Upload batch 11/14: sst2/mBERT/multi_branch_average/seed_43 ... vsfc/mBERT/cls/seed_43

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_42/all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.930511686670878,
+    "eval_loss": 0.2223110944032669,
+    "eval_runtime": 6.1594,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 257.004,
+    "eval_steps_per_second": 1.624,
+    "total_flos": 1153409273949996.0,
+    "train_loss": 0.22780775857138466,
+    "train_runtime": 62.1063,
+    "train_samples": 11426,
+    "train_samples_per_second": 551.925,
+    "train_steps_per_second": 6.908
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_42/eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.930511686670878,
+    "eval_loss": 0.2223110944032669,
+    "eval_runtime": 6.1594,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 257.004,
+    "eval_steps_per_second": 1.624
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_42/train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1153409273949996.0,
+    "train_loss": 0.22780775857138466,
+    "train_runtime": 62.1063,
+    "train_samples": 11426,
+    "train_samples_per_second": 551.925,
+    "train_steps_per_second": 6.908
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_42/trainer_state.json

@@ -0,0 +1,125 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 429,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.34965034965034963,
+      "grad_norm": 2.9751393795013428,
+      "learning_rate": 1.766899766899767e-05,
+      "loss": 0.4517,
+      "step": 50
+    },
+    {
+      "epoch": 0.6993006993006993,
+      "grad_norm": 4.006201267242432,
+      "learning_rate": 1.533799533799534e-05,
+      "loss": 0.2925,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9260897030953885,
+      "eval_loss": 0.22814147174358368,
+      "eval_runtime": 6.0441,
+      "eval_samples_per_second": 261.908,
+      "eval_steps_per_second": 1.655,
+      "step": 143
+    },
+    {
+      "epoch": 1.048951048951049,
+      "grad_norm": 5.519968509674072,
+      "learning_rate": 1.3006993006993008e-05,
+      "loss": 0.2498,
+      "step": 150
+    },
+    {
+      "epoch": 1.3986013986013985,
+      "grad_norm": 4.9121832847595215,
+      "learning_rate": 1.0675990675990678e-05,
+      "loss": 0.189,
+      "step": 200
+    },
+    {
+      "epoch": 1.7482517482517483,
+      "grad_norm": 3.908798933029175,
+      "learning_rate": 8.344988344988347e-06,
+      "loss": 0.2085,
+      "step": 250
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.930511686670878,
+      "eval_loss": 0.21119962632656097,
+      "eval_runtime": 6.5516,
+      "eval_samples_per_second": 241.62,
+      "eval_steps_per_second": 1.526,
+      "step": 286
+    },
+    {
+      "epoch": 2.097902097902098,
+      "grad_norm": 3.7348828315734863,
+      "learning_rate": 6.013986013986014e-06,
+      "loss": 0.1721,
+      "step": 300
+    },
+    {
+      "epoch": 2.4475524475524475,
+      "grad_norm": 2.9933745861053467,
+      "learning_rate": 3.6829836829836835e-06,
+      "loss": 0.1557,
+      "step": 350
+    },
+    {
+      "epoch": 2.797202797202797,
+      "grad_norm": 4.798852443695068,
+      "learning_rate": 1.3519813519813522e-06,
+      "loss": 0.1483,
+      "step": 400
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.930511686670878,
+      "eval_loss": 0.2223110944032669,
+      "eval_runtime": 6.398,
+      "eval_samples_per_second": 247.42,
+      "eval_steps_per_second": 1.563,
+      "step": 429
+    },
+    {
+      "epoch": 3.0,
+      "step": 429,
+      "total_flos": 1153409273949996.0,
+      "train_loss": 0.22780775857138466,
+      "train_runtime": 62.1063,
+      "train_samples_per_second": 551.925,
+      "train_steps_per_second": 6.908
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 429,
+  "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": 1153409273949996.0,
+  "train_batch_size": 80,
+  "trial_name": null,
+  "trial_params": null
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_43/all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9336702463676564,
+    "eval_loss": 0.21957743167877197,
+    "eval_runtime": 5.9827,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 264.596,
+    "eval_steps_per_second": 1.671,
+    "total_flos": 1153679794377588.0,
+    "train_loss": 0.22140424123732916,
+    "train_runtime": 60.9769,
+    "train_samples": 11426,
+    "train_samples_per_second": 562.147,
+    "train_steps_per_second": 7.035
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_43/eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9336702463676564,
+    "eval_loss": 0.21957743167877197,
+    "eval_runtime": 5.9827,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 264.596,
+    "eval_steps_per_second": 1.671
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_43/train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1153679794377588.0,
+    "train_loss": 0.22140424123732916,
+    "train_runtime": 60.9769,
+    "train_samples": 11426,
+    "train_samples_per_second": 562.147,
+    "train_steps_per_second": 7.035
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/attention/seed_43/trainer_state.json

@@ -0,0 +1,125 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 429,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.34965034965034963,
+      "grad_norm": 5.305752754211426,
+      "learning_rate": 1.766899766899767e-05,
+      "loss": 0.4126,
+      "step": 50
+    },
+    {
+      "epoch": 0.6993006993006993,
+      "grad_norm": 2.7622997760772705,
+      "learning_rate": 1.533799533799534e-05,
+      "loss": 0.2827,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9121920404295641,
+      "eval_loss": 0.24175824224948883,
+      "eval_runtime": 6.1218,
+      "eval_samples_per_second": 258.586,
+      "eval_steps_per_second": 1.634,
+      "step": 143
+    },
+    {
+      "epoch": 1.048951048951049,
+      "grad_norm": 2.03625750541687,
+      "learning_rate": 1.3006993006993008e-05,
+      "loss": 0.2435,
+      "step": 150
+    },
+    {
+      "epoch": 1.3986013986013985,
+      "grad_norm": 6.116645336151123,
+      "learning_rate": 1.0675990675990678e-05,
+      "loss": 0.211,
+      "step": 200
+    },
+    {
+      "epoch": 1.7482517482517483,
+      "grad_norm": 2.7598345279693604,
+      "learning_rate": 8.344988344988347e-06,
+      "loss": 0.1971,
+      "step": 250
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9336702463676564,
+      "eval_loss": 0.2090362310409546,
+      "eval_runtime": 6.8705,
+      "eval_samples_per_second": 230.406,
+      "eval_steps_per_second": 1.455,
+      "step": 286
+    },
+    {
+      "epoch": 2.097902097902098,
+      "grad_norm": 4.960213661193848,
+      "learning_rate": 6.013986013986014e-06,
+      "loss": 0.1732,
+      "step": 300
+    },
+    {
+      "epoch": 2.4475524475524475,
+      "grad_norm": 4.121504783630371,
+      "learning_rate": 3.6829836829836835e-06,
+      "loss": 0.1475,
+      "step": 350
+    },
+    {
+      "epoch": 2.797202797202797,
+      "grad_norm": 4.008220195770264,
+      "learning_rate": 1.3519813519813522e-06,
+      "loss": 0.1514,
+      "step": 400
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9336702463676564,
+      "eval_loss": 0.21957743167877197,
+      "eval_runtime": 6.849,
+      "eval_samples_per_second": 231.128,
+      "eval_steps_per_second": 1.46,
+      "step": 429
+    },
+    {
+      "epoch": 3.0,
+      "step": 429,
+      "total_flos": 1153679794377588.0,
+      "train_loss": 0.22140424123732916,
+      "train_runtime": 60.9769,
+      "train_samples_per_second": 562.147,
+      "train_steps_per_second": 7.035
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 429,
+  "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": 1153679794377588.0,
+  "train_batch_size": 80,
+  "trial_name": null,
+  "trial_params": null
+}