End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 69.71875,
+    "eval_f1": 84.25954102103677,
+    "eval_runtime": 118.2106,
+    "eval_samples": 5111,
+    "eval_samples_per_second": 43.236,
+    "eval_steps_per_second": 1.548,
+    "total_flos": 7.910493982198272e+16,
+    "train_loss": 0.5004441698196487,
+    "train_runtime": 7390.7618,
+    "train_samples": 22714,
+    "train_samples_per_second": 15.366,
+    "train_steps_per_second": 0.549
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 69.71875,
+    "eval_f1": 84.25954102103677,
+    "eval_runtime": 118.2106,
+    "eval_samples": 5111,
+    "eval_samples_per_second": 43.236,
+    "eval_steps_per_second": 1.548
+}

runs/May11_11-46-01_ml-machine/events.out.tfevents.1715435640.ml-machine.286461.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1980451bc755e8d6d416112726773759008b54009c20c496e5899bbd257c8934
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 7.910493982198272e+16,
+    "train_loss": 0.5004441698196487,
+    "train_runtime": 7390.7618,
+    "train_samples": 22714,
+    "train_samples_per_second": 15.366,
+    "train_steps_per_second": 0.549
+}

trainer_state.json

@@ -0,0 +1,155 @@
+{
+  "best_metric": 84.25954102103677,
+  "best_model_checkpoint": "/root/turkic_qa/tr_uzn_models/tr_uzn_xlm_roberta_large_squad_model/checkpoint-4060",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 4060,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 812,
+      "train_exact_match": 67.33266733266733,
+      "train_f1": 84.16624209950214,
+      "train_runtime": 38.4449,
+      "train_samples_per_second": 43.179,
+      "train_steps_per_second": 1.561
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 38.27867126464844,
+      "learning_rate": 1e-05,
+      "loss": 0.96,
+      "step": 812
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 65.96875,
+      "eval_f1": 82.56564918433806,
+      "eval_runtime": 118.2731,
+      "eval_samples_per_second": 43.214,
+      "eval_steps_per_second": 1.547,
+      "step": 812
+    },
+    {
+      "epoch": 2.0,
+      "step": 1624,
+      "train_exact_match": 79.32067932067932,
+      "train_f1": 90.47712960156771,
+      "train_runtime": 37.8398,
+      "train_samples_per_second": 42.997,
+      "train_steps_per_second": 1.559
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 50.23493576049805,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.6242,
+      "step": 1624
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 68.4375,
+      "eval_f1": 84.06100133020266,
+      "eval_runtime": 118.0295,
+      "eval_samples_per_second": 43.303,
+      "eval_steps_per_second": 1.55,
+      "step": 1624
+    },
+    {
+      "epoch": 3.0,
+      "step": 2436,
+      "train_exact_match": 81.31868131868131,
+      "train_f1": 91.7552771568634,
+      "train_runtime": 36.6816,
+      "train_samples_per_second": 43.046,
+      "train_steps_per_second": 1.554
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 10.1982421875,
+      "learning_rate": 5e-06,
+      "loss": 0.4159,
+      "step": 2436
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 67.96875,
+      "eval_f1": 83.74509639678304,
+      "eval_runtime": 118.288,
+      "eval_samples_per_second": 43.208,
+      "eval_steps_per_second": 1.547,
+      "step": 2436
+    },
+    {
+      "epoch": 4.0,
+      "step": 3248,
+      "train_exact_match": 86.81318681318682,
+      "train_f1": 94.84718111795424,
+      "train_runtime": 36.8975,
+      "train_samples_per_second": 43.174,
+      "train_steps_per_second": 1.545
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 14.403440475463867,
+      "learning_rate": 2.5e-06,
+      "loss": 0.2868,
+      "step": 3248
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 69.21875,
+      "eval_f1": 83.92338168335839,
+      "eval_runtime": 118.3768,
+      "eval_samples_per_second": 43.176,
+      "eval_steps_per_second": 1.546,
+      "step": 3248
+    },
+    {
+      "epoch": 5.0,
+      "step": 4060,
+      "train_exact_match": 90.40959040959041,
+      "train_f1": 95.51651922991793,
+      "train_runtime": 36.9341,
+      "train_samples_per_second": 43.212,
+      "train_steps_per_second": 1.543
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 36.00727081298828,
+      "learning_rate": 0.0,
+      "loss": 0.2153,
+      "step": 4060
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 69.71875,
+      "eval_f1": 84.25954102103677,
+      "eval_runtime": 118.2561,
+      "eval_samples_per_second": 43.22,
+      "eval_steps_per_second": 1.547,
+      "step": 4060
+    },
+    {
+      "epoch": 5.0,
+      "step": 4060,
+      "total_flos": 7.910493982198272e+16,
+      "train_loss": 0.5004441698196487,
+      "train_runtime": 7390.7618,
+      "train_samples_per_second": 15.366,
+      "train_steps_per_second": 0.549
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 4060,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 7.910493982198272e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}