End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 63.96875,
+    "eval_f1": 80.76073976371114,
+    "eval_runtime": 74.826,
+    "eval_samples": 3281,
+    "eval_samples_per_second": 43.848,
+    "eval_steps_per_second": 1.577,
+    "total_flos": 5.376168689054976e+16,
+    "train_loss": 1.1795940786168195,
+    "train_runtime": 4973.0116,
+    "train_samples": 15437,
+    "train_samples_per_second": 15.521,
+    "train_steps_per_second": 0.555
+}

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:121c180ce77f71f6af553682aca8c835d714085c204bd75515f8b90fb476669d
-size 23382387
+oid sha256:c42a6d1bb817379c460711d83de38aca90d92824800622ffc8e9d8998f987859
+size 23993364

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 63.96875,
+    "eval_f1": 80.76073976371114,
+    "eval_runtime": 74.826,
+    "eval_samples": 3281,
+    "eval_samples_per_second": 43.848,
+    "eval_steps_per_second": 1.577
+}

runs/May10_10-57-05_ml-machine/events.out.tfevents.1715343825.ml-machine.43133.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55e00d37a4b3abc08071578cfd45d241ae58f50e3b1a5beeb106dc3ca70a6875
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 5.376168689054976e+16,
+    "train_loss": 1.1795940786168195,
+    "train_runtime": 4973.0116,
+    "train_samples": 15437,
+    "train_samples_per_second": 15.521,
+    "train_steps_per_second": 0.555
+}

trainer_state.json

@@ -0,0 +1,155 @@
+{
+  "best_metric": 80.76073976371114,
+  "best_model_checkpoint": "/root/turkic_qa/ru_kaz_models/ru_kaz_xlm_roberta_large_model/checkpoint-2208",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 2760,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 552,
+      "train_exact_match": 58.94105894105894,
+      "train_f1": 78.59308014289637,
+      "train_runtime": 23.2375,
+      "train_samples_per_second": 43.722,
+      "train_steps_per_second": 1.592
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 83.0545425415039,
+      "learning_rate": 1e-05,
+      "loss": 2.9982,
+      "step": 552
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 57.84375,
+      "eval_f1": 76.41032874262423,
+      "eval_runtime": 74.7526,
+      "eval_samples_per_second": 43.891,
+      "eval_steps_per_second": 1.579,
+      "step": 552
+    },
+    {
+      "epoch": 2.0,
+      "step": 1104,
+      "train_exact_match": 73.52647352647352,
+      "train_f1": 87.35432090211079,
+      "train_runtime": 23.6437,
+      "train_samples_per_second": 43.479,
+      "train_steps_per_second": 1.565
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 65.99756622314453,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 1.0802,
+      "step": 1104
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 62.90625,
+      "eval_f1": 79.63501935340823,
+      "eval_runtime": 75.6962,
+      "eval_samples_per_second": 43.344,
+      "eval_steps_per_second": 1.559,
+      "step": 1104
+    },
+    {
+      "epoch": 3.0,
+      "step": 1656,
+      "train_exact_match": 77.22277722277722,
+      "train_f1": 90.8541682004209,
+      "train_runtime": 23.3536,
+      "train_samples_per_second": 43.548,
+      "train_steps_per_second": 1.584
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 94.93318939208984,
+      "learning_rate": 5e-06,
+      "loss": 0.7755,
+      "step": 1656
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 63.40625,
+      "eval_f1": 79.91103921274045,
+      "eval_runtime": 75.0081,
+      "eval_samples_per_second": 43.742,
+      "eval_steps_per_second": 1.573,
+      "step": 1656
+    },
+    {
+      "epoch": 4.0,
+      "step": 2208,
+      "train_exact_match": 82.81718281718281,
+      "train_f1": 93.85897538338976,
+      "train_runtime": 23.3033,
+      "train_samples_per_second": 43.513,
+      "train_steps_per_second": 1.588
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 181.9861297607422,
+      "learning_rate": 2.5e-06,
+      "loss": 0.584,
+      "step": 2208
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 63.96875,
+      "eval_f1": 80.76073976371114,
+      "eval_runtime": 74.9415,
+      "eval_samples_per_second": 43.781,
+      "eval_steps_per_second": 1.575,
+      "step": 2208
+    },
+    {
+      "epoch": 5.0,
+      "step": 2760,
+      "train_exact_match": 85.81418581418582,
+      "train_f1": 95.18484768485531,
+      "train_runtime": 23.5964,
+      "train_samples_per_second": 43.312,
+      "train_steps_per_second": 1.568
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 45.12015914916992,
+      "learning_rate": 0.0,
+      "loss": 0.4601,
+      "step": 2760
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 64.4375,
+      "eval_f1": 80.75596391907597,
+      "eval_runtime": 75.7078,
+      "eval_samples_per_second": 43.338,
+      "eval_steps_per_second": 1.559,
+      "step": 2760
+    },
+    {
+      "epoch": 5.0,
+      "step": 2760,
+      "total_flos": 5.376168689054976e+16,
+      "train_loss": 1.1795940786168195,
+      "train_runtime": 4973.0116,
+      "train_samples_per_second": 15.521,
+      "train_steps_per_second": 0.555
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 2760,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 5.376168689054976e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}