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: 0.0797
-- Bleu: 0.2112
-- Rouge1: 0.5592
-- Rouge2: 0.3546
-- Rougel: 0.5569
+- Loss: 0.0679
+- Bleu: 0.1681
+- Rouge1: 0.5343
+- Rouge2: 0.3158
+- Rougel: 0.5319
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 8.0,
+    "eval_bleu": 0.16813566870280167,
+    "eval_loss": 0.0678805559873581,
+    "eval_rouge1": 0.5342839149814531,
+    "eval_rouge2": 0.3157655613133348,
+    "eval_rougeL": 0.5319180103197795,
+    "eval_runtime": 2024.0604,
+    "eval_samples": 33746,
+    "eval_samples_per_second": 16.672,
+    "eval_steps_per_second": 4.168,
+    "perplexity": 1.0702374673864155,
+    "total_flos": 2.0060826554471547e+18,
+    "train_loss": 0.04765514560038514,
+    "train_runtime": 203650.1496,
+    "train_samples": 135006,
+    "train_samples_per_second": 13.259,
+    "train_steps_per_second": 3.315
+}

egy_training_log.txt

@@ -156,3 +156,5 @@ INFO:root:Epoch 7.0: Train Loss = 0.0345, Eval Loss = 0.07342757284641266
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 8.0: Train Loss = 0.0302, Eval Loss = 0.0768970176577568
 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": 0.16813566870280167,
+    "eval_loss": 0.0678805559873581,
+    "eval_rouge1": 0.5342839149814531,
+    "eval_rouge2": 0.3157655613133348,
+    "eval_rougeL": 0.5319180103197795,
+    "eval_runtime": 2024.0604,
+    "eval_samples": 33746,
+    "eval_samples_per_second": 16.672,
+    "eval_steps_per_second": 4.168,
+    "perplexity": 1.0702374673864155
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.0,
+    "total_flos": 2.0060826554471547e+18,
+    "train_loss": 0.04765514560038514,
+    "train_runtime": 203650.1496,
+    "train_samples": 135006,
+    "train_samples_per_second": 13.259,
+    "train_steps_per_second": 3.315
+}

trainer_state.json

@@ -0,0 +1,203 @@
+{
+  "best_metric": 0.0678805559873581,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_arabiangpt0.3/checkpoint-101256",
+  "epoch": 8.0,
+  "eval_steps": 500,
+  "global_step": 270016,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.5094387531280518,
+      "learning_rate": 4.753520917958906e-05,
+      "loss": 0.0845,
+      "step": 33752
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.1360083183193832,
+      "eval_loss": 0.0708920955657959,
+      "eval_rouge1": 0.48412628006639347,
+      "eval_rouge2": 0.2595444928790475,
+      "eval_rougeL": 0.4811152550714598,
+      "eval_runtime": 2076.8563,
+      "eval_samples_per_second": 16.249,
+      "eval_steps_per_second": 4.062,
+      "step": 33752
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.2100098133087158,
+      "learning_rate": 4.503335606487384e-05,
+      "loss": 0.0638,
+      "step": 67504
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.15568253204889895,
+      "eval_loss": 0.06831026077270508,
+      "eval_rouge1": 0.516649446787459,
+      "eval_rouge2": 0.2948018415714037,
+      "eval_rougeL": 0.514024802556861,
+      "eval_runtime": 2012.7332,
+      "eval_samples_per_second": 16.766,
+      "eval_steps_per_second": 4.192,
+      "step": 67504
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.4080255627632141,
+      "learning_rate": 4.253150295015863e-05,
+      "loss": 0.0546,
+      "step": 101256
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.16813566870280167,
+      "eval_loss": 0.0678805559873581,
+      "eval_rouge1": 0.5342839149814531,
+      "eval_rouge2": 0.3157655613133348,
+      "eval_rougeL": 0.5319180103197795,
+      "eval_runtime": 1952.412,
+      "eval_samples_per_second": 17.284,
+      "eval_steps_per_second": 4.321,
+      "step": 101256
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.40639498829841614,
+      "learning_rate": 4.002964983544342e-05,
+      "loss": 0.0468,
+      "step": 135008
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.18301542857870237,
+      "eval_loss": 0.06886506825685501,
+      "eval_rouge1": 0.5476940640774866,
+      "eval_rouge2": 0.33203989935219014,
+      "eval_rougeL": 0.5452209555156204,
+      "eval_runtime": 1954.1613,
+      "eval_samples_per_second": 17.269,
+      "eval_steps_per_second": 4.317,
+      "step": 135008
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.5491040945053101,
+      "learning_rate": 3.7527796720728204e-05,
+      "loss": 0.0401,
+      "step": 168760
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.19405846737344298,
+      "eval_loss": 0.0706464946269989,
+      "eval_rouge1": 0.5548980705277954,
+      "eval_rouge2": 0.34218497796474945,
+      "eval_rougeL": 0.55265645352457,
+      "eval_runtime": 2015.2556,
+      "eval_samples_per_second": 16.745,
+      "eval_steps_per_second": 4.187,
+      "step": 168760
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 0.2615252435207367,
+      "learning_rate": 3.502594360601299e-05,
+      "loss": 0.0345,
+      "step": 202512
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.20164731422462528,
+      "eval_loss": 0.07342757284641266,
+      "eval_rouge1": 0.5579768475101489,
+      "eval_rouge2": 0.34858517959726787,
+      "eval_rougeL": 0.5558276150233377,
+      "eval_runtime": 2028.3934,
+      "eval_samples_per_second": 16.637,
+      "eval_steps_per_second": 4.159,
+      "step": 202512
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 0.24876318871974945,
+      "learning_rate": 3.252409049129777e-05,
+      "loss": 0.0302,
+      "step": 236264
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.2083227447079373,
+      "eval_loss": 0.0768970176577568,
+      "eval_rouge1": 0.5600411658774835,
+      "eval_rouge2": 0.3535281728359022,
+      "eval_rougeL": 0.5576670140997706,
+      "eval_runtime": 2016.181,
+      "eval_samples_per_second": 16.738,
+      "eval_steps_per_second": 4.185,
+      "step": 236264
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.5295557379722595,
+      "learning_rate": 3.0022237376582564e-05,
+      "loss": 0.0269,
+      "step": 270016
+    },
+    {
+      "epoch": 8.0,
+      "eval_bleu": 0.21123631944078344,
+      "eval_loss": 0.07967726141214371,
+      "eval_rouge1": 0.5591904067243532,
+      "eval_rouge2": 0.3545914045960048,
+      "eval_rougeL": 0.5569180306629817,
+      "eval_runtime": 2017.1957,
+      "eval_samples_per_second": 16.729,
+      "eval_steps_per_second": 4.183,
+      "step": 270016
+    },
+    {
+      "epoch": 8.0,
+      "step": 270016,
+      "total_flos": 2.0060826554471547e+18,
+      "train_loss": 0.04765514560038514,
+      "train_runtime": 203650.1496,
+      "train_samples_per_second": 13.259,
+      "train_steps_per_second": 3.315
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 675040,
+  "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": 2.0060826554471547e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}