End of training

README.md

@@ -18,11 +18,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [aubmindlab/aragpt2-large](https://huggingface.co/aubmindlab/aragpt2-large) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0578
-- Bleu: 0.1113
-- Rouge1: 0.4976
-- Rouge2: 0.2714
-- Rougel: 0.4940
+- Loss: 0.0469
+- Bleu: 0.0733
+- Rouge1: 0.3956
+- Rouge2: 0.1677
+- Rougel: 0.3901
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 7.0,
+    "eval_bleu": 0.07334464750125981,
+    "eval_loss": 0.04691193997859955,
+    "eval_rouge1": 0.39556592669658314,
+    "eval_rouge2": 0.1676736409423039,
+    "eval_rougeL": 0.3901085723006601,
+    "eval_runtime": 41.1462,
+    "eval_samples": 304,
+    "eval_samples_per_second": 7.388,
+    "eval_steps_per_second": 0.924,
+    "perplexity": 1.048029715496853,
+    "total_flos": 3.71386078199808e+16,
+    "train_loss": 0.1430642572443156,
+    "train_runtime": 4921.8294,
+    "train_samples": 1219,
+    "train_samples_per_second": 4.953,
+    "train_steps_per_second": 1.239
+}

egy_training_log.txt

@@ -568,3 +568,5 @@ INFO:root:Epoch 6.0: Train Loss = 0.016, Eval Loss = 0.054129794239997864
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 7.0: Train Loss = 0.0134, Eval Loss = 0.05314570292830467
 INFO:absl:Using default tokenizer.
+INFO:__main__:*** Evaluate ***
+INFO:absl:Using default tokenizer.

eval_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 7.0,
+    "eval_bleu": 0.07334464750125981,
+    "eval_loss": 0.04691193997859955,
+    "eval_rouge1": 0.39556592669658314,
+    "eval_rouge2": 0.1676736409423039,
+    "eval_rougeL": 0.3901085723006601,
+    "eval_runtime": 41.1462,
+    "eval_samples": 304,
+    "eval_samples_per_second": 7.388,
+    "eval_steps_per_second": 0.924,
+    "perplexity": 1.048029715496853
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 7.0,
+    "total_flos": 3.71386078199808e+16,
+    "train_loss": 0.1430642572443156,
+    "train_runtime": 4921.8294,
+    "train_samples": 1219,
+    "train_samples_per_second": 4.953,
+    "train_steps_per_second": 1.239
+}

trainer_state.json

@@ -0,0 +1,184 @@
+{
+  "best_metric": 0.04691193997859955,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_yem_aragpt2-large/checkpoint-610",
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 2135,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.40213119983673096,
+      "learning_rate": 3.05e-05,
+      "loss": 0.8644,
+      "step": 305
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.03103545034655718,
+      "eval_loss": 0.049881722778081894,
+      "eval_rouge1": 0.2976310465047741,
+      "eval_rouge2": 0.0825114524566873,
+      "eval_rougeL": 0.2931870351050212,
+      "eval_runtime": 101.6572,
+      "eval_samples_per_second": 2.99,
+      "eval_steps_per_second": 0.374,
+      "step": 305
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.45833268761634827,
+      "learning_rate": 4.901785714285714e-05,
+      "loss": 0.0439,
+      "step": 610
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.07334464750125981,
+      "eval_loss": 0.04691193997859955,
+      "eval_rouge1": 0.39556592669658314,
+      "eval_rouge2": 0.1676736409423039,
+      "eval_rougeL": 0.3901085723006601,
+      "eval_runtime": 41.076,
+      "eval_samples_per_second": 7.401,
+      "eval_steps_per_second": 0.925,
+      "step": 610
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.3850250840187073,
+      "learning_rate": 4.629464285714286e-05,
+      "loss": 0.0307,
+      "step": 915
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.09008751782596823,
+      "eval_loss": 0.0473761111497879,
+      "eval_rouge1": 0.44112670621442845,
+      "eval_rouge2": 0.20925026290023724,
+      "eval_rougeL": 0.43611440413552455,
+      "eval_runtime": 41.095,
+      "eval_samples_per_second": 7.397,
+      "eval_steps_per_second": 0.925,
+      "step": 915
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.4408362805843353,
+      "learning_rate": 4.3571428571428576e-05,
+      "loss": 0.0212,
+      "step": 1220
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.10385450068202749,
+      "eval_loss": 0.04973715916275978,
+      "eval_rouge1": 0.4643434470122675,
+      "eval_rouge2": 0.23150126678662394,
+      "eval_rougeL": 0.45914976519742434,
+      "eval_runtime": 41.0763,
+      "eval_samples_per_second": 7.401,
+      "eval_steps_per_second": 0.925,
+      "step": 1220
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.4677392840385437,
+      "learning_rate": 4.084821428571429e-05,
+      "loss": 0.016,
+      "step": 1525
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.09227592943056641,
+      "eval_loss": 0.054129794239997864,
+      "eval_rouge1": 0.464136474952609,
+      "eval_rouge2": 0.22291426741508574,
+      "eval_rougeL": 0.4599627344232192,
+      "eval_runtime": 41.1781,
+      "eval_samples_per_second": 7.383,
+      "eval_steps_per_second": 0.923,
+      "step": 1525
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.30016008019447327,
+      "learning_rate": 3.8125e-05,
+      "loss": 0.0134,
+      "step": 1830
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.11604582902883592,
+      "eval_loss": 0.05314570292830467,
+      "eval_rouge1": 0.47459996105255675,
+      "eval_rouge2": 0.25028361013073314,
+      "eval_rougeL": 0.46988615577124576,
+      "eval_runtime": 61.9343,
+      "eval_samples_per_second": 4.908,
+      "eval_steps_per_second": 0.614,
+      "step": 1830
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.24085062742233276,
+      "learning_rate": 3.5401785714285716e-05,
+      "loss": 0.0118,
+      "step": 2135
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.11125669152362899,
+      "eval_loss": 0.057827215641736984,
+      "eval_rouge1": 0.4975937747836845,
+      "eval_rouge2": 0.27138037997797015,
+      "eval_rougeL": 0.4939744686596661,
+      "eval_runtime": 101.6221,
+      "eval_samples_per_second": 2.991,
+      "eval_steps_per_second": 0.374,
+      "step": 2135
+    },
+    {
+      "epoch": 7.0,
+      "step": 2135,
+      "total_flos": 3.71386078199808e+16,
+      "train_loss": 0.1430642572443156,
+      "train_runtime": 4921.8294,
+      "train_samples_per_second": 4.953,
+      "train_steps_per_second": 1.239
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 6100,
+  "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": 3.71386078199808e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}