{
  "mode": "hook",
  "metadata": {
    "created_unix": 1781909611.9385035,
    "created_utc": "2026-06-19T22:53:31Z",
    "python": "3.12.3 (main, Aug 14 2025, 17:47:21) [GCC 13.3.0]",
    "platform": "Linux-6.8.0-62-generic-x86_64-with-glibc2.39",
    "torch_version": "2.8.0+cu128",
    "transformers_version": "5.12.1",
    "bitsandbytes_version": "0.49.2",
    "cuda": {
      "available": true,
      "torch_cuda": "12.8",
      "device_count": 1,
      "devices": [
        {
          "index": 0,
          "name": "NVIDIA RTX 6000 Ada Generation",
          "total_memory_gb": 47.38275146484375,
          "major": 8,
          "minor": 9
        }
      ]
    },
    "loader_name": "AutoModelForMultimodalLM",
    "script": "/workspace/latent_qwen_fastweight_experiment.py",
    "script_sha256": "4b46a19f2b91b7acd9d1849dfbf0962822faf615aba712a449b93cf241013392",
    "args": {
      "model_id": "Qwen/Qwen3.5-4B",
      "text_model_fallback": "Qwen/Qwen3-4B",
      "model_class": "auto",
      "mode": "hook",
      "load_in_4bit": true,
      "torch_dtype": "bf16",
      "device_map": "auto",
      "seed": 7,
      "verbose": false,
      "modulus": 97,
      "num_choices": 5,
      "min_steps": 3,
      "max_steps": 8,
      "max_length": 384,
      "hook_layer": -4,
      "runtime_dim": 256,
      "workspace_tokens": 8,
      "runtime_heads": 4,
      "num_bases": 12,
      "adapter_rank": 16,
      "mem_dim": 128,
      "dropout": 0.05,
      "init_scale": 0.05,
      "disable_fast_memory": false,
      "disable_dynamic_lowrank": false,
      "train_steps": 300,
      "batch_size": 4,
      "grad_accum": 1,
      "lr": 0.0002,
      "weight_decay": 0.01,
      "max_grad_norm": 1.0,
      "train_k": "1,2,4",
      "eval_k": "0,1,2,4,8",
      "gate_entropy_bonus": 0.0,
      "log_every": 10,
      "eval_every": 50,
      "eval_batches": 25,
      "eval_batch_size": 4,
      "candidate_batch_size": 20,
      "skip_initial_eval": false,
      "output_dir": "runs/main_qwen35_hook_full_seed7"
    }
  },
  "architecture": {
    "layer_path": "model.language_model.layers",
    "num_layers": 32,
    "hook_layer_index": 28,
    "hidden_dim": 2560,
    "runtime_dim": 256,
    "workspace_tokens": 8,
    "runtime_heads": 4,
    "num_bases": 12,
    "adapter_rank": 16,
    "mem_dim": 128,
    "disable_fast_memory": false,
    "disable_dynamic_lowrank": false
  },
  "train": [
    {
      "step": 1,
      "loss": 4.1179070472717285,
      "delta_rms": 1.3955565918877255e-05,
      "scale": 0.05000000074505806,
      "gate_entropy": 2.461815357208252,
      "mem_rms": 0.431740403175354
    },
    {
      "step": 10,
      "loss": 3.4086313247680664,
      "delta_rms": 0.012721618637442589,
      "scale": 0.050080280750989914,
      "gate_entropy": 2.4701006412506104,
      "mem_rms": 0.3411058485507965
    },
    {
      "step": 20,
      "loss": 1.7132983207702637,
      "delta_rms": 0.08033158630132675,
      "scale": 0.05018859729170799,
      "gate_entropy": 2.464916229248047,
      "mem_rms": 0.14326994121074677
    },
    {
      "step": 30,
      "loss": 1.5163843631744385,
      "delta_rms": 0.18750716745853424,
      "scale": 0.050285711884498596,
      "gate_entropy": 2.4601478576660156,
      "mem_rms": 0.119301438331604
    },
    {
      "step": 40,
      "loss": 1.5525873899459839,
      "delta_rms": 0.248764768242836,
      "scale": 0.050316039472818375,
      "gate_entropy": 2.450620412826538,
      "mem_rms": 0.15725922584533691
    },
    {
      "step": 50,
      "loss": 1.4839179515838623,
      "delta_rms": 0.2093852162361145,
      "scale": 0.050315164029598236,
      "gate_entropy": 2.423682689666748,
      "mem_rms": 0.20455847680568695
    },
    {
      "step": 60,
      "loss": 1.5924322605133057,
      "delta_rms": 0.16070067882537842,
      "scale": 0.05030990019440651,
      "gate_entropy": 2.3451859951019287,
      "mem_rms": 0.1522810161113739
    },
    {
      "step": 70,
      "loss": 1.5945180654525757,
      "delta_rms": 0.1905737668275833,
      "scale": 0.05031182989478111,
      "gate_entropy": 2.34202241897583,
      "mem_rms": 0.058437999337911606
    },
    {
      "step": 80,
      "loss": 1.5633034706115723,
      "delta_rms": 0.22426706552505493,
      "scale": 0.050317276269197464,
      "gate_entropy": 2.363554000854492,
      "mem_rms": 0.030576979741454124
    },
    {
      "step": 90,
      "loss": 1.6747488975524902,
      "delta_rms": 0.23694391548633575,
      "scale": 0.05032183229923248,
      "gate_entropy": 2.3696722984313965,
      "mem_rms": 0.029808247461915016
    },
    {
      "step": 100,
      "loss": 1.5857124328613281,
      "delta_rms": 0.210680291056633,
      "scale": 0.05031435936689377,
      "gate_entropy": 2.3244524002075195,
      "mem_rms": 0.022638775408267975
    },
    {
      "step": 110,
      "loss": 1.4105584621429443,
      "delta_rms": 0.21820737421512604,
      "scale": 0.05031595751643181,
      "gate_entropy": 2.329282283782959,
      "mem_rms": 0.02627759799361229
    },
    {
      "step": 120,
      "loss": 1.564353346824646,
      "delta_rms": 0.22774063050746918,
      "scale": 0.050319209694862366,
      "gate_entropy": 2.375620126724243,
      "mem_rms": 0.02275921404361725
    },
    {
      "step": 130,
      "loss": 1.6887460947036743,
      "delta_rms": 0.1857813149690628,
      "scale": 0.05031263083219528,
      "gate_entropy": 2.3732571601867676,
      "mem_rms": 0.03188639506697655
    },
    {
      "step": 140,
      "loss": 1.4794766902923584,
      "delta_rms": 0.15100950002670288,
      "scale": 0.050300270318984985,
      "gate_entropy": 2.371335744857788,
      "mem_rms": 0.027854913845658302
    },
    {
      "step": 150,
      "loss": 1.4182813167572021,
      "delta_rms": 0.1503198891878128,
      "scale": 0.05029267445206642,
      "gate_entropy": 2.3611574172973633,
      "mem_rms": 0.033884141594171524
    },
    {
      "step": 160,
      "loss": 1.5088914632797241,
      "delta_rms": 0.15904290974140167,
      "scale": 0.050296153873205185,
      "gate_entropy": 2.360133647918701,
      "mem_rms": 0.03212420642375946
    },
    {
      "step": 170,
      "loss": 1.6486318111419678,
      "delta_rms": 0.19038191437721252,
      "scale": 0.0503113828599453,
      "gate_entropy": 2.373371124267578,
      "mem_rms": 0.03539883345365524
    },
    {
      "step": 180,
      "loss": 1.4767765998840332,
      "delta_rms": 0.18687804043293,
      "scale": 0.05031079798936844,
      "gate_entropy": 2.3620920181274414,
      "mem_rms": 0.03350771591067314
    },
    {
      "step": 190,
      "loss": 1.5936626195907593,
      "delta_rms": 0.1671178936958313,
      "scale": 0.050301384180784225,
      "gate_entropy": 2.3203866481781006,
      "mem_rms": 0.04224289953708649
    },
    {
      "step": 200,
      "loss": 1.6749637126922607,
      "delta_rms": 0.11977526545524597,
      "scale": 0.050277892500162125,
      "gate_entropy": 2.2138314247131348,
      "mem_rms": 0.050439268350601196
    },
    {
      "step": 210,
      "loss": 1.518019437789917,
      "delta_rms": 0.13686852157115936,
      "scale": 0.05028434470295906,
      "gate_entropy": 2.257627487182617,
      "mem_rms": 0.03846197575330734
    },
    {
      "step": 220,
      "loss": 1.506201148033142,
      "delta_rms": 0.15822093188762665,
      "scale": 0.0503004714846611,
      "gate_entropy": 2.273688316345215,
      "mem_rms": 0.038562893867492676
    },
    {
      "step": 230,
      "loss": 1.7350894212722778,
      "delta_rms": 0.199442520737648,
      "scale": 0.05031630024313927,
      "gate_entropy": 2.2926459312438965,
      "mem_rms": 0.03590347617864609
    },
    {
      "step": 240,
      "loss": 1.6355184316635132,
      "delta_rms": 0.21655425429344177,
      "scale": 0.05032328516244888,
      "gate_entropy": 2.300745725631714,
      "mem_rms": 0.03473370149731636
    },
    {
      "step": 250,
      "loss": 1.4418703317642212,
      "delta_rms": 0.19791527092456818,
      "scale": 0.05031963810324669,
      "gate_entropy": 2.308685302734375,
      "mem_rms": 0.030125273391604424
    },
    {
      "step": 260,
      "loss": 1.1781446933746338,
      "delta_rms": 0.21698921918869019,
      "scale": 0.05033188685774803,
      "gate_entropy": 2.32378888130188,
      "mem_rms": 0.030886933207511902
    },
    {
      "step": 270,
      "loss": 1.6503232717514038,
      "delta_rms": 0.2092040479183197,
      "scale": 0.05032594874501228,
      "gate_entropy": 2.339421510696411,
      "mem_rms": 0.027097051963210106
    },
    {
      "step": 280,
      "loss": 1.5445903539657593,
      "delta_rms": 0.167969211935997,
      "scale": 0.05030287057161331,
      "gate_entropy": 2.3359415531158447,
      "mem_rms": 0.03109603561460972
    },
    {
      "step": 290,
      "loss": 1.6370047330856323,
      "delta_rms": 0.1608499437570572,
      "scale": 0.05030300095677376,
      "gate_entropy": 2.335318088531494,
      "mem_rms": 0.030492864549160004
    },
    {
      "step": 300,
      "loss": 1.673874020576477,
      "delta_rms": 0.1472472995519638,
      "scale": 0.05029742792248726,
      "gate_entropy": 2.3424131870269775,
      "mem_rms": 0.02865893393754959
    }
  ],
  "eval": [
    {
      "step": 0,
      "base_val": 0.18,
      "base_hard": 0.21,
      "val": {
        "0": 0.19,
        "1": 0.19,
        "2": 0.18,
        "4": 0.18,
        "8": 0.19
      },
      "hard": {
        "0": 0.28,
        "1": 0.26,
        "2": 0.26,
        "4": 0.29,
        "8": 0.29
      }
    },
    {
      "step": 50,
      "val": {
        "0": 0.24,
        "1": 0.26,
        "2": 0.27,
        "4": 0.27,
        "8": 0.25
      },
      "hard": {
        "0": 0.26,
        "1": 0.28,
        "2": 0.28,
        "4": 0.27,
        "8": 0.27
      }
    },
    {
      "step": 100,
      "val": {
        "0": 0.21,
        "1": 0.23,
        "2": 0.23,
        "4": 0.22,
        "8": 0.22
      },
      "hard": {
        "0": 0.25,
        "1": 0.24,
        "2": 0.23,
        "4": 0.24,
        "8": 0.23
      }
    },
    {
      "step": 150,
      "val": {
        "0": 0.22,
        "1": 0.25,
        "2": 0.28,
        "4": 0.25,
        "8": 0.24
      },
      "hard": {
        "0": 0.29,
        "1": 0.32,
        "2": 0.31,
        "4": 0.27,
        "8": 0.28
      }
    },
    {
      "step": 200,
      "val": {
        "0": 0.23,
        "1": 0.27,
        "2": 0.26,
        "4": 0.25,
        "8": 0.27
      },
      "hard": {
        "0": 0.22,
        "1": 0.27,
        "2": 0.24,
        "4": 0.26,
        "8": 0.26
      }
    },
    {
      "step": 250,
      "val": {
        "0": 0.18,
        "1": 0.24,
        "2": 0.21,
        "4": 0.21,
        "8": 0.24
      },
      "hard": {
        "0": 0.24,
        "1": 0.23,
        "2": 0.23,
        "4": 0.21,
        "8": 0.24
      }
    },
    {
      "step": 300,
      "val": {
        "0": 0.26,
        "1": 0.22,
        "2": 0.21,
        "4": 0.22,
        "8": 0.24
      },
      "hard": {
        "0": 0.15,
        "1": 0.15,
        "2": 0.15,
        "4": 0.14,
        "8": 0.15
      }
    }
  ]
}