Upload batch 23/24: vsfc/mBERT/mean/seed_43 ... vsfc/mBERT/mha_attention/seed_43

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mean/seed_43/all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9349336702463676,
+    "eval_loss": 0.23304706811904907,
+    "eval_runtime": 6.0845,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 260.17,
+    "eval_steps_per_second": 2.794,
+    "total_flos": 1022750124392448.0,
+    "train_loss": 0.20186518625235458,
+    "train_runtime": 124.707,
+    "train_samples": 11426,
+    "train_samples_per_second": 274.868,
+    "train_steps_per_second": 5.749
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mean/seed_43/eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9349336702463676,
+    "eval_loss": 0.23304706811904907,
+    "eval_runtime": 6.0845,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 260.17,
+    "eval_steps_per_second": 2.794
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mean/seed_43/train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1022750124392448.0,
+    "train_loss": 0.20186518625235458,
+    "train_runtime": 124.707,
+    "train_samples": 11426,
+    "train_samples_per_second": 274.868,
+    "train_steps_per_second": 5.749
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mean/seed_43/trainer_state.json

@@ -0,0 +1,265 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 717,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.10460251046025104,
+      "grad_norm": 7.961418628692627,
+      "learning_rate": 2.8953974895397488e-05,
+      "loss": 0.5142,
+      "step": 25
+    },
+    {
+      "epoch": 0.20920502092050208,
+      "grad_norm": 4.761366844177246,
+      "learning_rate": 2.7907949790794982e-05,
+      "loss": 0.3188,
+      "step": 50
+    },
+    {
+      "epoch": 0.3138075313807531,
+      "grad_norm": 5.633045673370361,
+      "learning_rate": 2.686192468619247e-05,
+      "loss": 0.2954,
+      "step": 75
+    },
+    {
+      "epoch": 0.41841004184100417,
+      "grad_norm": 3.909522771835327,
+      "learning_rate": 2.5815899581589957e-05,
+      "loss": 0.2775,
+      "step": 100
+    },
+    {
+      "epoch": 0.5230125523012552,
+      "grad_norm": 7.437557697296143,
+      "learning_rate": 2.4769874476987448e-05,
+      "loss": 0.3261,
+      "step": 125
+    },
+    {
+      "epoch": 0.6276150627615062,
+      "grad_norm": 5.51772403717041,
+      "learning_rate": 2.372384937238494e-05,
+      "loss": 0.2706,
+      "step": 150
+    },
+    {
+      "epoch": 0.7322175732217573,
+      "grad_norm": 6.031228065490723,
+      "learning_rate": 2.2677824267782426e-05,
+      "loss": 0.2662,
+      "step": 175
+    },
+    {
+      "epoch": 0.8368200836820083,
+      "grad_norm": 6.335776329040527,
+      "learning_rate": 2.1631799163179917e-05,
+      "loss": 0.2251,
+      "step": 200
+    },
+    {
+      "epoch": 0.9414225941422594,
+      "grad_norm": 5.41457986831665,
+      "learning_rate": 2.0585774058577408e-05,
+      "loss": 0.2258,
+      "step": 225
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9166140240050537,
+      "eval_loss": 0.23878706991672516,
+      "eval_runtime": 6.4509,
+      "eval_samples_per_second": 245.39,
+      "eval_steps_per_second": 2.635,
+      "step": 239
+    },
+    {
+      "epoch": 1.0460251046025104,
+      "grad_norm": 5.0510077476501465,
+      "learning_rate": 1.9539748953974895e-05,
+      "loss": 0.245,
+      "step": 250
+    },
+    {
+      "epoch": 1.1506276150627615,
+      "grad_norm": 4.800015449523926,
+      "learning_rate": 1.8493723849372386e-05,
+      "loss": 0.186,
+      "step": 275
+    },
+    {
+      "epoch": 1.2552301255230125,
+      "grad_norm": 6.838464260101318,
+      "learning_rate": 1.7447698744769877e-05,
+      "loss": 0.1985,
+      "step": 300
+    },
+    {
+      "epoch": 1.3598326359832635,
+      "grad_norm": 9.14803695678711,
+      "learning_rate": 1.6401673640167365e-05,
+      "loss": 0.1933,
+      "step": 325
+    },
+    {
+      "epoch": 1.4644351464435146,
+      "grad_norm": 5.208489418029785,
+      "learning_rate": 1.5355648535564852e-05,
+      "loss": 0.1722,
+      "step": 350
+    },
+    {
+      "epoch": 1.5690376569037658,
+      "grad_norm": 8.175561904907227,
+      "learning_rate": 1.4309623430962344e-05,
+      "loss": 0.1801,
+      "step": 375
+    },
+    {
+      "epoch": 1.6736401673640167,
+      "grad_norm": 5.153775215148926,
+      "learning_rate": 1.3263598326359832e-05,
+      "loss": 0.1638,
+      "step": 400
+    },
+    {
+      "epoch": 1.778242677824268,
+      "grad_norm": 5.193408012390137,
+      "learning_rate": 1.2217573221757323e-05,
+      "loss": 0.1903,
+      "step": 425
+    },
+    {
+      "epoch": 1.8828451882845187,
+      "grad_norm": 4.024351596832275,
+      "learning_rate": 1.1171548117154812e-05,
+      "loss": 0.168,
+      "step": 450
+    },
+    {
+      "epoch": 1.98744769874477,
+      "grad_norm": 5.724711894989014,
+      "learning_rate": 1.0125523012552301e-05,
+      "loss": 0.1652,
+      "step": 475
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.930511686670878,
+      "eval_loss": 0.2159256637096405,
+      "eval_runtime": 6.1246,
+      "eval_samples_per_second": 258.464,
+      "eval_steps_per_second": 2.776,
+      "step": 478
+    },
+    {
+      "epoch": 2.092050209205021,
+      "grad_norm": 5.1875081062316895,
+      "learning_rate": 9.079497907949792e-06,
+      "loss": 0.1067,
+      "step": 500
+    },
+    {
+      "epoch": 2.196652719665272,
+      "grad_norm": 8.933672904968262,
+      "learning_rate": 8.03347280334728e-06,
+      "loss": 0.1266,
+      "step": 525
+    },
+    {
+      "epoch": 2.301255230125523,
+      "grad_norm": 0.36261019110679626,
+      "learning_rate": 6.98744769874477e-06,
+      "loss": 0.1238,
+      "step": 550
+    },
+    {
+      "epoch": 2.405857740585774,
+      "grad_norm": 7.418014049530029,
+      "learning_rate": 5.941422594142259e-06,
+      "loss": 0.1313,
+      "step": 575
+    },
+    {
+      "epoch": 2.510460251046025,
+      "grad_norm": 3.8534774780273438,
+      "learning_rate": 4.895397489539749e-06,
+      "loss": 0.1408,
+      "step": 600
+    },
+    {
+      "epoch": 2.6150627615062763,
+      "grad_norm": 2.7833070755004883,
+      "learning_rate": 3.849372384937238e-06,
+      "loss": 0.132,
+      "step": 625
+    },
+    {
+      "epoch": 2.719665271966527,
+      "grad_norm": 0.3560211658477783,
+      "learning_rate": 2.803347280334728e-06,
+      "loss": 0.1169,
+      "step": 650
+    },
+    {
+      "epoch": 2.8242677824267783,
+      "grad_norm": 3.333390712738037,
+      "learning_rate": 1.7573221757322176e-06,
+      "loss": 0.1277,
+      "step": 675
+    },
+    {
+      "epoch": 2.928870292887029,
+      "grad_norm": 5.79809045791626,
+      "learning_rate": 7.112970711297071e-07,
+      "loss": 0.1303,
+      "step": 700
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9349336702463676,
+      "eval_loss": 0.23304706811904907,
+      "eval_runtime": 6.1255,
+      "eval_samples_per_second": 258.427,
+      "eval_steps_per_second": 2.775,
+      "step": 717
+    },
+    {
+      "epoch": 3.0,
+      "step": 717,
+      "total_flos": 1022750124392448.0,
+      "train_loss": 0.20186518625235458,
+      "train_runtime": 124.707,
+      "train_samples_per_second": 274.868,
+      "train_steps_per_second": 5.749
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 717,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1022750124392448.0,
+  "train_batch_size": 48,
+  "trial_name": null,
+  "trial_params": null
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mha_attention/seed_42/all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9368288060644346,
+    "eval_loss": 0.21603453159332275,
+    "eval_runtime": 6.0304,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 262.502,
+    "eval_steps_per_second": 2.819,
+    "total_flos": 1028856339912744.0,
+    "train_loss": 0.2058336245963763,
+    "train_runtime": 124.06,
+    "train_samples": 11426,
+    "train_samples_per_second": 276.302,
+    "train_steps_per_second": 5.779
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mha_attention/seed_42/eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9368288060644346,
+    "eval_loss": 0.21603453159332275,
+    "eval_runtime": 6.0304,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 262.502,
+    "eval_steps_per_second": 2.819
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mha_attention/seed_42/train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1028856339912744.0,
+    "train_loss": 0.2058336245963763,
+    "train_runtime": 124.06,
+    "train_samples": 11426,
+    "train_samples_per_second": 276.302,
+    "train_steps_per_second": 5.779
+}

mbert_rtx6000_metrics/result_ablation_mbert/vsfc/mBERT/mha_attention/seed_42/trainer_state.json

@@ -0,0 +1,265 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 717,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.10460251046025104,
+      "grad_norm": 10.088802337646484,
+      "learning_rate": 2.8953974895397488e-05,
+      "loss": 0.5373,
+      "step": 25
+    },
+    {
+      "epoch": 0.20920502092050208,
+      "grad_norm": 5.5244975090026855,
+      "learning_rate": 2.7907949790794982e-05,
+      "loss": 0.3461,
+      "step": 50
+    },
+    {
+      "epoch": 0.3138075313807531,
+      "grad_norm": 3.1109814643859863,
+      "learning_rate": 2.686192468619247e-05,
+      "loss": 0.3159,
+      "step": 75
+    },
+    {
+      "epoch": 0.41841004184100417,
+      "grad_norm": 5.578044891357422,
+      "learning_rate": 2.5815899581589957e-05,
+      "loss": 0.2771,
+      "step": 100
+    },
+    {
+      "epoch": 0.5230125523012552,
+      "grad_norm": 7.939838409423828,
+      "learning_rate": 2.4769874476987448e-05,
+      "loss": 0.2752,
+      "step": 125
+    },
+    {
+      "epoch": 0.6276150627615062,
+      "grad_norm": 5.911745548248291,
+      "learning_rate": 2.372384937238494e-05,
+      "loss": 0.2693,
+      "step": 150
+    },
+    {
+      "epoch": 0.7322175732217573,
+      "grad_norm": 5.975719451904297,
+      "learning_rate": 2.2677824267782426e-05,
+      "loss": 0.2555,
+      "step": 175
+    },
+    {
+      "epoch": 0.8368200836820083,
+      "grad_norm": 3.8831708431243896,
+      "learning_rate": 2.1631799163179917e-05,
+      "loss": 0.2843,
+      "step": 200
+    },
+    {
+      "epoch": 0.9414225941422594,
+      "grad_norm": 6.38673734664917,
+      "learning_rate": 2.0585774058577408e-05,
+      "loss": 0.2425,
+      "step": 225
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9286165508528111,
+      "eval_loss": 0.22932249307632446,
+      "eval_runtime": 5.5396,
+      "eval_samples_per_second": 285.762,
+      "eval_steps_per_second": 3.069,
+      "step": 239
+    },
+    {
+      "epoch": 1.0460251046025104,
+      "grad_norm": 7.314534664154053,
+      "learning_rate": 1.9539748953974895e-05,
+      "loss": 0.2014,
+      "step": 250
+    },
+    {
+      "epoch": 1.1506276150627615,
+      "grad_norm": 7.990490436553955,
+      "learning_rate": 1.8493723849372386e-05,
+      "loss": 0.1795,
+      "step": 275
+    },
+    {
+      "epoch": 1.2552301255230125,
+      "grad_norm": 4.692038059234619,
+      "learning_rate": 1.7447698744769877e-05,
+      "loss": 0.1441,
+      "step": 300
+    },
+    {
+      "epoch": 1.3598326359832635,
+      "grad_norm": 5.073693752288818,
+      "learning_rate": 1.6401673640167365e-05,
+      "loss": 0.2142,
+      "step": 325
+    },
+    {
+      "epoch": 1.4644351464435146,
+      "grad_norm": 3.2842681407928467,
+      "learning_rate": 1.5355648535564852e-05,
+      "loss": 0.2001,
+      "step": 350
+    },
+    {
+      "epoch": 1.5690376569037658,
+      "grad_norm": 8.114819526672363,
+      "learning_rate": 1.4309623430962344e-05,
+      "loss": 0.217,
+      "step": 375
+    },
+    {
+      "epoch": 1.6736401673640167,
+      "grad_norm": 2.59684681892395,
+      "learning_rate": 1.3263598326359832e-05,
+      "loss": 0.1884,
+      "step": 400
+    },
+    {
+      "epoch": 1.778242677824268,
+      "grad_norm": 6.29541015625,
+      "learning_rate": 1.2217573221757323e-05,
+      "loss": 0.1667,
+      "step": 425
+    },
+    {
+      "epoch": 1.8828451882845187,
+      "grad_norm": 5.3400044441223145,
+      "learning_rate": 1.1171548117154812e-05,
+      "loss": 0.1645,
+      "step": 450
+    },
+    {
+      "epoch": 1.98744769874477,
+      "grad_norm": 3.297173261642456,
+      "learning_rate": 1.0125523012552301e-05,
+      "loss": 0.2051,
+      "step": 475
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.930511686670878,
+      "eval_loss": 0.21334458887577057,
+      "eval_runtime": 5.9273,
+      "eval_samples_per_second": 267.07,
+      "eval_steps_per_second": 2.868,
+      "step": 478
+    },
+    {
+      "epoch": 2.092050209205021,
+      "grad_norm": 5.004485607147217,
+      "learning_rate": 9.079497907949792e-06,
+      "loss": 0.1254,
+      "step": 500
+    },
+    {
+      "epoch": 2.196652719665272,
+      "grad_norm": 5.266722202301025,
+      "learning_rate": 8.03347280334728e-06,
+      "loss": 0.1307,
+      "step": 525
+    },
+    {
+      "epoch": 2.301255230125523,
+      "grad_norm": 4.00379753112793,
+      "learning_rate": 6.98744769874477e-06,
+      "loss": 0.1438,
+      "step": 550
+    },
+    {
+      "epoch": 2.405857740585774,
+      "grad_norm": 3.726776599884033,
+      "learning_rate": 5.941422594142259e-06,
+      "loss": 0.1244,
+      "step": 575
+    },
+    {
+      "epoch": 2.510460251046025,
+      "grad_norm": 3.872097969055176,
+      "learning_rate": 4.895397489539749e-06,
+      "loss": 0.1006,
+      "step": 600
+    },
+    {
+      "epoch": 2.6150627615062763,
+      "grad_norm": 6.372356414794922,
+      "learning_rate": 3.849372384937238e-06,
+      "loss": 0.1452,
+      "step": 625
+    },
+    {
+      "epoch": 2.719665271966527,
+      "grad_norm": 6.529725551605225,
+      "learning_rate": 2.803347280334728e-06,
+      "loss": 0.1203,
+      "step": 650
+    },
+    {
+      "epoch": 2.8242677824267783,
+      "grad_norm": 1.954482913017273,
+      "learning_rate": 1.7573221757322176e-06,
+      "loss": 0.1316,
+      "step": 675
+    },
+    {
+      "epoch": 2.928870292887029,
+      "grad_norm": 2.8128585815429688,
+      "learning_rate": 7.112970711297071e-07,
+      "loss": 0.1125,
+      "step": 700
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9368288060644346,
+      "eval_loss": 0.21603453159332275,
+      "eval_runtime": 5.7799,
+      "eval_samples_per_second": 273.878,
+      "eval_steps_per_second": 2.941,
+      "step": 717
+    },
+    {
+      "epoch": 3.0,
+      "step": 717,
+      "total_flos": 1028856339912744.0,
+      "train_loss": 0.2058336245963763,
+      "train_runtime": 124.06,
+      "train_samples_per_second": 276.302,
+      "train_steps_per_second": 5.779
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 717,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1028856339912744.0,
+  "train_batch_size": 48,
+  "trial_name": null,
+  "trial_params": null
+}