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-01B](https://huggingface.co/riotu-lab/ArabianGPT-01B) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4525
-- Bleu: 0.2539
-- Rouge1: 0.6407
-- Rouge2: 0.3977
-- Rougel: 0.6393
+- Loss: 0.4183
+- Bleu: 0.2374
+- Rouge1: 0.6213
+- Rouge2: 0.3630
+- Rougel: 0.6195
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 9.0,
+    "eval_bleu": 0.23736106874390867,
+    "eval_loss": 0.4183219075202942,
+    "eval_rouge1": 0.6213217050496613,
+    "eval_rouge2": 0.3629928034660268,
+    "eval_rougeL": 0.6194945230812778,
+    "eval_runtime": 7.4528,
+    "eval_samples": 1675,
+    "eval_samples_per_second": 224.748,
+    "eval_steps_per_second": 28.177,
+    "perplexity": 1.5194097050971664,
+    "total_flos": 3938977382400000.0,
+    "train_loss": 0.41718409883979146,
+    "train_runtime": 2958.6328,
+    "train_samples": 6700,
+    "train_samples_per_second": 45.291,
+    "train_steps_per_second": 5.665
+}

egy_training_log.txt

@@ -162,3 +162,5 @@ INFO:root:Epoch 8.0: Train Loss = 0.148, Eval Loss = 0.4369480013847351
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 9.0: Train Loss = 0.1285, Eval Loss = 0.4447513520717621
 INFO:absl:Using default tokenizer.
+INFO:__main__:*** Evaluate ***
+INFO:absl:Using default tokenizer.

eval_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 9.0,
+    "eval_bleu": 0.23736106874390867,
+    "eval_loss": 0.4183219075202942,
+    "eval_rouge1": 0.6213217050496613,
+    "eval_rouge2": 0.3629928034660268,
+    "eval_rougeL": 0.6194945230812778,
+    "eval_runtime": 7.4528,
+    "eval_samples": 1675,
+    "eval_samples_per_second": 224.748,
+    "eval_steps_per_second": 28.177,
+    "perplexity": 1.5194097050971664
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.0,
+    "total_flos": 3938977382400000.0,
+    "train_loss": 0.41718409883979146,
+    "train_runtime": 2958.6328,
+    "train_samples": 6700,
+    "train_samples_per_second": 45.291,
+    "train_steps_per_second": 5.665
+}

trainer_state.json

@@ -0,0 +1,222 @@
+{
+  "best_metric": 0.4183219075202942,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_gulf/checkpoint-3352",
+  "epoch": 9.0,
+  "eval_steps": 500,
+  "global_step": 7542,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.1270569562911987,
+      "learning_rate": 4.8960639606396064e-05,
+      "loss": 1.9575,
+      "step": 838
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.17899151773967856,
+      "eval_loss": 0.4805418848991394,
+      "eval_rouge1": 0.5300700267573193,
+      "eval_rouge2": 0.25846419107204655,
+      "eval_rougeL": 0.5287371290578213,
+      "eval_runtime": 26.4765,
+      "eval_samples_per_second": 63.264,
+      "eval_steps_per_second": 7.932,
+      "step": 838
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.8918192982673645,
+      "learning_rate": 4.6383763837638376e-05,
+      "loss": 0.422,
+      "step": 1676
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.2111987576275603,
+      "eval_loss": 0.4326799809932709,
+      "eval_rouge1": 0.5789346222407353,
+      "eval_rouge2": 0.3115433580134526,
+      "eval_rougeL": 0.5777838651600141,
+      "eval_runtime": 36.5489,
+      "eval_samples_per_second": 45.829,
+      "eval_steps_per_second": 5.746,
+      "step": 1676
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.1533405780792236,
+      "learning_rate": 4.380688806888069e-05,
+      "loss": 0.331,
+      "step": 2514
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.22887638181669745,
+      "eval_loss": 0.41845184564590454,
+      "eval_rouge1": 0.6066983915980586,
+      "eval_rouge2": 0.34626660980048474,
+      "eval_rougeL": 0.6055721645591957,
+      "eval_runtime": 25.3383,
+      "eval_samples_per_second": 66.106,
+      "eval_steps_per_second": 8.288,
+      "step": 2514
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 1.3615330457687378,
+      "learning_rate": 4.123001230012301e-05,
+      "loss": 0.2638,
+      "step": 3352
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.23736106874390867,
+      "eval_loss": 0.4183219075202942,
+      "eval_rouge1": 0.6213217050496613,
+      "eval_rouge2": 0.3629928034660268,
+      "eval_rougeL": 0.6194945230812778,
+      "eval_runtime": 16.1799,
+      "eval_samples_per_second": 103.524,
+      "eval_steps_per_second": 12.979,
+      "step": 3352
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 1.29606294631958,
+      "learning_rate": 3.865313653136531e-05,
+      "loss": 0.2131,
+      "step": 4190
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.24225546513762886,
+      "eval_loss": 0.42484816908836365,
+      "eval_rouge1": 0.6280394265150125,
+      "eval_rouge2": 0.3718667234148674,
+      "eval_rougeL": 0.6264526155291095,
+      "eval_runtime": 25.4283,
+      "eval_samples_per_second": 65.872,
+      "eval_steps_per_second": 8.259,
+      "step": 4190
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 1.9162726402282715,
+      "learning_rate": 3.6076260762607624e-05,
+      "loss": 0.1756,
+      "step": 5028
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.25110200021461276,
+      "eval_loss": 0.4322951138019562,
+      "eval_rouge1": 0.6329099900871186,
+      "eval_rouge2": 0.3818443278310246,
+      "eval_rougeL": 0.6311516008372622,
+      "eval_runtime": 13.3457,
+      "eval_samples_per_second": 125.508,
+      "eval_steps_per_second": 15.735,
+      "step": 5028
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.2213388681411743,
+      "learning_rate": 3.349938499384994e-05,
+      "loss": 0.148,
+      "step": 5866
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.24584223218153298,
+      "eval_loss": 0.4369480013847351,
+      "eval_rouge1": 0.6346640667719812,
+      "eval_rouge2": 0.3890963668887653,
+      "eval_rougeL": 0.6326978702709232,
+      "eval_runtime": 6.3313,
+      "eval_samples_per_second": 264.559,
+      "eval_steps_per_second": 33.169,
+      "step": 5866
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 1.774276852607727,
+      "learning_rate": 3.092250922509225e-05,
+      "loss": 0.1285,
+      "step": 6704
+    },
+    {
+      "epoch": 8.0,
+      "eval_bleu": 0.2499016538097514,
+      "eval_loss": 0.4447513520717621,
+      "eval_rouge1": 0.6380175687280993,
+      "eval_rouge2": 0.38845767776215345,
+      "eval_rougeL": 0.6368182514454324,
+      "eval_runtime": 5.4681,
+      "eval_samples_per_second": 306.322,
+      "eval_steps_per_second": 38.405,
+      "step": 6704
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 0.779574990272522,
+      "learning_rate": 2.8345633456334564e-05,
+      "loss": 0.1152,
+      "step": 7542
+    },
+    {
+      "epoch": 9.0,
+      "eval_bleu": 0.2538776211394201,
+      "eval_loss": 0.45251065492630005,
+      "eval_rouge1": 0.6407467681145795,
+      "eval_rouge2": 0.3976657405838769,
+      "eval_rougeL": 0.6393461853572592,
+      "eval_runtime": 6.3546,
+      "eval_samples_per_second": 263.59,
+      "eval_steps_per_second": 33.047,
+      "step": 7542
+    },
+    {
+      "epoch": 9.0,
+      "step": 7542,
+      "total_flos": 3938977382400000.0,
+      "train_loss": 0.41718409883979146,
+      "train_runtime": 2958.6328,
+      "train_samples_per_second": 45.291,
+      "train_steps_per_second": 5.665
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 16760,
+  "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": 3938977382400000.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}