End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 10.0,
+    "eval_exact_match": 65.21875,
+    "eval_f1": 80.9666820924435,
+    "eval_runtime": 50.9751,
+    "eval_samples": 4618,
+    "eval_samples_per_second": 90.593,
+    "eval_steps_per_second": 3.237,
+    "total_flos": 4.108760851295232e+16,
+    "train_loss": 0.6347312815198911,
+    "train_runtime": 4825.6442,
+    "train_samples": 20966,
+    "train_samples_per_second": 43.447,
+    "train_steps_per_second": 1.552
+}

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a13827fa57814229f1a50bbc38277e962ff6fb3dcdccb67c221c1890cc6f66db
-size 23286918
+oid sha256:a019f598b37c2c4cbd659694561f27c06e8c428bffe55a0310470ca1601c34e5
+size 23878588

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "eval_exact_match": 65.21875,
+    "eval_f1": 80.9666820924435,
+    "eval_runtime": 50.9751,
+    "eval_samples": 4618,
+    "eval_samples_per_second": 90.593,
+    "eval_steps_per_second": 3.237
+}

runs/May11_00-00-23_ml-machine/events.out.tfevents.1715390614.ml-machine.159063.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:143fb1f46eae742200ab2c462b918dff55b8e72d9c64e6d513a4029968125980
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 4.108760851295232e+16,
+    "train_loss": 0.6347312815198911,
+    "train_runtime": 4825.6442,
+    "train_samples": 20966,
+    "train_samples_per_second": 43.447,
+    "train_steps_per_second": 1.552
+}

trainer_state.json

@@ -0,0 +1,280 @@
+{
+  "best_metric": 80.9666820924435,
+  "best_model_checkpoint": "/root/turkic_qa/tr_kaz_models/tr_kaz_xlm_roberta_base_squad_model/checkpoint-5243",
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 7490,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 749,
+      "train_exact_match": 58.74125874125874,
+      "train_f1": 76.91933391150722,
+      "train_runtime": 17.3698,
+      "train_samples_per_second": 87.508,
+      "train_steps_per_second": 3.166
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 24.767520904541016,
+      "learning_rate": 5e-06,
+      "loss": 1.366,
+      "step": 749
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 56.03125,
+      "eval_f1": 75.5726701661797,
+      "eval_runtime": 52.3069,
+      "eval_samples_per_second": 88.287,
+      "eval_steps_per_second": 3.154,
+      "step": 749
+    },
+    {
+      "epoch": 2.0,
+      "step": 1498,
+      "train_exact_match": 65.33466533466533,
+      "train_f1": 81.16405259356212,
+      "train_runtime": 16.8317,
+      "train_samples_per_second": 88.048,
+      "train_steps_per_second": 3.149
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 22.722991943359375,
+      "learning_rate": 1e-05,
+      "loss": 1.0296,
+      "step": 1498
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 60.6875,
+      "eval_f1": 77.7357851907795,
+      "eval_runtime": 52.4104,
+      "eval_samples_per_second": 88.112,
+      "eval_steps_per_second": 3.148,
+      "step": 1498
+    },
+    {
+      "epoch": 3.0,
+      "step": 2247,
+      "train_exact_match": 74.52547452547452,
+      "train_f1": 87.61463327451462,
+      "train_runtime": 16.7088,
+      "train_samples_per_second": 89.593,
+      "train_steps_per_second": 3.232
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 18.895648956298828,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 0.8407,
+      "step": 2247
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 63.0625,
+      "eval_f1": 79.25734099418638,
+      "eval_runtime": 51.9473,
+      "eval_samples_per_second": 88.898,
+      "eval_steps_per_second": 3.176,
+      "step": 2247
+    },
+    {
+      "epoch": 4.0,
+      "step": 2996,
+      "train_exact_match": 75.02497502497502,
+      "train_f1": 88.1924695005796,
+      "train_runtime": 16.7212,
+      "train_samples_per_second": 90.065,
+      "train_steps_per_second": 3.229
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 21.92975425720215,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.6761,
+      "step": 2996
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 64.3125,
+      "eval_f1": 80.41214611674394,
+      "eval_runtime": 51.1453,
+      "eval_samples_per_second": 90.292,
+      "eval_steps_per_second": 3.226,
+      "step": 2996
+    },
+    {
+      "epoch": 5.0,
+      "step": 3745,
+      "train_exact_match": 80.71928071928072,
+      "train_f1": 91.80017345637452,
+      "train_runtime": 16.6722,
+      "train_samples_per_second": 89.73,
+      "train_steps_per_second": 3.239
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 10.589529991149902,
+      "learning_rate": 6.25e-06,
+      "loss": 0.5642,
+      "step": 3745
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 65.0625,
+      "eval_f1": 80.50496259716563,
+      "eval_runtime": 51.2576,
+      "eval_samples_per_second": 90.094,
+      "eval_steps_per_second": 3.219,
+      "step": 3745
+    },
+    {
+      "epoch": 6.0,
+      "step": 4494,
+      "train_exact_match": 82.71728271728271,
+      "train_f1": 92.8088019961768,
+      "train_runtime": 16.7647,
+      "train_samples_per_second": 88.997,
+      "train_steps_per_second": 3.221
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 17.399965286254883,
+      "learning_rate": 5e-06,
+      "loss": 0.4769,
+      "step": 4494
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 65.1875,
+      "eval_f1": 80.88709112615079,
+      "eval_runtime": 51.1815,
+      "eval_samples_per_second": 90.228,
+      "eval_steps_per_second": 3.224,
+      "step": 4494
+    },
+    {
+      "epoch": 7.0,
+      "step": 5243,
+      "train_exact_match": 84.41558441558442,
+      "train_f1": 93.70818045593467,
+      "train_runtime": 16.3897,
+      "train_samples_per_second": 90.118,
+      "train_steps_per_second": 3.234
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 15.382094383239746,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.4089,
+      "step": 5243
+    },
+    {
+      "epoch": 7.0,
+      "eval_exact_match": 65.21875,
+      "eval_f1": 80.9666820924435,
+      "eval_runtime": 51.2038,
+      "eval_samples_per_second": 90.189,
+      "eval_steps_per_second": 3.222,
+      "step": 5243
+    },
+    {
+      "epoch": 8.0,
+      "step": 5992,
+      "train_exact_match": 84.41558441558442,
+      "train_f1": 94.00170296456992,
+      "train_runtime": 17.5397,
+      "train_samples_per_second": 87.573,
+      "train_steps_per_second": 3.136
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 48.553558349609375,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3646,
+      "step": 5992
+    },
+    {
+      "epoch": 8.0,
+      "eval_exact_match": 64.71875,
+      "eval_f1": 80.35415857951902,
+      "eval_runtime": 52.7202,
+      "eval_samples_per_second": 87.594,
+      "eval_steps_per_second": 3.13,
+      "step": 5992
+    },
+    {
+      "epoch": 9.0,
+      "step": 6741,
+      "train_exact_match": 87.61238761238761,
+      "train_f1": 95.10743821560438,
+      "train_runtime": 16.7145,
+      "train_samples_per_second": 88.008,
+      "train_steps_per_second": 3.171
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 54.76875305175781,
+      "learning_rate": 1.25e-06,
+      "loss": 0.3226,
+      "step": 6741
+    },
+    {
+      "epoch": 9.0,
+      "eval_exact_match": 65.40625,
+      "eval_f1": 80.62880989809192,
+      "eval_runtime": 52.4199,
+      "eval_samples_per_second": 88.096,
+      "eval_steps_per_second": 3.148,
+      "step": 6741
+    },
+    {
+      "epoch": 10.0,
+      "step": 7490,
+      "train_exact_match": 88.1118881118881,
+      "train_f1": 95.35905521584445,
+      "train_runtime": 16.8298,
+      "train_samples_per_second": 88.771,
+      "train_steps_per_second": 3.209
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 20.260765075683594,
+      "learning_rate": 0.0,
+      "loss": 0.2978,
+      "step": 7490
+    },
+    {
+      "epoch": 10.0,
+      "eval_exact_match": 65.59375,
+      "eval_f1": 80.93066187820529,
+      "eval_runtime": 51.2112,
+      "eval_samples_per_second": 90.176,
+      "eval_steps_per_second": 3.222,
+      "step": 7490
+    },
+    {
+      "epoch": 10.0,
+      "step": 7490,
+      "total_flos": 4.108760851295232e+16,
+      "train_loss": 0.6347312815198911,
+      "train_runtime": 4825.6442,
+      "train_samples_per_second": 43.447,
+      "train_steps_per_second": 1.552
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 7490,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "total_flos": 4.108760851295232e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}