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: 0.3947
-- Bleu: 0.4715
-- Rouge1: 0.7229
-- Rouge2: 0.4895
-- Rougel: 0.7189
+- Loss: 0.3767
+- Bleu: 0.4279
+- Rouge1: 0.6599
+- Rouge2: 0.3971
+- Rougel: 0.6585
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 8.0,
+    "eval_bleu": 0.427914494660841,
+    "eval_loss": 0.3767148554325104,
+    "eval_rouge1": 0.6598531398910279,
+    "eval_rouge2": 0.3971047774692673,
+    "eval_rougeL": 0.6584832139558556,
+    "eval_runtime": 2.4117,
+    "eval_samples": 304,
+    "eval_samples_per_second": 126.05,
+    "eval_steps_per_second": 15.756,
+    "perplexity": 1.4574886551410782,
+    "total_flos": 2264172293259264.0,
+    "train_loss": 0.719257706910177,
+    "train_runtime": 2187.326,
+    "train_samples": 1219,
+    "train_samples_per_second": 11.146,
+    "train_steps_per_second": 1.399
+}

egy_training_log.txt

@@ -156,3 +156,5 @@ INFO:root:Epoch 7.0: Train Loss = 0.0999, Eval Loss = 0.39377203583717346
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 8.0: Train Loss = 0.0844, Eval Loss = 0.39711037278175354
 INFO:absl:Using default tokenizer.
+INFO:__main__:*** Evaluate ***
+INFO:absl:Using default tokenizer.

eval_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 8.0,
+    "eval_bleu": 0.427914494660841,
+    "eval_loss": 0.3767148554325104,
+    "eval_rouge1": 0.6598531398910279,
+    "eval_rouge2": 0.3971047774692673,
+    "eval_rougeL": 0.6584832139558556,
+    "eval_runtime": 2.4117,
+    "eval_samples": 304,
+    "eval_samples_per_second": 126.05,
+    "eval_steps_per_second": 15.756,
+    "perplexity": 1.4574886551410782
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "total_flos": 2264172293259264.0,
+    "train_loss": 0.719257706910177,
+    "train_runtime": 2187.326,
+    "train_samples": 1219,
+    "train_samples_per_second": 11.146,
+    "train_steps_per_second": 1.399
+}

trainer_state.json

@@ -0,0 +1,203 @@
+{
+  "best_metric": 0.3767148554325104,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_yem_03/checkpoint-459",
+  "epoch": 8.0,
+  "eval_steps": 500,
+  "global_step": 1224,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 5.96675968170166,
+      "learning_rate": 1.53e-05,
+      "loss": 4.4383,
+      "step": 153
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.015528584651533126,
+      "eval_loss": 0.7436053156852722,
+      "eval_rouge1": 0.46982391145733055,
+      "eval_rouge2": 0.19207986874781602,
+      "eval_rougeL": 0.46930459291115545,
+      "eval_runtime": 2.3655,
+      "eval_samples_per_second": 128.516,
+      "eval_steps_per_second": 16.064,
+      "step": 153
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 3.5768725872039795,
+      "learning_rate": 3.06e-05,
+      "loss": 0.4733,
+      "step": 306
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.39670936867008716,
+      "eval_loss": 0.39333394169807434,
+      "eval_rouge1": 0.5938039437844622,
+      "eval_rouge2": 0.3153907013759276,
+      "eval_rougeL": 0.5909085576639257,
+      "eval_runtime": 9.4412,
+      "eval_samples_per_second": 32.199,
+      "eval_steps_per_second": 4.025,
+      "step": 306
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 3.4199442863464355,
+      "learning_rate": 4.5900000000000004e-05,
+      "loss": 0.2714,
+      "step": 459
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.427914494660841,
+      "eval_loss": 0.3767148554325104,
+      "eval_rouge1": 0.6598531398910279,
+      "eval_rouge2": 0.3971047774692673,
+      "eval_rougeL": 0.6584832139558556,
+      "eval_runtime": 9.4581,
+      "eval_samples_per_second": 32.142,
+      "eval_steps_per_second": 4.018,
+      "step": 459
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 2.4085845947265625,
+      "learning_rate": 4.7812500000000003e-05,
+      "loss": 0.1837,
+      "step": 612
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.43801708242621135,
+      "eval_loss": 0.3863435983657837,
+      "eval_rouge1": 0.6761692862282009,
+      "eval_rouge2": 0.41837300418726475,
+      "eval_rougeL": 0.6725637079403999,
+      "eval_runtime": 127.7574,
+      "eval_samples_per_second": 2.38,
+      "eval_steps_per_second": 0.297,
+      "step": 612
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.990710973739624,
+      "learning_rate": 4.482421875e-05,
+      "loss": 0.129,
+      "step": 765
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.4536061292967932,
+      "eval_loss": 0.38353848457336426,
+      "eval_rouge1": 0.6993772519271104,
+      "eval_rouge2": 0.4575839398398158,
+      "eval_rougeL": 0.6967680711913622,
+      "eval_runtime": 4.8885,
+      "eval_samples_per_second": 62.186,
+      "eval_steps_per_second": 7.773,
+      "step": 765
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.9595922827720642,
+      "learning_rate": 4.18359375e-05,
+      "loss": 0.0999,
+      "step": 918
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.44934275646340344,
+      "eval_loss": 0.39377203583717346,
+      "eval_rouge1": 0.6978745123044579,
+      "eval_rouge2": 0.45411326233563776,
+      "eval_rougeL": 0.693627688943746,
+      "eval_runtime": 2.8238,
+      "eval_samples_per_second": 107.657,
+      "eval_steps_per_second": 13.457,
+      "step": 918
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.0308947563171387,
+      "learning_rate": 3.884765625e-05,
+      "loss": 0.0844,
+      "step": 1071
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.46229210583360764,
+      "eval_loss": 0.39711037278175354,
+      "eval_rouge1": 0.7106823471740515,
+      "eval_rouge2": 0.4766332327393774,
+      "eval_rougeL": 0.7069100507308859,
+      "eval_runtime": 2.6065,
+      "eval_samples_per_second": 116.631,
+      "eval_steps_per_second": 14.579,
+      "step": 1071
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.5569229125976562,
+      "learning_rate": 3.5859375e-05,
+      "loss": 0.0741,
+      "step": 1224
+    },
+    {
+      "epoch": 8.0,
+      "eval_bleu": 0.4714652829533398,
+      "eval_loss": 0.3946935832500458,
+      "eval_rouge1": 0.7228506957910504,
+      "eval_rouge2": 0.48947047623829953,
+      "eval_rougeL": 0.7188907468330396,
+      "eval_runtime": 3.6852,
+      "eval_samples_per_second": 82.492,
+      "eval_steps_per_second": 10.311,
+      "step": 1224
+    },
+    {
+      "epoch": 8.0,
+      "step": 1224,
+      "total_flos": 2264172293259264.0,
+      "train_loss": 0.719257706910177,
+      "train_runtime": 2187.326,
+      "train_samples_per_second": 11.146,
+      "train_steps_per_second": 1.399
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 3060,
+  "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": 2264172293259264.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}