End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 69.4375,
+    "eval_f1": 84.3335778331317,
+    "eval_runtime": 104.7966,
+    "eval_samples": 4618,
+    "eval_samples_per_second": 44.066,
+    "eval_steps_per_second": 1.574,
+    "total_flos": 7.301726548858368e+16,
+    "train_loss": 0.5262851322922751,
+    "train_runtime": 6744.6122,
+    "train_samples": 20966,
+    "train_samples_per_second": 15.543,
+    "train_steps_per_second": 0.555
+}

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:74de49b492e9dfa869c39cf1175102158482eb0396d64ec29d8dadf9948d3add
-size 23050707
+oid sha256:e170baad37cf8f7d02587ad5789bc2ee8a833b735d6399481276284c1c64820a
+size 23460193

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 69.4375,
+    "eval_f1": 84.3335778331317,
+    "eval_runtime": 104.7966,
+    "eval_samples": 4618,
+    "eval_samples_per_second": 44.066,
+    "eval_steps_per_second": 1.574
+}

runs/May11_01-29-06_ml-machine/events.out.tfevents.1715397958.ml-machine.184279.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53112f1a14c70109b117cf373e1ba5db27225df45e51a9f9f6a239b928a7a018
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 7.301726548858368e+16,
+    "train_loss": 0.5262851322922751,
+    "train_runtime": 6744.6122,
+    "train_samples": 20966,
+    "train_samples_per_second": 15.543,
+    "train_steps_per_second": 0.555
+}

trainer_state.json

@@ -0,0 +1,155 @@
+{
+  "best_metric": 84.3335778331317,
+  "best_model_checkpoint": "/root/turkic_qa/tr_kaz_models/tr_kaz_xlm_roberta_large_squad_model/checkpoint-2996",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 3745,
+  "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": 69.13086913086913,
+      "train_f1": 85.05221652384066,
+      "train_runtime": 34.8082,
+      "train_samples_per_second": 43.668,
+      "train_steps_per_second": 1.58
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 26.24435043334961,
+      "learning_rate": 1e-05,
+      "loss": 1.0066,
+      "step": 749
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 66.0625,
+      "eval_f1": 82.35823940468298,
+      "eval_runtime": 105.1616,
+      "eval_samples_per_second": 43.913,
+      "eval_steps_per_second": 1.569,
+      "step": 749
+    },
+    {
+      "epoch": 2.0,
+      "step": 1498,
+      "train_exact_match": 76.42357642357642,
+      "train_f1": 89.30189341154144,
+      "train_runtime": 33.8507,
+      "train_samples_per_second": 43.78,
+      "train_steps_per_second": 1.566
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 19.309417724609375,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.6623,
+      "step": 1498
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 67.0625,
+      "eval_f1": 83.28485925766783,
+      "eval_runtime": 105.3597,
+      "eval_samples_per_second": 43.831,
+      "eval_steps_per_second": 1.566,
+      "step": 1498
+    },
+    {
+      "epoch": 3.0,
+      "step": 2247,
+      "train_exact_match": 83.71628371628371,
+      "train_f1": 93.99788828536703,
+      "train_runtime": 34.2893,
+      "train_samples_per_second": 43.658,
+      "train_steps_per_second": 1.575
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 30.92243766784668,
+      "learning_rate": 5e-06,
+      "loss": 0.4407,
+      "step": 2247
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 69.03125,
+      "eval_f1": 84.12904259140414,
+      "eval_runtime": 105.3447,
+      "eval_samples_per_second": 43.837,
+      "eval_steps_per_second": 1.566,
+      "step": 2247
+    },
+    {
+      "epoch": 4.0,
+      "step": 2996,
+      "train_exact_match": 88.51148851148851,
+      "train_f1": 95.38995803272034,
+      "train_runtime": 34.4321,
+      "train_samples_per_second": 43.738,
+      "train_steps_per_second": 1.568
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 16.18670082092285,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3048,
+      "step": 2996
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 69.4375,
+      "eval_f1": 84.3335778331317,
+      "eval_runtime": 105.3761,
+      "eval_samples_per_second": 43.824,
+      "eval_steps_per_second": 1.566,
+      "step": 2996
+    },
+    {
+      "epoch": 5.0,
+      "step": 3745,
+      "train_exact_match": 90.10989010989012,
+      "train_f1": 96.46340834910971,
+      "train_runtime": 34.2854,
+      "train_samples_per_second": 43.634,
+      "train_steps_per_second": 1.575
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 10.52661418914795,
+      "learning_rate": 0.0,
+      "loss": 0.2169,
+      "step": 3745
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 69.78125,
+      "eval_f1": 84.15544283092075,
+      "eval_runtime": 105.3513,
+      "eval_samples_per_second": 43.834,
+      "eval_steps_per_second": 1.566,
+      "step": 3745
+    },
+    {
+      "epoch": 5.0,
+      "step": 3745,
+      "total_flos": 7.301726548858368e+16,
+      "train_loss": 0.5262851322922751,
+      "train_runtime": 6744.6122,
+      "train_samples_per_second": 15.543,
+      "train_steps_per_second": 0.555
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 3745,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 7.301726548858368e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}