End of training

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 63.40625,
+    "eval_f1": 77.37530328335532,
+    "eval_runtime": 87.6909,
+    "eval_samples": 3858,
+    "eval_samples_per_second": 43.995,
+    "eval_steps_per_second": 1.574,
+    "total_flos": 6.354793682009856e+16,
+    "train_loss": 0.6681148458843582,
+    "train_runtime": 5891.9748,
+    "train_samples": 18247,
+    "train_samples_per_second": 15.485,
+    "train_steps_per_second": 0.553
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "eval_exact_match": 63.40625,
+    "eval_f1": 77.37530328335532,
+    "eval_runtime": 87.6909,
+    "eval_samples": 3858,
+    "eval_samples_per_second": 43.995,
+    "eval_steps_per_second": 1.574
+}

runs/May09_15-43-33_ml-machine/events.out.tfevents.1715275618.ml-machine.55741.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be112d4b86616029b03d90753f06378cce5926fb1e16101e60476e14043a3fa1
+size 412

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 6.354793682009856e+16,
+    "train_loss": 0.6681148458843582,
+    "train_runtime": 5891.9748,
+    "train_samples": 18247,
+    "train_samples_per_second": 15.485,
+    "train_steps_per_second": 0.553
+}

trainer_state.json

@@ -0,0 +1,155 @@
+{
+  "best_metric": 77.37530328335532,
+  "best_model_checkpoint": "/root/turkic_qa/en_kaz_models/en_kaz_xlm_roberta_large_squad_model/checkpoint-3260",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 3260,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "step": 652,
+      "train_exact_match": 68.63136863136863,
+      "train_f1": 83.20724970751985,
+      "train_runtime": 28.0635,
+      "train_samples_per_second": 43.722,
+      "train_steps_per_second": 1.568
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 45.333194732666016,
+      "learning_rate": 1e-05,
+      "loss": 1.2106,
+      "step": 652
+    },
+    {
+      "epoch": 1.0,
+      "eval_exact_match": 60.65625,
+      "eval_f1": 75.59481049132827,
+      "eval_runtime": 87.9399,
+      "eval_samples_per_second": 43.871,
+      "eval_steps_per_second": 1.569,
+      "step": 652
+    },
+    {
+      "epoch": 2.0,
+      "step": 1304,
+      "train_exact_match": 75.52447552447552,
+      "train_f1": 89.04467674156216,
+      "train_runtime": 28.5171,
+      "train_samples_per_second": 43.518,
+      "train_steps_per_second": 1.578
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 26.666345596313477,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.8497,
+      "step": 1304
+    },
+    {
+      "epoch": 2.0,
+      "eval_exact_match": 62.71875,
+      "eval_f1": 77.20303172730397,
+      "eval_runtime": 88.007,
+      "eval_samples_per_second": 43.837,
+      "eval_steps_per_second": 1.568,
+      "step": 1304
+    },
+    {
+      "epoch": 3.0,
+      "step": 1956,
+      "train_exact_match": 83.91608391608392,
+      "train_f1": 93.33096969615053,
+      "train_runtime": 27.7404,
+      "train_samples_per_second": 43.474,
+      "train_steps_per_second": 1.586
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 30.600830078125,
+      "learning_rate": 5e-06,
+      "loss": 0.5689,
+      "step": 1956
+    },
+    {
+      "epoch": 3.0,
+      "eval_exact_match": 63.40625,
+      "eval_f1": 77.30509811671915,
+      "eval_runtime": 88.124,
+      "eval_samples_per_second": 43.779,
+      "eval_steps_per_second": 1.566,
+      "step": 1956
+    },
+    {
+      "epoch": 4.0,
+      "step": 2608,
+      "train_exact_match": 86.61338661338661,
+      "train_f1": 95.19866566649596,
+      "train_runtime": 28.7793,
+      "train_samples_per_second": 43.782,
+      "train_steps_per_second": 1.564
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 76.71251678466797,
+      "learning_rate": 2.5e-06,
+      "loss": 0.4001,
+      "step": 2608
+    },
+    {
+      "epoch": 4.0,
+      "eval_exact_match": 62.71875,
+      "eval_f1": 77.01339638544334,
+      "eval_runtime": 88.1308,
+      "eval_samples_per_second": 43.776,
+      "eval_steps_per_second": 1.566,
+      "step": 2608
+    },
+    {
+      "epoch": 5.0,
+      "step": 3260,
+      "train_exact_match": 90.20979020979021,
+      "train_f1": 96.69185942456909,
+      "train_runtime": 27.889,
+      "train_samples_per_second": 43.386,
+      "train_steps_per_second": 1.578
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 29.443225860595703,
+      "learning_rate": 0.0,
+      "loss": 0.3113,
+      "step": 3260
+    },
+    {
+      "epoch": 5.0,
+      "eval_exact_match": 63.40625,
+      "eval_f1": 77.37530328335532,
+      "eval_runtime": 88.2621,
+      "eval_samples_per_second": 43.711,
+      "eval_steps_per_second": 1.564,
+      "step": 3260
+    },
+    {
+      "epoch": 5.0,
+      "step": 3260,
+      "total_flos": 6.354793682009856e+16,
+      "train_loss": 0.6681148458843582,
+      "train_runtime": 5891.9748,
+      "train_samples_per_second": 15.485,
+      "train_steps_per_second": 0.553
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 3260,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 6.354793682009856e+16,
+  "train_batch_size": 28,
+  "trial_name": null,
+  "trial_params": null
+}