{
  "mode": "eval_hook_checkpoint",
  "metadata": {
    "created_unix": 1781912001.9351847,
    "created_utc": "2026-06-19T23:33:21Z",
    "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": "bd43bada40db273eb38fdaf7ddb7076f2dda5ef230255068b3e6ad4a77d39a76",
    "args": {
      "model_id": "Qwen/Qwen3.5-4B",
      "text_model_fallback": "Qwen/Qwen3-4B",
      "model_class": "auto",
      "mode": "hook",
      "eval_only_checkpoint": "runs/main_qwen35_hook_aux02_seed7/latent_adapter.pt",
      "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": 1,
      "grad_accum": 8,
      "lr": 0.0002,
      "weight_decay": 0.01,
      "max_grad_norm": 1.0,
      "train_k": "1,2,4",
      "eval_k": "0,2,4,8",
      "aux_value_loss": 0.0,
      "gate_entropy_bonus": 0.0,
      "log_every": 10,
      "eval_every": 50,
      "eval_batches": 50,
      "eval_batch_size": 5,
      "candidate_batch_size": 20,
      "skip_initial_eval": false,
      "output_dir": "runs/eval_aux_final_n250"
    }
  },
  "checkpoint": "runs/main_qwen35_hook_aux02_seed7/latent_adapter.pt",
  "checkpoint_saved_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",
    "aux_value_loss": 0.2,
    "gate_entropy_bonus": 0.0,
    "log_every": 10,
    "eval_every": 100,
    "eval_batches": 25,
    "eval_batch_size": 4,
    "candidate_batch_size": 20,
    "skip_initial_eval": true,
    "output_dir": "runs/main_qwen35_hook_aux02_seed7"
  },
  "architecture": {
    "layer_path": "model.language_model.layers",
    "num_layers": 32,
    "hook_layer_index": 28,
    "hidden_dim": 2560
  },
  "eval": {
    "val": {
      "0": 0.176,
      "2": 0.152,
      "4": 0.168,
      "8": 0.156
    },
    "hard": {
      "0": 0.212,
      "2": 0.228,
      "4": 0.204,
      "8": 0.228
    },
    "eval_batches": 50,
    "eval_batch_size": 5,
    "num_examples_per_split": 250
  }
}