_training",
    "accepted_layers": [
      0
    ],
    "current_layer": 1,
    "rounds_completed": 12,
    "last_report": {
      "layer": 1,
      "accepted": false,
      "steps": 300,
      "nmse": 0.04072427377104759,
      "cosine": 0.9802741408348083,
      "probe_nll": 2.9506284594535828,
      "incremental_delta_nll": 0.023577749729156494,
      "cumulative_delta_nll": 0.03668522834777832
    },
    "message": "No next Transformer layer was replaced. Rerun to continue this same layer."
  },
  "important": "Only formally accepted layers are included. Any sequential_in_progress layer is excluded."
}