End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 10.0,
+    "eval_exact_match": 65.03125,
+    "eval_f1": 78.54224063649897,
+    "eval_runtime": 59.4914,
+    "eval_samples": 4372,
+    "eval_samples_per_second": 73.49,
+    "eval_steps_per_second": 2.639,
+    "total_flos": 4.073559349625856e+16,
+    "train_loss": 0.5783894863616409,
+    "train_runtime": 6220.8095,
+    "train_samples": 20786,
+    "train_samples_per_second": 33.414,
+    "train_steps_per_second": 1.194
+}

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:98b3578cafa7f98f010171616d06ec1e9f636e2295d2205b705fc1ce51a210ab
-size 14022721
+oid sha256:ba0d27b6f33a27ae32f51accc2a21a5cfc98ec439f6787cd21860b68f52538e2
+size 14169531

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "eval_exact_match": 65.03125,
+    "eval_f1": 78.54224063649897,
+    "eval_runtime": 59.4914,
+    "eval_samples": 4372,
+    "eval_samples_per_second": 73.49,
+    "eval_steps_per_second": 2.639
+}

runs/May15_19-34-00_ml-machine/events.out.tfevents.1715808044.ml-machine.1192.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61a5a91804da52c0b11fb9f91213bb10fd3b259540e84e063578e9f68d336e89
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 4.073559349625856e+16,
+    "train_loss": 0.5783894863616409,
+    "train_runtime": 6220.8095,
+    "train_samples": 20786,
+    "train_samples_per_second": 33.414,
+    "train_steps_per_second": 1.194
+}

trainer_state.json

@@ -0,0 +1,280 @@
+{
+  "best_metric": 78.54224063649897,
+  "best_model_checkpoint": "/root/turkic_qa/en_uzn_models/en_uzn_mdeberta_base_squad_model/checkpoint-2229",
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 7430,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 743,
+      "train_exact_match": 60.739260739260736,
+      "train_f1": 76.79740346661637,
+      "train_runtime": 18.4418,
+      "train_samples_per_second": 73.529,
+      "train_steps_per_second": 2.657
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 12.275191307067871,
+      "learning_rate": 5e-06,
+      "loss": 1.1938,
+      "step": 743
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 59.40625,
+      "eval_f1": 74.4654162201273,
+      "eval_runtime": 58.6195,
+      "eval_samples_per_second": 74.583,
+      "eval_steps_per_second": 2.678,
+      "step": 743
+    },
+    {
+      "epoch": 2.0,
+      "step": 1486,
+      "train_exact_match": 71.32867132867133,
+      "train_f1": 82.88126787145742,
+      "train_runtime": 18.7819,
+      "train_samples_per_second": 73.901,
+      "train_steps_per_second": 2.662
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 17.339393615722656,
+      "learning_rate": 1e-05,
+      "loss": 0.9538,
+      "step": 1486
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 63.21875,
+      "eval_f1": 77.06909842105833,
+      "eval_runtime": 59.7376,
+      "eval_samples_per_second": 73.187,
+      "eval_steps_per_second": 2.628,
+      "step": 1486
+    },
+    {
+      "epoch": 3.0,
+      "step": 2229,
+      "train_exact_match": 72.02797202797203,
+      "train_f1": 85.06033245209376,
+      "train_runtime": 19.1902,
+      "train_samples_per_second": 72.954,
+      "train_steps_per_second": 2.605
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 31.182722091674805,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 0.7754,
+      "step": 2229
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 65.03125,
+      "eval_f1": 78.54224063649897,
+      "eval_runtime": 60.2459,
+      "eval_samples_per_second": 72.569,
+      "eval_steps_per_second": 2.606,
+      "step": 2229
+    },
+    {
+      "epoch": 4.0,
+      "step": 2972,
+      "train_exact_match": 77.42257742257742,
+      "train_f1": 89.5672832756712,
+      "train_runtime": 18.9023,
+      "train_samples_per_second": 73.483,
+      "train_steps_per_second": 2.645
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 31.596288681030273,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.6219,
+      "step": 2972
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 65.0,
+      "eval_f1": 78.32854487938292,
+      "eval_runtime": 59.9399,
+      "eval_samples_per_second": 72.94,
+      "eval_steps_per_second": 2.619,
+      "step": 2972
+    },
+    {
+      "epoch": 5.0,
+      "step": 3715,
+      "train_exact_match": 78.92107892107892,
+      "train_f1": 90.10872697449445,
+      "train_runtime": 18.4815,
+      "train_samples_per_second": 71.964,
+      "train_steps_per_second": 2.597
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 18.16082000732422,
+      "learning_rate": 6.25e-06,
+      "loss": 0.5195,
+      "step": 3715
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 64.40625,
+      "eval_f1": 78.09634918973954,
+      "eval_runtime": 59.7181,
+      "eval_samples_per_second": 73.211,
+      "eval_steps_per_second": 2.629,
+      "step": 3715
+    },
+    {
+      "epoch": 6.0,
+      "step": 4458,
+      "train_exact_match": 82.61738261738262,
+      "train_f1": 91.68526538584675,
+      "train_runtime": 18.7765,
+      "train_samples_per_second": 72.005,
+      "train_steps_per_second": 2.61
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 23.594913482666016,
+      "learning_rate": 5e-06,
+      "loss": 0.4385,
+      "step": 4458
+    },
+    {
+      "epoch": 6.0,
+      "eval_exact_match": 64.59375,
+      "eval_f1": 77.90198302625599,
+      "eval_runtime": 60.4426,
+      "eval_samples_per_second": 72.333,
+      "eval_steps_per_second": 2.598,
+      "step": 4458
+    },
+    {
+      "epoch": 7.0,
+      "step": 5201,
+      "train_exact_match": 84.81518481518482,
+      "train_f1": 93.31685648855274,
+      "train_runtime": 18.5303,
+      "train_samples_per_second": 72.206,
+      "train_steps_per_second": 2.59
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 28.661422729492188,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.3756,
+      "step": 5201
+    },
+    {
+      "epoch": 7.0,
+      "eval_exact_match": 64.34375,
+      "eval_f1": 77.75398362289235,
+      "eval_runtime": 60.0044,
+      "eval_samples_per_second": 72.861,
+      "eval_steps_per_second": 2.616,
+      "step": 5201
+    },
+    {
+      "epoch": 8.0,
+      "step": 5944,
+      "train_exact_match": 84.01598401598402,
+      "train_f1": 92.96191804754487,
+      "train_runtime": 19.1375,
+      "train_samples_per_second": 71.953,
+      "train_steps_per_second": 2.613
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 36.12919998168945,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3313,
+      "step": 5944
+    },
+    {
+      "epoch": 8.0,
+      "eval_exact_match": 64.40625,
+      "eval_f1": 77.95679362066461,
+      "eval_runtime": 59.8689,
+      "eval_samples_per_second": 73.026,
+      "eval_steps_per_second": 2.622,
+      "step": 5944
+    },
+    {
+      "epoch": 9.0,
+      "step": 6687,
+      "train_exact_match": 87.01298701298701,
+      "train_f1": 94.75304040815107,
+      "train_runtime": 19.6099,
+      "train_samples_per_second": 71.443,
+      "train_steps_per_second": 2.601
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 7.826128959655762,
+      "learning_rate": 1.25e-06,
+      "loss": 0.2966,
+      "step": 6687
+    },
+    {
+      "epoch": 9.0,
+      "eval_exact_match": 64.40625,
+      "eval_f1": 77.80886757487892,
+      "eval_runtime": 59.9613,
+      "eval_samples_per_second": 72.914,
+      "eval_steps_per_second": 2.618,
+      "step": 6687
+    },
+    {
+      "epoch": 10.0,
+      "step": 7430,
+      "train_exact_match": 87.81218781218782,
+      "train_f1": 95.03010006914994,
+      "train_runtime": 19.1654,
+      "train_samples_per_second": 71.274,
+      "train_steps_per_second": 2.557
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 16.050716400146484,
+      "learning_rate": 0.0,
+      "loss": 0.2775,
+      "step": 7430
+    },
+    {
+      "epoch": 10.0,
+      "eval_exact_match": 64.21875,
+      "eval_f1": 77.80834688072528,
+      "eval_runtime": 60.7473,
+      "eval_samples_per_second": 71.97,
+      "eval_steps_per_second": 2.584,
+      "step": 7430
+    },
+    {
+      "epoch": 10.0,
+      "step": 7430,
+      "total_flos": 4.073559349625856e+16,
+      "train_loss": 0.5783894863616409,
+      "train_runtime": 6220.8095,
+      "train_samples_per_second": 33.414,
+      "train_steps_per_second": 1.194
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 7430,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "total_flos": 4.073559349625856e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}