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: 1.3693
-- Bleu: 0.4248
-- Rouge1: 0.6534
-- Rouge2: 0.4079
-- Rougel: 0.6523
+- Loss: 0.9735
+- Bleu: 0.3759
+- Rouge1: 0.5621
+- Rouge2: 0.3073
+- Rougel: 0.5603
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 6.0,
+    "eval_bleu": 0.3758671129852906,
+    "eval_loss": 0.9734994769096375,
+    "eval_rouge1": 0.5620694784691431,
+    "eval_rouge2": 0.3072634138927888,
+    "eval_rougeL": 0.5602967261743523,
+    "eval_runtime": 136.3007,
+    "eval_samples": 2113,
+    "eval_samples_per_second": 15.502,
+    "eval_steps_per_second": 1.944,
+    "perplexity": 2.647192056517508,
+    "total_flos": 1.1774067405225984e+16,
+    "train_loss": 0.37803844186042396,
+    "train_runtime": 4074.3615,
+    "train_samples": 8452,
+    "train_samples_per_second": 41.489,
+    "train_steps_per_second": 5.189
+}

egy_training_log.txt

@@ -152,3 +152,5 @@ INFO:root:Epoch 5.0: Train Loss = 0.1771, Eval Loss = 1.3314274549484253
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 6.0: Train Loss = 0.1337, Eval Loss = 1.344425916671753
 INFO:absl:Using default tokenizer.
+INFO:__main__:*** Evaluate ***
+INFO:absl:Using default tokenizer.

eval_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 6.0,
+    "eval_bleu": 0.3758671129852906,
+    "eval_loss": 0.9734994769096375,
+    "eval_rouge1": 0.5620694784691431,
+    "eval_rouge2": 0.3072634138927888,
+    "eval_rougeL": 0.5602967261743523,
+    "eval_runtime": 136.3007,
+    "eval_samples": 2113,
+    "eval_samples_per_second": 15.502,
+    "eval_steps_per_second": 1.944,
+    "perplexity": 2.647192056517508
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 6.0,
+    "total_flos": 1.1774067405225984e+16,
+    "train_loss": 0.37803844186042396,
+    "train_runtime": 4074.3615,
+    "train_samples": 8452,
+    "train_samples_per_second": 41.489,
+    "train_steps_per_second": 5.189
+}

trainer_state.json

@@ -0,0 +1,165 @@
+{
+  "best_metric": 0.9734994769096375,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_irq_03/checkpoint-1057",
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 6342,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 2.316096782684326,
+      "learning_rate": 4.865067829457365e-05,
+      "loss": 1.1923,
+      "step": 1057
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.3758671129852906,
+      "eval_loss": 0.9734994769096375,
+      "eval_rouge1": 0.5620694784691431,
+      "eval_rouge2": 0.3072634138927888,
+      "eval_rougeL": 0.5602967261743523,
+      "eval_runtime": 18.3819,
+      "eval_samples_per_second": 114.95,
+      "eval_steps_per_second": 14.416,
+      "step": 1057
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.9589498043060303,
+      "learning_rate": 4.609011627906977e-05,
+      "loss": 0.3956,
+      "step": 2114
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.41456228931689637,
+      "eval_loss": 1.2529053688049316,
+      "eval_rouge1": 0.6342121068042533,
+      "eval_rouge2": 0.3777098530906886,
+      "eval_rougeL": 0.6329704343270921,
+      "eval_runtime": 27.6701,
+      "eval_samples_per_second": 76.364,
+      "eval_steps_per_second": 9.577,
+      "step": 2114
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.9037576913833618,
+      "learning_rate": 4.3529554263565894e-05,
+      "loss": 0.2578,
+      "step": 3171
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.4177567847266788,
+      "eval_loss": 1.1432485580444336,
+      "eval_rouge1": 0.6356354703691816,
+      "eval_rouge2": 0.38873743710479597,
+      "eval_rougeL": 0.6340522473648312,
+      "eval_runtime": 16.0819,
+      "eval_samples_per_second": 131.39,
+      "eval_steps_per_second": 16.478,
+      "step": 3171
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 3.466954469680786,
+      "learning_rate": 4.096899224806201e-05,
+      "loss": 0.1771,
+      "step": 4228
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.420916242293927,
+      "eval_loss": 1.3314274549484253,
+      "eval_rouge1": 0.6564326494399506,
+      "eval_rouge2": 0.4082972964966378,
+      "eval_rougeL": 0.6545946336979762,
+      "eval_runtime": 15.9444,
+      "eval_samples_per_second": 132.523,
+      "eval_steps_per_second": 16.62,
+      "step": 4228
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.057302236557007,
+      "learning_rate": 3.840843023255814e-05,
+      "loss": 0.1337,
+      "step": 5285
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.42563539215296015,
+      "eval_loss": 1.344425916671753,
+      "eval_rouge1": 0.6508055039664861,
+      "eval_rouge2": 0.40613669991922435,
+      "eval_rougeL": 0.6494650833785085,
+      "eval_runtime": 139.4783,
+      "eval_samples_per_second": 15.149,
+      "eval_steps_per_second": 1.9,
+      "step": 5285
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 1.6741355657577515,
+      "learning_rate": 3.5847868217054265e-05,
+      "loss": 0.1118,
+      "step": 6342
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.42475546499314026,
+      "eval_loss": 1.369273066520691,
+      "eval_rouge1": 0.6534368445259939,
+      "eval_rouge2": 0.40793487128864964,
+      "eval_rougeL": 0.6523435258107617,
+      "eval_runtime": 22.7821,
+      "eval_samples_per_second": 92.748,
+      "eval_steps_per_second": 11.632,
+      "step": 6342
+    },
+    {
+      "epoch": 6.0,
+      "step": 6342,
+      "total_flos": 1.1774067405225984e+16,
+      "train_loss": 0.37803844186042396,
+      "train_runtime": 4074.3615,
+      "train_samples_per_second": 41.489,
+      "train_steps_per_second": 5.189
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 21140,
+  "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": 1.1774067405225984e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}