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 [riotu-lab/ArabianGPT-03B](https://huggingface.co/riotu-lab/ArabianGPT-03B) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.9511
-- Bleu: 0.3620
-- Rouge1: 0.6421
-- Rouge2: 0.3875
-- Rougel: 0.6388
+- Loss: 0.7603
+- Bleu: 0.3632
+- Rouge1: 0.6403
+- Rouge2: 0.3829
+- Rougel: 0.6368
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 7.0,
+    "eval_bleu": 0.3631689720024005,
+    "eval_loss": 0.7602559328079224,
+    "eval_rouge1": 0.6402930839888462,
+    "eval_rouge2": 0.3828942418456889,
+    "eval_rougeL": 0.6367647874361373,
+    "eval_runtime": 105.7511,
+    "eval_samples": 14209,
+    "eval_samples_per_second": 134.363,
+    "eval_steps_per_second": 16.804,
+    "perplexity": 2.1388235455702778,
+    "total_flos": 9.237135717983846e+16,
+    "train_loss": 0.40426278318054437,
+    "train_runtime": 12413.183,
+    "train_samples": 56836,
+    "train_samples_per_second": 91.574,
+    "train_steps_per_second": 11.448
+}

egy_training_log.txt

@@ -154,3 +154,5 @@ INFO:root:Epoch 6.0: Train Loss = 0.2518, Eval Loss = 0.8583759069442749
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 7.0: Train Loss = 0.1991, Eval Loss = 0.9106101989746094
 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": 0.3631689720024005,
+    "eval_loss": 0.7602559328079224,
+    "eval_rouge1": 0.6402930839888462,
+    "eval_rouge2": 0.3828942418456889,
+    "eval_rougeL": 0.6367647874361373,
+    "eval_runtime": 105.7511,
+    "eval_samples": 14209,
+    "eval_samples_per_second": 134.363,
+    "eval_steps_per_second": 16.804,
+    "perplexity": 2.1388235455702778
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 7.0,
+    "total_flos": 9.237135717983846e+16,
+    "train_loss": 0.40426278318054437,
+    "train_runtime": 12413.183,
+    "train_samples": 56836,
+    "train_samples_per_second": 91.574,
+    "train_steps_per_second": 11.448
+}

trainer_state.json

@@ -0,0 +1,184 @@
+{
+  "best_metric": 0.7602559328079224,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_eg_03/checkpoint-14210",
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 49735,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.4059462547302246,
+      "learning_rate": 4.766772598870057e-05,
+      "loss": 0.874,
+      "step": 7105
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.34333508777542915,
+      "eval_loss": 0.7882477641105652,
+      "eval_rouge1": 0.6159142183825631,
+      "eval_rouge2": 0.3508423646141942,
+      "eval_rougeL": 0.6121110717073233,
+      "eval_runtime": 132.93,
+      "eval_samples_per_second": 106.891,
+      "eval_steps_per_second": 13.368,
+      "step": 7105
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 3.7706522941589355,
+      "learning_rate": 4.515889830508475e-05,
+      "loss": 0.5689,
+      "step": 14210
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.3631689720024005,
+      "eval_loss": 0.7602559328079224,
+      "eval_rouge1": 0.6402930839888462,
+      "eval_rouge2": 0.3828942418456889,
+      "eval_rougeL": 0.6367647874361373,
+      "eval_runtime": 219.0505,
+      "eval_samples_per_second": 64.866,
+      "eval_steps_per_second": 8.112,
+      "step": 14210
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 2.4260447025299072,
+      "learning_rate": 4.265007062146893e-05,
+      "loss": 0.4369,
+      "step": 21315
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.36610303407947103,
+      "eval_loss": 0.7770611643791199,
+      "eval_rouge1": 0.6205180467810911,
+      "eval_rouge2": 0.28726786502140805,
+      "eval_rougeL": 0.6174286813730623,
+      "eval_runtime": 229.3738,
+      "eval_samples_per_second": 61.947,
+      "eval_steps_per_second": 7.747,
+      "step": 21315
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 4.454061031341553,
+      "learning_rate": 4.014124293785311e-05,
+      "loss": 0.3314,
+      "step": 28420
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.365696341752005,
+      "eval_loss": 0.8175964951515198,
+      "eval_rouge1": 0.6468122957416143,
+      "eval_rouge2": 0.3918607021930349,
+      "eval_rougeL": 0.6436111807449674,
+      "eval_runtime": 229.5565,
+      "eval_samples_per_second": 61.898,
+      "eval_steps_per_second": 7.741,
+      "step": 28420
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.0373406410217285,
+      "learning_rate": 3.763241525423729e-05,
+      "loss": 0.2518,
+      "step": 35525
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.3619747558826484,
+      "eval_loss": 0.8583759069442749,
+      "eval_rouge1": 0.6465870583022197,
+      "eval_rouge2": 0.3915693838544316,
+      "eval_rougeL": 0.6434157590310816,
+      "eval_runtime": 139.7631,
+      "eval_samples_per_second": 101.665,
+      "eval_steps_per_second": 12.714,
+      "step": 35525
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 2.6252825260162354,
+      "learning_rate": 3.5123587570621466e-05,
+      "loss": 0.1991,
+      "step": 42630
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.3615834909766938,
+      "eval_loss": 0.9106101989746094,
+      "eval_rouge1": 0.6428131756085969,
+      "eval_rouge2": 0.38795383008457324,
+      "eval_rougeL": 0.6396112320774026,
+      "eval_runtime": 228.74,
+      "eval_samples_per_second": 62.119,
+      "eval_steps_per_second": 7.769,
+      "step": 42630
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.9299352169036865,
+      "learning_rate": 3.261475988700565e-05,
+      "loss": 0.1678,
+      "step": 49735
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.3620025104118138,
+      "eval_loss": 0.9511250257492065,
+      "eval_rouge1": 0.6421476450933186,
+      "eval_rouge2": 0.38754079327242963,
+      "eval_rougeL": 0.6388356718753687,
+      "eval_runtime": 228.7592,
+      "eval_samples_per_second": 62.113,
+      "eval_steps_per_second": 7.768,
+      "step": 49735
+    },
+    {
+      "epoch": 7.0,
+      "step": 49735,
+      "total_flos": 9.237135717983846e+16,
+      "train_loss": 0.40426278318054437,
+      "train_runtime": 12413.183,
+      "train_samples_per_second": 91.574,
+      "train_steps_per_second": 11.448
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 142100,
+  "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": 9.237135717983846e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}