bge-small-hotpotwa-matryoshka-fine-tuned-50/checkpoint-500/trainer_state.json

{
  "best_metric": 0.8829862754377662,
  "best_model_checkpoint": "bge-small-hotpotwa-matryoshka-fine-tuned-50/checkpoint-500",
  "epoch": 3.3655868742111905,
  "eval_steps": 50,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.33655868742111905,
      "grad_norm": 1.810544490814209,
      "learning_rate": 3.3783783783783788e-06,
      "loss": 19.5492,
      "step": 50
    },
    {
      "epoch": 0.33655868742111905,
      "eval_dim_128_cosine_accuracy": 0.9584713677236157,
      "eval_dim_128_dot_accuracy": 0.06874112636062471,
      "eval_dim_128_euclidean_accuracy": 0.9569332702318978,
      "eval_dim_128_manhattan_accuracy": 0.9557501183151916,
      "eval_dim_128_max_accuracy": 0.9584713677236157,
      "eval_dim_256_cosine_accuracy": 0.9656885944155229,
      "eval_dim_256_dot_accuracy": 0.03726928537624231,
      "eval_dim_256_euclidean_accuracy": 0.9667534311405585,
      "eval_dim_256_manhattan_accuracy": 0.9635589209654519,
      "eval_dim_256_max_accuracy": 0.9667534311405585,
      "eval_dim_384_cosine_accuracy": 0.966280170373876,
      "eval_dim_384_dot_accuracy": 0.03371982962612399,
      "eval_dim_384_euclidean_accuracy": 0.966280170373876,
      "eval_dim_384_manhattan_accuracy": 0.9650970184571699,
      "eval_dim_384_max_accuracy": 0.966280170373876,
      "eval_dim_64_cosine_accuracy": 0.943208707998107,
      "eval_dim_64_dot_accuracy": 0.1025792711784193,
      "eval_dim_64_euclidean_accuracy": 0.941670610506389,
      "eval_dim_64_manhattan_accuracy": 0.9364647420728821,
      "eval_dim_64_max_accuracy": 0.943208707998107,
      "eval_loss": 19.260427474975586,
      "eval_runtime": 113.0655,
      "eval_samples_per_second": 74.753,
      "eval_sequential_score": 0.943208707998107,
      "eval_steps_per_second": 2.344,
      "step": 50
    },
    {
      "epoch": 0.6731173748422381,
      "grad_norm": 1.9286330938339233,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 19.1976,
      "step": 100
    },
    {
      "epoch": 0.6731173748422381,
      "eval_dim_128_cosine_accuracy": 0.9358731661145291,
      "eval_dim_128_dot_accuracy": 0.0698059630856602,
      "eval_dim_128_euclidean_accuracy": 0.9371746332229058,
      "eval_dim_128_manhattan_accuracy": 0.9326786559394227,
      "eval_dim_128_max_accuracy": 0.9371746332229058,
      "eval_dim_256_cosine_accuracy": 0.9391859914813062,
      "eval_dim_256_dot_accuracy": 0.06282536677709417,
      "eval_dim_256_euclidean_accuracy": 0.9390676762896356,
      "eval_dim_256_manhattan_accuracy": 0.9387127307146238,
      "eval_dim_256_max_accuracy": 0.9391859914813062,
      "eval_dim_384_cosine_accuracy": 0.9424988168480833,
      "eval_dim_384_dot_accuracy": 0.057501183151916706,
      "eval_dim_384_euclidean_accuracy": 0.9424988168480833,
      "eval_dim_384_manhattan_accuracy": 0.9403691433980123,
      "eval_dim_384_max_accuracy": 0.9424988168480833,
      "eval_dim_64_cosine_accuracy": 0.9275911026975864,
      "eval_dim_64_dot_accuracy": 0.08021769995267393,
      "eval_dim_64_euclidean_accuracy": 0.92830099384761,
      "eval_dim_64_manhattan_accuracy": 0.9260530052058684,
      "eval_dim_64_max_accuracy": 0.92830099384761,
      "eval_loss": 18.295848846435547,
      "eval_runtime": 112.3139,
      "eval_samples_per_second": 75.253,
      "eval_sequential_score": 0.9275911026975864,
      "eval_steps_per_second": 2.359,
      "step": 100
    },
    {
      "epoch": 1.0096760622633572,
      "grad_norm": 1.4801104068756104,
      "learning_rate": 1.0135135135135136e-05,
      "loss": 18.4746,
      "step": 150
    },
    {
      "epoch": 1.0096760622633572,
      "eval_dim_128_cosine_accuracy": 0.9053478466635116,
      "eval_dim_128_dot_accuracy": 0.09548035967818268,
      "eval_dim_128_euclidean_accuracy": 0.9085423568386181,
      "eval_dim_128_manhattan_accuracy": 0.9060577378135353,
      "eval_dim_128_max_accuracy": 0.9085423568386181,
      "eval_dim_256_cosine_accuracy": 0.9074775201135826,
      "eval_dim_256_dot_accuracy": 0.09264079507808802,
      "eval_dim_256_euclidean_accuracy": 0.9093705631803124,
      "eval_dim_256_manhattan_accuracy": 0.907950780880265,
      "eval_dim_256_max_accuracy": 0.9093705631803124,
      "eval_dim_384_cosine_accuracy": 0.9085423568386181,
      "eval_dim_384_dot_accuracy": 0.09145764316138193,
      "eval_dim_384_euclidean_accuracy": 0.9085423568386181,
      "eval_dim_384_manhattan_accuracy": 0.9081874112636062,
      "eval_dim_384_max_accuracy": 0.9085423568386181,
      "eval_dim_64_cosine_accuracy": 0.8995504022716517,
      "eval_dim_64_dot_accuracy": 0.10234264079507809,
      "eval_dim_64_euclidean_accuracy": 0.9028632276384287,
      "eval_dim_64_manhattan_accuracy": 0.900970184571699,
      "eval_dim_64_max_accuracy": 0.9028632276384287,
      "eval_loss": 16.984577178955078,
      "eval_runtime": 112.2207,
      "eval_samples_per_second": 75.316,
      "eval_sequential_score": 0.8995504022716517,
      "eval_steps_per_second": 2.361,
      "step": 150
    },
    {
      "epoch": 1.3462347496844762,
      "grad_norm": 1.4812753200531006,
      "learning_rate": 1.3513513513513515e-05,
      "loss": 18.0684,
      "step": 200
    },
    {
      "epoch": 1.3462347496844762,
      "eval_dim_128_cosine_accuracy": 0.9029815428300993,
      "eval_dim_128_dot_accuracy": 0.09737340274491245,
      "eval_dim_128_euclidean_accuracy": 0.900378608613346,
      "eval_dim_128_manhattan_accuracy": 0.9004969238050166,
      "eval_dim_128_max_accuracy": 0.9029815428300993,
      "eval_dim_256_cosine_accuracy": 0.9051112162801703,
      "eval_dim_256_dot_accuracy": 0.09512541410317085,
      "eval_dim_256_euclidean_accuracy": 0.9034548035967819,
      "eval_dim_256_manhattan_accuracy": 0.9021533364884051,
      "eval_dim_256_max_accuracy": 0.9051112162801703,
      "eval_dim_384_cosine_accuracy": 0.9048745858968291,
      "eval_dim_384_dot_accuracy": 0.09512541410317085,
      "eval_dim_384_euclidean_accuracy": 0.9048745858968291,
      "eval_dim_384_manhattan_accuracy": 0.9029815428300993,
      "eval_dim_384_max_accuracy": 0.9048745858968291,
      "eval_dim_64_cosine_accuracy": 0.8958826313298628,
      "eval_dim_64_dot_accuracy": 0.10506389020350212,
      "eval_dim_64_euclidean_accuracy": 0.8997870326549929,
      "eval_dim_64_manhattan_accuracy": 0.8989588263132986,
      "eval_dim_64_max_accuracy": 0.8997870326549929,
      "eval_loss": 16.686861038208008,
      "eval_runtime": 112.2576,
      "eval_samples_per_second": 75.291,
      "eval_sequential_score": 0.8958826313298628,
      "eval_steps_per_second": 2.361,
      "step": 200
    },
    {
      "epoch": 1.6827934371055953,
      "grad_norm": 1.361006736755371,
      "learning_rate": 1.6891891891891896e-05,
      "loss": 17.8979,
      "step": 250
    },
    {
      "epoch": 1.6827934371055953,
      "eval_dim_128_cosine_accuracy": 0.9016800757217227,
      "eval_dim_128_dot_accuracy": 0.09867486985328916,
      "eval_dim_128_euclidean_accuracy": 0.9020350212967345,
      "eval_dim_128_manhattan_accuracy": 0.9000236630383341,
      "eval_dim_128_max_accuracy": 0.9020350212967345,
      "eval_dim_256_cosine_accuracy": 0.9029815428300993,
      "eval_dim_256_dot_accuracy": 0.09831992427827733,
      "eval_dim_256_euclidean_accuracy": 0.900970184571699,
      "eval_dim_256_manhattan_accuracy": 0.9026265972550875,
      "eval_dim_256_max_accuracy": 0.9029815428300993,
      "eval_dim_384_cosine_accuracy": 0.9015617605300521,
      "eval_dim_384_dot_accuracy": 0.09843823946994794,
      "eval_dim_384_euclidean_accuracy": 0.9015617605300521,
      "eval_dim_384_manhattan_accuracy": 0.9032181732134406,
      "eval_dim_384_max_accuracy": 0.9032181732134406,
      "eval_dim_64_cosine_accuracy": 0.8954093705631803,
      "eval_dim_64_dot_accuracy": 0.10518220539517274,
      "eval_dim_64_euclidean_accuracy": 0.8995504022716517,
      "eval_dim_64_manhattan_accuracy": 0.8978939895882632,
      "eval_dim_64_max_accuracy": 0.8995504022716517,
      "eval_loss": 16.577987670898438,
      "eval_runtime": 112.319,
      "eval_samples_per_second": 75.25,
      "eval_sequential_score": 0.8954093705631803,
      "eval_steps_per_second": 2.359,
      "step": 250
    },
    {
      "epoch": 2.0193521245267143,
      "grad_norm": 1.656162142753601,
      "learning_rate": 1.9999888744757143e-05,
      "loss": 17.7545,
      "step": 300
    },
    {
      "epoch": 2.0193521245267143,
      "eval_dim_128_cosine_accuracy": 0.897657359204922,
      "eval_dim_128_dot_accuracy": 0.10328916232844297,
      "eval_dim_128_euclidean_accuracy": 0.8977756743965926,
      "eval_dim_128_manhattan_accuracy": 0.896710837671557,
      "eval_dim_128_max_accuracy": 0.8977756743965926,
      "eval_dim_256_cosine_accuracy": 0.8990771415049692,
      "eval_dim_256_dot_accuracy": 0.10210601041173686,
      "eval_dim_256_euclidean_accuracy": 0.8981306199716044,
      "eval_dim_256_manhattan_accuracy": 0.8968291528632276,
      "eval_dim_256_max_accuracy": 0.8990771415049692,
      "eval_dim_384_cosine_accuracy": 0.8983672503549456,
      "eval_dim_384_dot_accuracy": 0.10163274964505442,
      "eval_dim_384_euclidean_accuracy": 0.8983672503549456,
      "eval_dim_384_manhattan_accuracy": 0.897657359204922,
      "eval_dim_384_max_accuracy": 0.8983672503549456,
      "eval_dim_64_cosine_accuracy": 0.8924514907714151,
      "eval_dim_64_dot_accuracy": 0.10884997633696167,
      "eval_dim_64_euclidean_accuracy": 0.8923331755797445,
      "eval_dim_64_manhattan_accuracy": 0.8930430667297681,
      "eval_dim_64_max_accuracy": 0.8930430667297681,
      "eval_loss": 16.513458251953125,
      "eval_runtime": 111.7362,
      "eval_samples_per_second": 75.642,
      "eval_sequential_score": 0.8924514907714151,
      "eval_steps_per_second": 2.372,
      "step": 300
    },
    {
      "epoch": 2.3559108119478336,
      "grad_norm": 2.0629849433898926,
      "learning_rate": 1.9979730545608128e-05,
      "loss": 17.6046,
      "step": 350
    },
    {
      "epoch": 2.3559108119478336,
      "eval_dim_128_cosine_accuracy": 0.8893752957879791,
      "eval_dim_128_dot_accuracy": 0.11074301940369144,
      "eval_dim_128_euclidean_accuracy": 0.8889020350212967,
      "eval_dim_128_manhattan_accuracy": 0.8898485565546617,
      "eval_dim_128_max_accuracy": 0.8898485565546617,
      "eval_dim_256_cosine_accuracy": 0.8893752957879791,
      "eval_dim_256_dot_accuracy": 0.11086133459536204,
      "eval_dim_256_euclidean_accuracy": 0.8896119261713203,
      "eval_dim_256_manhattan_accuracy": 0.8904401325130147,
      "eval_dim_256_max_accuracy": 0.8904401325130147,
      "eval_dim_384_cosine_accuracy": 0.8906767628963559,
      "eval_dim_384_dot_accuracy": 0.10932323710364411,
      "eval_dim_384_euclidean_accuracy": 0.8906767628963559,
      "eval_dim_384_manhattan_accuracy": 0.8912683388547089,
      "eval_dim_384_max_accuracy": 0.8912683388547089,
      "eval_dim_64_cosine_accuracy": 0.8861807856128727,
      "eval_dim_64_dot_accuracy": 0.11784193090392807,
      "eval_dim_64_euclidean_accuracy": 0.8860624704212021,
      "eval_dim_64_manhattan_accuracy": 0.8861807856128727,
      "eval_dim_64_max_accuracy": 0.8861807856128727,
      "eval_loss": 16.491697311401367,
      "eval_runtime": 112.5812,
      "eval_samples_per_second": 75.075,
      "eval_sequential_score": 0.8861807856128727,
      "eval_steps_per_second": 2.354,
      "step": 350
    },
    {
      "epoch": 2.6924694993689524,
      "grad_norm": 2.6013519763946533,
      "learning_rate": 1.992488554155135e-05,
      "loss": 17.4434,
      "step": 400
    },
    {
      "epoch": 2.6924694993689524,
      "eval_dim_128_cosine_accuracy": 0.8873639375295788,
      "eval_dim_128_dot_accuracy": 0.11512068149550403,
      "eval_dim_128_euclidean_accuracy": 0.8857075248461902,
      "eval_dim_128_manhattan_accuracy": 0.8874822527212494,
      "eval_dim_128_max_accuracy": 0.8874822527212494,
      "eval_dim_256_cosine_accuracy": 0.8861807856128727,
      "eval_dim_256_dot_accuracy": 0.11393752957879792,
      "eval_dim_256_euclidean_accuracy": 0.8868906767628963,
      "eval_dim_256_manhattan_accuracy": 0.8862991008045433,
      "eval_dim_256_max_accuracy": 0.8868906767628963,
      "eval_dim_384_cosine_accuracy": 0.8874822527212494,
      "eval_dim_384_dot_accuracy": 0.1125177472787506,
      "eval_dim_384_euclidean_accuracy": 0.8874822527212494,
      "eval_dim_384_manhattan_accuracy": 0.8865357311878845,
      "eval_dim_384_max_accuracy": 0.8874822527212494,
      "eval_dim_64_cosine_accuracy": 0.8858258400378609,
      "eval_dim_64_dot_accuracy": 0.12068149550402271,
      "eval_dim_64_euclidean_accuracy": 0.8855892096545196,
      "eval_dim_64_manhattan_accuracy": 0.885470894462849,
      "eval_dim_64_max_accuracy": 0.8858258400378609,
      "eval_loss": 16.492637634277344,
      "eval_runtime": 111.0561,
      "eval_samples_per_second": 76.106,
      "eval_sequential_score": 0.8858258400378609,
      "eval_steps_per_second": 2.386,
      "step": 400
    },
    {
      "epoch": 3.0290281867900717,
      "grad_norm": 3.0714638233184814,
      "learning_rate": 1.983554435877128e-05,
      "loss": 17.3278,
      "step": 450
    },
    {
      "epoch": 3.0290281867900717,
      "eval_dim_128_cosine_accuracy": 0.8853525792711784,
      "eval_dim_128_dot_accuracy": 0.1160672030288689,
      "eval_dim_128_euclidean_accuracy": 0.8867723615712257,
      "eval_dim_128_manhattan_accuracy": 0.8855892096545196,
      "eval_dim_128_max_accuracy": 0.8867723615712257,
      "eval_dim_256_cosine_accuracy": 0.8860624704212021,
      "eval_dim_256_dot_accuracy": 0.11429247515380975,
      "eval_dim_256_euclidean_accuracy": 0.885470894462849,
      "eval_dim_256_manhattan_accuracy": 0.8864174159962139,
      "eval_dim_256_max_accuracy": 0.8864174159962139,
      "eval_dim_384_cosine_accuracy": 0.8868906767628963,
      "eval_dim_384_dot_accuracy": 0.11310932323710364,
      "eval_dim_384_euclidean_accuracy": 0.8868906767628963,
      "eval_dim_384_manhattan_accuracy": 0.8867723615712257,
      "eval_dim_384_max_accuracy": 0.8868906767628963,
      "eval_dim_64_cosine_accuracy": 0.8859441552295315,
      "eval_dim_64_dot_accuracy": 0.12091812588736393,
      "eval_dim_64_euclidean_accuracy": 0.884287742546143,
      "eval_dim_64_manhattan_accuracy": 0.8847610033128254,
      "eval_dim_64_max_accuracy": 0.8859441552295315,
      "eval_loss": 16.475650787353516,
      "eval_runtime": 112.3325,
      "eval_samples_per_second": 75.241,
      "eval_sequential_score": 0.8859441552295315,
      "eval_steps_per_second": 2.359,
      "step": 450
    },
    {
      "epoch": 3.3655868742111905,
      "grad_norm": 3.793804407119751,
      "learning_rate": 1.9712017522703764e-05,
      "loss": 17.247,
      "step": 500
    },
    {
      "epoch": 3.3655868742111905,
      "eval_dim_128_cosine_accuracy": 0.8829862754377662,
      "eval_dim_128_dot_accuracy": 0.11831519167061051,
      "eval_dim_128_euclidean_accuracy": 0.8836961665877898,
      "eval_dim_128_manhattan_accuracy": 0.8823946994794132,
      "eval_dim_128_max_accuracy": 0.8836961665877898,
      "eval_dim_256_cosine_accuracy": 0.8840511121628017,
      "eval_dim_256_dot_accuracy": 0.11571225745385708,
      "eval_dim_256_euclidean_accuracy": 0.8841694273544723,
      "eval_dim_256_manhattan_accuracy": 0.8851159488878372,
      "eval_dim_256_max_accuracy": 0.8851159488878372,
      "eval_dim_384_cosine_accuracy": 0.8853525792711784,
      "eval_dim_384_dot_accuracy": 0.11464742072882159,
      "eval_dim_384_euclidean_accuracy": 0.8853525792711784,
      "eval_dim_384_manhattan_accuracy": 0.8862991008045433,
      "eval_dim_384_max_accuracy": 0.8862991008045433,
      "eval_dim_64_cosine_accuracy": 0.8815664931377188,
      "eval_dim_64_dot_accuracy": 0.12434926644581164,
      "eval_dim_64_euclidean_accuracy": 0.88180312352106,
      "eval_dim_64_manhattan_accuracy": 0.88180312352106,
      "eval_dim_64_max_accuracy": 0.88180312352106,
      "eval_loss": 16.47345542907715,
      "eval_runtime": 111.9611,
      "eval_samples_per_second": 75.49,
      "eval_sequential_score": 0.8815664931377188,
      "eval_steps_per_second": 2.367,
      "step": 500
    }
  ],
  "logging_steps": 50,
  "max_steps": 2960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}