End of training

README.md

@@ -17,11 +17,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [aubmindlab/aragpt2-base](https://huggingface.co/aubmindlab/aragpt2-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0798
-- Bleu: 0.1112
-- Rouge1: 0.4634
-- Rouge2: 0.2356
-- Rougel: 0.4600
+- Loss: 0.0752
+- Bleu: 0.0928
+- Rouge1: 0.4343
+- Rouge2: 0.2043
+- Rougel: 0.4304
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 11.0,
+    "eval_bleu": 0.09284298963391195,
+    "eval_loss": 0.07523266971111298,
+    "eval_rouge1": 0.434313103224844,
+    "eval_rouge2": 0.2043067195259623,
+    "eval_rougeL": 0.430449964990971,
+    "eval_runtime": 119.8336,
+    "eval_samples": 5380,
+    "eval_samples_per_second": 44.896,
+    "eval_steps_per_second": 5.616,
+    "perplexity": 1.0781349710566281,
+    "total_flos": 1.23182992982016e+17,
+    "train_loss": 0.09744292338859069,
+    "train_runtime": 17256.5383,
+    "train_samples": 21429,
+    "train_samples_per_second": 24.836,
+    "train_steps_per_second": 3.105
+}

egy_training_log.txt

@@ -298,3 +298,5 @@ INFO:root:Epoch 10.0: Train Loss = 0.0517, Eval Loss = 0.07660207897424698
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 11.0: Train Loss = 0.0486, Eval Loss = 0.07839509844779968
 INFO:absl:Using default tokenizer.
+INFO:__main__:*** Evaluate ***
+INFO:absl:Using default tokenizer.

eval_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 11.0,
+    "eval_bleu": 0.09284298963391195,
+    "eval_loss": 0.07523266971111298,
+    "eval_rouge1": 0.434313103224844,
+    "eval_rouge2": 0.2043067195259623,
+    "eval_rougeL": 0.430449964990971,
+    "eval_runtime": 119.8336,
+    "eval_samples": 5380,
+    "eval_samples_per_second": 44.896,
+    "eval_steps_per_second": 5.616,
+    "perplexity": 1.0781349710566281
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 11.0,
+    "total_flos": 1.23182992982016e+17,
+    "train_loss": 0.09744292338859069,
+    "train_runtime": 17256.5383,
+    "train_samples": 21429,
+    "train_samples_per_second": 24.836,
+    "train_steps_per_second": 3.105
+}

trainer_state.json

@@ -0,0 +1,260 @@
+{
+  "best_metric": 0.07523266971111298,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_dj_aragpt2-base/checkpoint-16074",
+  "epoch": 11.0,
+  "eval_steps": 500,
+  "global_step": 29469,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.1490328013896942,
+      "learning_rate": 4.7947437829691034e-05,
+      "loss": 0.435,
+      "step": 2679
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.022539363481404024,
+      "eval_loss": 0.08990845084190369,
+      "eval_rouge1": 0.26967504263175696,
+      "eval_rouge2": 0.06811111838722575,
+      "eval_rougeL": 0.2648760311441921,
+      "eval_runtime": 118.8639,
+      "eval_samples_per_second": 45.262,
+      "eval_steps_per_second": 5.662,
+      "step": 2679
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.1779935508966446,
+      "learning_rate": 4.542388847023361e-05,
+      "loss": 0.0908,
+      "step": 5358
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.0494923940085718,
+      "eval_loss": 0.08216600120067596,
+      "eval_rouge1": 0.34383088385557475,
+      "eval_rouge2": 0.12325553058649232,
+      "eval_rougeL": 0.33940964680072816,
+      "eval_runtime": 181.022,
+      "eval_samples_per_second": 29.72,
+      "eval_steps_per_second": 3.718,
+      "step": 5358
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.13422226905822754,
+      "learning_rate": 4.290033911077619e-05,
+      "loss": 0.0808,
+      "step": 8037
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.06700443598211842,
+      "eval_loss": 0.0785822868347168,
+      "eval_rouge1": 0.38347464472946946,
+      "eval_rouge2": 0.15815766995910013,
+      "eval_rougeL": 0.37897410206784377,
+      "eval_runtime": 118.7728,
+      "eval_samples_per_second": 45.297,
+      "eval_steps_per_second": 5.666,
+      "step": 8037
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.17974382638931274,
+      "learning_rate": 4.0376789751318766e-05,
+      "loss": 0.0738,
+      "step": 10716
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.0782017656533744,
+      "eval_loss": 0.07652640342712402,
+      "eval_rouge1": 0.4066218577182069,
+      "eval_rouge2": 0.1797711825940213,
+      "eval_rougeL": 0.4024533186612807,
+      "eval_runtime": 302.5488,
+      "eval_samples_per_second": 17.782,
+      "eval_steps_per_second": 2.224,
+      "step": 10716
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.1600717157125473,
+      "learning_rate": 3.785324039186134e-05,
+      "loss": 0.0681,
+      "step": 13395
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.088003678878398,
+      "eval_loss": 0.07556667178869247,
+      "eval_rouge1": 0.42424169752606233,
+      "eval_rouge2": 0.19635936454770994,
+      "eval_rougeL": 0.42040996405014475,
+      "eval_runtime": 119.8742,
+      "eval_samples_per_second": 44.88,
+      "eval_steps_per_second": 5.614,
+      "step": 13395
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.17717573046684265,
+      "learning_rate": 3.532969103240392e-05,
+      "loss": 0.0632,
+      "step": 16074
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.09284298963391195,
+      "eval_loss": 0.07523266971111298,
+      "eval_rouge1": 0.434313103224844,
+      "eval_rouge2": 0.2043067195259623,
+      "eval_rougeL": 0.430449964990971,
+      "eval_runtime": 119.9224,
+      "eval_samples_per_second": 44.862,
+      "eval_steps_per_second": 5.612,
+      "step": 16074
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.20939625799655914,
+      "learning_rate": 3.28061416729465e-05,
+      "loss": 0.059,
+      "step": 18753
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.09955364884197439,
+      "eval_loss": 0.07554977387189865,
+      "eval_rouge1": 0.4438832318000667,
+      "eval_rouge2": 0.21524219506914244,
+      "eval_rougeL": 0.44011453518152965,
+      "eval_runtime": 119.8903,
+      "eval_samples_per_second": 44.874,
+      "eval_steps_per_second": 5.613,
+      "step": 18753
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.18116699159145355,
+      "learning_rate": 3.028259231348907e-05,
+      "loss": 0.0552,
+      "step": 21432
+    },
+    {
+      "epoch": 8.0,
+      "eval_bleu": 0.10150066662504155,
+      "eval_loss": 0.07605580985546112,
+      "eval_rouge1": 0.45000685735983936,
+      "eval_rouge2": 0.2217308384368585,
+      "eval_rougeL": 0.44629486754481273,
+      "eval_runtime": 180.8168,
+      "eval_samples_per_second": 29.754,
+      "eval_steps_per_second": 3.722,
+      "step": 21432
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 0.20604003965854645,
+      "learning_rate": 2.775904295403165e-05,
+      "loss": 0.0517,
+      "step": 24111
+    },
+    {
+      "epoch": 9.0,
+      "eval_bleu": 0.10504101584036675,
+      "eval_loss": 0.07660207897424698,
+      "eval_rouge1": 0.4527224851378999,
+      "eval_rouge2": 0.2249827236829457,
+      "eval_rougeL": 0.44886876248725205,
+      "eval_runtime": 241.7568,
+      "eval_samples_per_second": 22.254,
+      "eval_steps_per_second": 2.784,
+      "step": 24111
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 0.17469635605812073,
+      "learning_rate": 2.523549359457423e-05,
+      "loss": 0.0486,
+      "step": 26790
+    },
+    {
+      "epoch": 10.0,
+      "eval_bleu": 0.10926649986209491,
+      "eval_loss": 0.07839509844779968,
+      "eval_rouge1": 0.4612349211854463,
+      "eval_rouge2": 0.23378642120255436,
+      "eval_rougeL": 0.457790117200362,
+      "eval_runtime": 180.2876,
+      "eval_samples_per_second": 29.841,
+      "eval_steps_per_second": 3.733,
+      "step": 26790
+    },
+    {
+      "epoch": 11.0,
+      "grad_norm": 0.19318881630897522,
+      "learning_rate": 2.2711944235116806e-05,
+      "loss": 0.0458,
+      "step": 29469
+    },
+    {
+      "epoch": 11.0,
+      "eval_bleu": 0.11121120716108289,
+      "eval_loss": 0.07983218133449554,
+      "eval_rouge1": 0.4633689566450456,
+      "eval_rouge2": 0.2356348733383184,
+      "eval_rougeL": 0.4600017445894788,
+      "eval_runtime": 242.0184,
+      "eval_samples_per_second": 22.23,
+      "eval_steps_per_second": 2.781,
+      "step": 29469
+    },
+    {
+      "epoch": 11.0,
+      "step": 29469,
+      "total_flos": 1.23182992982016e+17,
+      "train_loss": 0.09744292338859069,
+      "train_runtime": 17256.5383,
+      "train_samples_per_second": 24.836,
+      "train_steps_per_second": 3.105
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 53580,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 5,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.23182992982016e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}