Upload batch 12/14: vsfc/mBERT/cls/seed_44 ... vsfc/mBERT/hf_sequence_classifier/seed_44

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

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9374605180037903,
+    "eval_loss": 0.22037982940673828,
+    "eval_runtime": 6.0069,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 263.528,
+    "eval_steps_per_second": 1.665,
+    "total_flos": 1206171310766856.0,
+    "train_loss": 0.2221188189544322,
+    "train_runtime": 55.1604,
+    "train_samples": 11426,
+    "train_samples_per_second": 621.424,
+    "train_steps_per_second": 7.777
+}

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

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9374605180037903,
+    "eval_loss": 0.22037982940673828,
+    "eval_runtime": 6.0069,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 263.528,
+    "eval_steps_per_second": 1.665
+}

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

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1206171310766856.0,
+    "train_loss": 0.2221188189544322,
+    "train_runtime": 55.1604,
+    "train_samples": 11426,
+    "train_samples_per_second": 621.424,
+    "train_steps_per_second": 7.777
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/gated_multi_branch/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": 5.015577793121338,
+      "learning_rate": 1.766899766899767e-05,
+      "loss": 0.4275,
+      "step": 50
+    },
+    {
+      "epoch": 0.6993006993006993,
+      "grad_norm": 2.219026803970337,
+      "learning_rate": 1.533799533799534e-05,
+      "loss": 0.2963,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.930511686670878,
+      "eval_loss": 0.2218150943517685,
+      "eval_runtime": 6.2094,
+      "eval_samples_per_second": 254.936,
+      "eval_steps_per_second": 1.61,
+      "step": 143
+    },
+    {
+      "epoch": 1.048951048951049,
+      "grad_norm": 7.437995910644531,
+      "learning_rate": 1.3006993006993008e-05,
+      "loss": 0.2454,
+      "step": 150
+    },
+    {
+      "epoch": 1.3986013986013985,
+      "grad_norm": 5.58640718460083,
+      "learning_rate": 1.0675990675990678e-05,
+      "loss": 0.1832,
+      "step": 200
+    },
+    {
+      "epoch": 1.7482517482517483,
+      "grad_norm": 2.857930898666382,
+      "learning_rate": 8.344988344988347e-06,
+      "loss": 0.208,
+      "step": 250
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9273531269740998,
+      "eval_loss": 0.21884030103683472,
+      "eval_runtime": 6.0149,
+      "eval_samples_per_second": 263.18,
+      "eval_steps_per_second": 1.663,
+      "step": 286
+    },
+    {
+      "epoch": 2.097902097902098,
+      "grad_norm": 6.3794097900390625,
+      "learning_rate": 6.013986013986014e-06,
+      "loss": 0.1672,
+      "step": 300
+    },
+    {
+      "epoch": 2.4475524475524475,
+      "grad_norm": 4.895865440368652,
+      "learning_rate": 3.6829836829836835e-06,
+      "loss": 0.1516,
+      "step": 350
+    },
+    {
+      "epoch": 2.797202797202797,
+      "grad_norm": 4.498716831207275,
+      "learning_rate": 1.3519813519813522e-06,
+      "loss": 0.1418,
+      "step": 400
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9374605180037903,
+      "eval_loss": 0.22037982940673828,
+      "eval_runtime": 6.0235,
+      "eval_samples_per_second": 262.806,
+      "eval_steps_per_second": 1.66,
+      "step": 429
+    },
+    {
+      "epoch": 3.0,
+      "step": 429,
+      "total_flos": 1206171310766856.0,
+      "train_loss": 0.2221188189544322,
+      "train_runtime": 55.1604,
+      "train_samples_per_second": 621.424,
+      "train_steps_per_second": 7.777
+    }
+  ],
+  "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": 1206171310766856.0,
+  "train_batch_size": 80,
+  "trial_name": null,
+  "trial_params": null
+}

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

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.932406822488945,
+    "eval_loss": 0.2144102156162262,
+    "eval_runtime": 5.8604,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 270.118,
+    "eval_steps_per_second": 1.706,
+    "total_flos": 1206454205994168.0,
+    "train_loss": 0.22194261373061955,
+    "train_runtime": 55.2665,
+    "train_samples": 11426,
+    "train_samples_per_second": 620.232,
+    "train_steps_per_second": 7.762
+}

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

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.932406822488945,
+    "eval_loss": 0.2144102156162262,
+    "eval_runtime": 5.8604,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 270.118,
+    "eval_steps_per_second": 1.706
+}

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

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1206454205994168.0,
+    "train_loss": 0.22194261373061955,
+    "train_runtime": 55.2665,
+    "train_samples": 11426,
+    "train_samples_per_second": 620.232,
+    "train_steps_per_second": 7.762
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/gated_multi_branch/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": 7.546212196350098,
+      "learning_rate": 1.766899766899767e-05,
+      "loss": 0.4088,
+      "step": 50
+    },
+    {
+      "epoch": 0.6993006993006993,
+      "grad_norm": 3.4772791862487793,
+      "learning_rate": 1.533799533799534e-05,
+      "loss": 0.2802,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9115603284902085,
+      "eval_loss": 0.2434704303741455,
+      "eval_runtime": 6.4526,
+      "eval_samples_per_second": 245.326,
+      "eval_steps_per_second": 1.55,
+      "step": 143
+    },
+    {
+      "epoch": 1.048951048951049,
+      "grad_norm": 3.2683866024017334,
+      "learning_rate": 1.3006993006993008e-05,
+      "loss": 0.2468,
+      "step": 150
+    },
+    {
+      "epoch": 1.3986013986013985,
+      "grad_norm": 5.839634418487549,
+      "learning_rate": 1.0675990675990678e-05,
+      "loss": 0.2132,
+      "step": 200
+    },
+    {
+      "epoch": 1.7482517482517483,
+      "grad_norm": 2.952626943588257,
+      "learning_rate": 8.344988344988347e-06,
+      "loss": 0.2008,
+      "step": 250
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9317751105495894,
+      "eval_loss": 0.20996716618537903,
+      "eval_runtime": 5.9832,
+      "eval_samples_per_second": 264.576,
+      "eval_steps_per_second": 1.671,
+      "step": 286
+    },
+    {
+      "epoch": 2.097902097902098,
+      "grad_norm": 5.234353542327881,
+      "learning_rate": 6.013986013986014e-06,
+      "loss": 0.1694,
+      "step": 300
+    },
+    {
+      "epoch": 2.4475524475524475,
+      "grad_norm": 4.98758602142334,
+      "learning_rate": 3.6829836829836835e-06,
+      "loss": 0.1495,
+      "step": 350
+    },
+    {
+      "epoch": 2.797202797202797,
+      "grad_norm": 4.4728617668151855,
+      "learning_rate": 1.3519813519813522e-06,
+      "loss": 0.1518,
+      "step": 400
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.932406822488945,
+      "eval_loss": 0.2144102156162262,
+      "eval_runtime": 6.0937,
+      "eval_samples_per_second": 259.775,
+      "eval_steps_per_second": 1.641,
+      "step": 429
+    },
+    {
+      "epoch": 3.0,
+      "step": 429,
+      "total_flos": 1206454205994168.0,
+      "train_loss": 0.22194261373061955,
+      "train_runtime": 55.2665,
+      "train_samples_per_second": 620.232,
+      "train_steps_per_second": 7.762
+    }
+  ],
+  "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": 1206454205994168.0,
+  "train_batch_size": 80,
+  "trial_name": null,
+  "trial_params": null
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/hf_sequence_classifier/seed_44/all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9311433986102338,
+    "eval_loss": 0.22898335754871368,
+    "eval_runtime": 6.2253,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 254.286,
+    "eval_steps_per_second": 1.606,
+    "total_flos": 1146401605629504.0,
+    "train_loss": 0.23928687622497133,
+    "train_runtime": 65.0976,
+    "train_samples": 11426,
+    "train_samples_per_second": 526.563,
+    "train_steps_per_second": 6.59
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/hf_sequence_classifier/seed_44/eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.9311433986102338,
+    "eval_loss": 0.22898335754871368,
+    "eval_runtime": 6.2253,
+    "eval_samples": 1583,
+    "eval_samples_per_second": 254.286,
+    "eval_steps_per_second": 1.606
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/hf_sequence_classifier/seed_44/train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "total_flos": 1146401605629504.0,
+    "train_loss": 0.23928687622497133,
+    "train_runtime": 65.0976,
+    "train_samples": 11426,
+    "train_samples_per_second": 526.563,
+    "train_steps_per_second": 6.59
+}

mbert_paper_metrics/result_ablation_mbert_paper/vsfc/mBERT/hf_sequence_classifier/seed_44/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": 5.094893932342529,
+      "learning_rate": 1.766899766899767e-05,
+      "loss": 0.5009,
+      "step": 50
+    },
+    {
+      "epoch": 0.6993006993006993,
+      "grad_norm": 7.145944118499756,
+      "learning_rate": 1.533799533799534e-05,
+      "loss": 0.2919,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9222994314592545,
+      "eval_loss": 0.22854222357273102,
+      "eval_runtime": 6.4974,
+      "eval_samples_per_second": 243.637,
+      "eval_steps_per_second": 1.539,
+      "step": 143
+    },
+    {
+      "epoch": 1.048951048951049,
+      "grad_norm": 3.090003252029419,
+      "learning_rate": 1.3006993006993008e-05,
+      "loss": 0.2445,
+      "step": 150
+    },
+    {
+      "epoch": 1.3986013986013985,
+      "grad_norm": 2.242663621902466,
+      "learning_rate": 1.0675990675990678e-05,
+      "loss": 0.2098,
+      "step": 200
+    },
+    {
+      "epoch": 1.7482517482517483,
+      "grad_norm": 2.601102828979492,
+      "learning_rate": 8.344988344988347e-06,
+      "loss": 0.2058,
+      "step": 250
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9298799747315224,
+      "eval_loss": 0.2253260314464569,
+      "eval_runtime": 6.6904,
+      "eval_samples_per_second": 236.607,
+      "eval_steps_per_second": 1.495,
+      "step": 286
+    },
+    {
+      "epoch": 2.097902097902098,
+      "grad_norm": 4.26569128036499,
+      "learning_rate": 6.013986013986014e-06,
+      "loss": 0.1938,
+      "step": 300
+    },
+    {
+      "epoch": 2.4475524475524475,
+      "grad_norm": 3.8603479862213135,
+      "learning_rate": 3.6829836829836835e-06,
+      "loss": 0.1563,
+      "step": 350
+    },
+    {
+      "epoch": 2.797202797202797,
+      "grad_norm": 4.265142440795898,
+      "learning_rate": 1.3519813519813522e-06,
+      "loss": 0.1558,
+      "step": 400
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9311433986102338,
+      "eval_loss": 0.22898335754871368,
+      "eval_runtime": 6.5805,
+      "eval_samples_per_second": 240.559,
+      "eval_steps_per_second": 1.52,
+      "step": 429
+    },
+    {
+      "epoch": 3.0,
+      "step": 429,
+      "total_flos": 1146401605629504.0,
+      "train_loss": 0.23928687622497133,
+      "train_runtime": 65.0976,
+      "train_samples_per_second": 526.563,
+      "train_steps_per_second": 6.59
+    }
+  ],
+  "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": 1146401605629504.0,
+  "train_batch_size": 80,
+  "trial_name": null,
+  "trial_params": null
+}