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: 0.7974
-- Bleu: 0.2491
-- Rouge1: 0.6112
-- Rouge2: 0.3654
-- Rougel: 0.6074
+- Loss: 0.7316
+- Bleu: 0.2459
+- Rouge1: 0.6017
+- Rouge2: 0.3519
+- Rougel: 0.5983
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 9.0,
+    "eval_bleu": 0.24587227576979195,
+    "eval_loss": 0.7315686941146851,
+    "eval_rouge1": 0.6017197427075045,
+    "eval_rouge2": 0.3518746485163118,
+    "eval_rougeL": 0.5982542515796094,
+    "eval_runtime": 47.1591,
+    "eval_samples": 14212,
+    "eval_samples_per_second": 301.363,
+    "eval_steps_per_second": 37.681,
+    "perplexity": 2.078338328926906,
+    "total_flos": 3.3423104950272e+16,
+    "train_loss": 0.5779510789562912,
+    "train_runtime": 8660.6869,
+    "train_samples": 56851,
+    "train_samples_per_second": 131.285,
+    "train_steps_per_second": 16.412
+}

egy_training_log.txt

@@ -162,3 +162,5 @@ INFO:root:Epoch 8.0: Train Loss = 0.4118, Eval Loss = 0.7635838389396667
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 9.0: Train Loss = 0.3725, Eval Loss = 0.7796261310577393
 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": 0.24587227576979195,
+    "eval_loss": 0.7315686941146851,
+    "eval_rouge1": 0.6017197427075045,
+    "eval_rouge2": 0.3518746485163118,
+    "eval_rougeL": 0.5982542515796094,
+    "eval_runtime": 47.1591,
+    "eval_samples": 14212,
+    "eval_samples_per_second": 301.363,
+    "eval_steps_per_second": 37.681,
+    "perplexity": 2.078338328926906
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.0,
+    "total_flos": 3.3423104950272e+16,
+    "train_loss": 0.5779510789562912,
+    "train_runtime": 8660.6869,
+    "train_samples": 56851,
+    "train_samples_per_second": 131.285,
+    "train_steps_per_second": 16.412
+}

trainer_state.json

@@ -0,0 +1,222 @@
+{
+  "best_metric": 0.7315686941146851,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_eg/checkpoint-28428",
+  "epoch": 9.0,
+  "eval_steps": 500,
+  "global_step": 63963,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.742945671081543,
+      "learning_rate": 4.7667678621858235e-05,
+      "loss": 1.1436,
+      "step": 7107
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.1900415894207328,
+      "eval_loss": 0.8277140259742737,
+      "eval_rouge1": 0.5211536867388353,
+      "eval_rouge2": 0.2576275131704426,
+      "eval_rougeL": 0.5169189427573101,
+      "eval_runtime": 204.5959,
+      "eval_samples_per_second": 69.464,
+      "eval_steps_per_second": 8.685,
+      "step": 7107
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 3.031801223754883,
+      "learning_rate": 4.515885343123411e-05,
+      "loss": 0.7508,
+      "step": 14214
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.22138684401610842,
+      "eval_loss": 0.7543078064918518,
+      "eval_rouge1": 0.5674397247471176,
+      "eval_rouge2": 0.3108337383535441,
+      "eval_rougeL": 0.5636106781794015,
+      "eval_runtime": 171.1246,
+      "eval_samples_per_second": 83.051,
+      "eval_steps_per_second": 10.384,
+      "step": 14214
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.8590487241744995,
+      "learning_rate": 4.265002824061e-05,
+      "loss": 0.6471,
+      "step": 21321
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.2374960454489342,
+      "eval_loss": 0.7337948083877563,
+      "eval_rouge1": 0.5880985827608463,
+      "eval_rouge2": 0.33558513842625187,
+      "eval_rougeL": 0.5844518671510625,
+      "eval_runtime": 68.7002,
+      "eval_samples_per_second": 206.87,
+      "eval_steps_per_second": 25.866,
+      "step": 21321
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 1.3944027423858643,
+      "learning_rate": 4.0141203049985884e-05,
+      "loss": 0.5713,
+      "step": 28428
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.24587227576979195,
+      "eval_loss": 0.7315686941146851,
+      "eval_rouge1": 0.6017197427075045,
+      "eval_rouge2": 0.3518746485163118,
+      "eval_rougeL": 0.5982542515796094,
+      "eval_runtime": 62.0721,
+      "eval_samples_per_second": 228.959,
+      "eval_steps_per_second": 28.628,
+      "step": 28428
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.199220657348633,
+      "learning_rate": 3.763237785936176e-05,
+      "loss": 0.5097,
+      "step": 35535
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.24748155317226092,
+      "eval_loss": 0.7390380501747131,
+      "eval_rouge1": 0.6058102682046419,
+      "eval_rouge2": 0.357170685615976,
+      "eval_rougeL": 0.6021635755679425,
+      "eval_runtime": 67.8747,
+      "eval_samples_per_second": 209.386,
+      "eval_steps_per_second": 26.181,
+      "step": 35535
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 2.055725574493408,
+      "learning_rate": 3.512355266873765e-05,
+      "loss": 0.4573,
+      "step": 42642
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.25030630377831276,
+      "eval_loss": 0.748293399810791,
+      "eval_rouge1": 0.6103116816448397,
+      "eval_rouge2": 0.361846050958361,
+      "eval_rougeL": 0.6066395364597333,
+      "eval_runtime": 56.4418,
+      "eval_samples_per_second": 251.799,
+      "eval_steps_per_second": 31.484,
+      "step": 42642
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.6595733165740967,
+      "learning_rate": 3.2614727478113526e-05,
+      "loss": 0.4118,
+      "step": 49749
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.2494244558337241,
+      "eval_loss": 0.7635838389396667,
+      "eval_rouge1": 0.610621109140437,
+      "eval_rouge2": 0.3633959713058441,
+      "eval_rougeL": 0.6069537363647842,
+      "eval_runtime": 173.9311,
+      "eval_samples_per_second": 81.711,
+      "eval_steps_per_second": 10.217,
+      "step": 49749
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 3.863671064376831,
+      "learning_rate": 3.010590228748941e-05,
+      "loss": 0.3725,
+      "step": 56856
+    },
+    {
+      "epoch": 8.0,
+      "eval_bleu": 0.25065847486647275,
+      "eval_loss": 0.7796261310577393,
+      "eval_rouge1": 0.6126587801190159,
+      "eval_rouge2": 0.3659624175392553,
+      "eval_rougeL": 0.6088959046619336,
+      "eval_runtime": 170.86,
+      "eval_samples_per_second": 83.179,
+      "eval_steps_per_second": 10.4,
+      "step": 56856
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 3.546931266784668,
+      "learning_rate": 2.7597077096865293e-05,
+      "loss": 0.3375,
+      "step": 63963
+    },
+    {
+      "epoch": 9.0,
+      "eval_bleu": 0.24908190761452426,
+      "eval_loss": 0.7973926663398743,
+      "eval_rouge1": 0.6111967178899901,
+      "eval_rouge2": 0.36536691181853787,
+      "eval_rougeL": 0.6074289902749841,
+      "eval_runtime": 173.0755,
+      "eval_samples_per_second": 82.114,
+      "eval_steps_per_second": 10.267,
+      "step": 63963
+    },
+    {
+      "epoch": 9.0,
+      "step": 63963,
+      "total_flos": 3.3423104950272e+16,
+      "train_loss": 0.5779510789562912,
+      "train_runtime": 8660.6869,
+      "train_samples_per_second": 131.285,
+      "train_steps_per_second": 16.412
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 142140,
+  "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": 3.3423104950272e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}