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: 0.3964
-- Bleu: 0.4567
-- Rouge1: 0.7127
-- Rouge2: 0.4779
-- Rougel: 0.7120
+- Loss: 0.3495
+- Bleu: 0.4355
+- Rouge1: 0.6808
+- Rouge2: 0.4218
+- Rougel: 0.6799
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 7.0,
+    "eval_bleu": 0.43550162080921356,
+    "eval_loss": 0.3495321273803711,
+    "eval_rouge1": 0.6807566585005222,
+    "eval_rouge2": 0.42184236623206983,
+    "eval_rougeL": 0.6798697145255593,
+    "eval_runtime": 135.4262,
+    "eval_samples": 1672,
+    "eval_samples_per_second": 12.346,
+    "eval_steps_per_second": 1.543,
+    "perplexity": 1.4184037610379865,
+    "total_flos": 1.086463725010944e+16,
+    "train_loss": 0.29683067467794466,
+    "train_runtime": 4067.7631,
+    "train_samples": 6685,
+    "train_samples_per_second": 32.868,
+    "train_steps_per_second": 4.11
+}

egy_training_log.txt

@@ -154,3 +154,5 @@ INFO:root:Epoch 6.0: Train Loss = 0.11, Eval Loss = 0.3791240453720093
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 7.0: Train Loss = 0.0981, Eval Loss = 0.39224255084991455
 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": 0.43550162080921356,
+    "eval_loss": 0.3495321273803711,
+    "eval_rouge1": 0.6807566585005222,
+    "eval_rouge2": 0.42184236623206983,
+    "eval_rougeL": 0.6798697145255593,
+    "eval_runtime": 135.4262,
+    "eval_samples": 1672,
+    "eval_samples_per_second": 12.346,
+    "eval_steps_per_second": 1.543,
+    "perplexity": 1.4184037610379865
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 7.0,
+    "total_flos": 1.086463725010944e+16,
+    "train_loss": 0.29683067467794466,
+    "train_runtime": 4067.7631,
+    "train_samples": 6685,
+    "train_samples_per_second": 32.868,
+    "train_steps_per_second": 4.11
+}

trainer_state.json

@@ -0,0 +1,184 @@
+{
+  "best_metric": 0.3495321273803711,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_gulf_03/checkpoint-1672",
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 5852,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.8581334352493286,
+      "learning_rate": 4.896424167694204e-05,
+      "loss": 1.1759,
+      "step": 836
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.41556477975575595,
+      "eval_loss": 0.37735649943351746,
+      "eval_rouge1": 0.6364308285085805,
+      "eval_rouge2": 0.36296002645709646,
+      "eval_rougeL": 0.6351682159497727,
+      "eval_runtime": 16.4559,
+      "eval_samples_per_second": 101.605,
+      "eval_steps_per_second": 12.701,
+      "step": 836
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.0768027305603027,
+      "learning_rate": 4.638717632552405e-05,
+      "loss": 0.2824,
+      "step": 1672
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.43550162080921356,
+      "eval_loss": 0.3495321273803711,
+      "eval_rouge1": 0.6807566585005222,
+      "eval_rouge2": 0.42184236623206983,
+      "eval_rougeL": 0.6798697145255593,
+      "eval_runtime": 12.3396,
+      "eval_samples_per_second": 135.499,
+      "eval_steps_per_second": 16.937,
+      "step": 1672
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.4890309572219849,
+      "learning_rate": 4.3810110974106046e-05,
+      "loss": 0.1846,
+      "step": 2508
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.44798984557044996,
+      "eval_loss": 0.35791030526161194,
+      "eval_rouge1": 0.6988772317351744,
+      "eval_rouge2": 0.4541195944585126,
+      "eval_rougeL": 0.6983599524493045,
+      "eval_runtime": 13.092,
+      "eval_samples_per_second": 127.712,
+      "eval_steps_per_second": 15.964,
+      "step": 2508
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.9949209690093994,
+      "learning_rate": 4.1233045622688044e-05,
+      "loss": 0.1351,
+      "step": 3344
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.4449484201171957,
+      "eval_loss": 0.36742404103279114,
+      "eval_rouge1": 0.6992663461784911,
+      "eval_rouge2": 0.45557653256736347,
+      "eval_rougeL": 0.6984837283971637,
+      "eval_runtime": 12.3128,
+      "eval_samples_per_second": 135.794,
+      "eval_steps_per_second": 16.974,
+      "step": 3344
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 1.245552897453308,
+      "learning_rate": 3.8655980271270036e-05,
+      "loss": 0.11,
+      "step": 4180
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.4488174409372482,
+      "eval_loss": 0.3791240453720093,
+      "eval_rouge1": 0.7066508462378398,
+      "eval_rouge2": 0.46813923665655116,
+      "eval_rougeL": 0.7055578408590141,
+      "eval_runtime": 18.2521,
+      "eval_samples_per_second": 91.606,
+      "eval_steps_per_second": 11.451,
+      "step": 4180
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 1.253645896911621,
+      "learning_rate": 3.6078914919852034e-05,
+      "loss": 0.0981,
+      "step": 5016
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.4582553582864403,
+      "eval_loss": 0.39224255084991455,
+      "eval_rouge1": 0.7135223253955782,
+      "eval_rouge2": 0.4789126140106226,
+      "eval_rougeL": 0.7129027733719826,
+      "eval_runtime": 65.9112,
+      "eval_samples_per_second": 25.367,
+      "eval_steps_per_second": 3.171,
+      "step": 5016
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.6491497755050659,
+      "learning_rate": 3.350184956843403e-05,
+      "loss": 0.0917,
+      "step": 5852
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.4567065309085701,
+      "eval_loss": 0.3964388072490692,
+      "eval_rouge1": 0.7126519996998933,
+      "eval_rouge2": 0.47787123299802714,
+      "eval_rougeL": 0.7120354739021442,
+      "eval_runtime": 21.7898,
+      "eval_samples_per_second": 76.733,
+      "eval_steps_per_second": 9.592,
+      "step": 5852
+    },
+    {
+      "epoch": 7.0,
+      "step": 5852,
+      "total_flos": 1.086463725010944e+16,
+      "train_loss": 0.29683067467794466,
+      "train_runtime": 4067.7631,
+      "train_samples_per_second": 32.868,
+      "train_steps_per_second": 4.11
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 16720,
+  "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": 1.086463725010944e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}