End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 60.71875,
+    "eval_f1": 79.20930220037307,
+    "eval_runtime": 76.1639,
+    "eval_samples": 3359,
+    "eval_samples_per_second": 44.102,
+    "eval_steps_per_second": 1.576,
+    "total_flos": 5.432587638826752e+16,
+    "train_loss": 1.231632082265765,
+    "train_runtime": 5057.1596,
+    "train_samples": 15599,
+    "train_samples_per_second": 15.423,
+    "train_steps_per_second": 0.552
+}

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:344d0ab4fd82cb4b3e60d276961ab1fe2bcdbcc242b102744b3462fa1fa446da
-size 14158618
+oid sha256:b1607d2cc39aa4f799e7ba76f48b77d04654084edf6d1b424b42f15de07cb7f4
+size 14224444

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 60.71875,
+    "eval_f1": 79.20930220037307,
+    "eval_runtime": 76.1639,
+    "eval_samples": 3359,
+    "eval_samples_per_second": 44.102,
+    "eval_steps_per_second": 1.576
+}

runs/May10_17-02-53_ml-machine/events.out.tfevents.1715365875.ml-machine.99560.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89a0e57487154933b501e0b32ea793a80a11d19286841848f9b1f6a6c216b38d
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 5.432587638826752e+16,
+    "train_loss": 1.231632082265765,
+    "train_runtime": 5057.1596,
+    "train_samples": 15599,
+    "train_samples_per_second": 15.423,
+    "train_steps_per_second": 0.552
+}

trainer_state.json

@@ -0,0 +1,155 @@
+{
+  "best_metric": 79.20930220037307,
+  "best_model_checkpoint": "/root/turkic_qa/ru_uzn_models/ru_uzn_xlm_roberta_large_model/checkpoint-1674",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 2790,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 558,
+      "train_exact_match": 58.64135864135864,
+      "train_f1": 78.03999717537812,
+      "train_runtime": 24.1085,
+      "train_samples_per_second": 43.802,
+      "train_steps_per_second": 1.576
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 71.476318359375,
+      "learning_rate": 1e-05,
+      "loss": 3.094,
+      "step": 558
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 55.96875,
+      "eval_f1": 75.1917557548502,
+      "eval_runtime": 76.2534,
+      "eval_samples_per_second": 44.05,
+      "eval_steps_per_second": 1.574,
+      "step": 558
+    },
+    {
+      "epoch": 2.0,
+      "step": 1116,
+      "train_exact_match": 69.43056943056943,
+      "train_f1": 86.54359079483112,
+      "train_runtime": 23.8497,
+      "train_samples_per_second": 43.648,
+      "train_steps_per_second": 1.593
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 64.74529266357422,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 1.1571,
+      "step": 1116
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 60.25,
+      "eval_f1": 78.34951496921676,
+      "eval_runtime": 76.3189,
+      "eval_samples_per_second": 44.013,
+      "eval_steps_per_second": 1.572,
+      "step": 1116
+    },
+    {
+      "epoch": 3.0,
+      "step": 1674,
+      "train_exact_match": 77.32267732267732,
+      "train_f1": 90.90190658126197,
+      "train_runtime": 24.1109,
+      "train_samples_per_second": 43.632,
+      "train_steps_per_second": 1.576
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 59.769203186035156,
+      "learning_rate": 5e-06,
+      "loss": 0.8046,
+      "step": 1674
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 60.71875,
+      "eval_f1": 79.20930220037307,
+      "eval_runtime": 76.4543,
+      "eval_samples_per_second": 43.935,
+      "eval_steps_per_second": 1.57,
+      "step": 1674
+    },
+    {
+      "epoch": 4.0,
+      "step": 2232,
+      "train_exact_match": 79.12087912087912,
+      "train_f1": 91.91790747073001,
+      "train_runtime": 24.69,
+      "train_samples_per_second": 42.365,
+      "train_steps_per_second": 1.539
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 10.912280082702637,
+      "learning_rate": 2.5e-06,
+      "loss": 0.612,
+      "step": 2232
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 61.15625,
+      "eval_f1": 79.20198937913115,
+      "eval_runtime": 76.6624,
+      "eval_samples_per_second": 43.816,
+      "eval_steps_per_second": 1.565,
+      "step": 2232
+    },
+    {
+      "epoch": 5.0,
+      "step": 2790,
+      "train_exact_match": 82.71728271728271,
+      "train_f1": 92.99953055921905,
+      "train_runtime": 23.9309,
+      "train_samples_per_second": 43.542,
+      "train_steps_per_second": 1.588
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 39.581809997558594,
+      "learning_rate": 0.0,
+      "loss": 0.4905,
+      "step": 2790
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 61.03125,
+      "eval_f1": 78.98703726982994,
+      "eval_runtime": 76.5756,
+      "eval_samples_per_second": 43.865,
+      "eval_steps_per_second": 1.567,
+      "step": 2790
+    },
+    {
+      "epoch": 5.0,
+      "step": 2790,
+      "total_flos": 5.432587638826752e+16,
+      "train_loss": 1.231632082265765,
+      "train_runtime": 5057.1596,
+      "train_samples_per_second": 15.423,
+      "train_steps_per_second": 0.552
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 2790,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 5.432587638826752e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}