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: 0.6491
-- Bleu: 0.4280
-- Rouge1: 0.6731
-- Rouge2: 0.4383
-- Rougel: 0.6715
+- Loss: 0.5335
+- Bleu: 0.4225
+- Rouge1: 0.6608
+- Rouge2: 0.4163
+- Rougel: 0.6590
 
 ## Model description
 

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 7.0,
+    "eval_bleu": 0.42251895839624176,
+    "eval_loss": 0.5335076451301575,
+    "eval_rouge1": 0.6607626579317042,
+    "eval_rouge2": 0.4163462041346451,
+    "eval_rougeL": 0.6589800378443118,
+    "eval_runtime": 57.7071,
+    "eval_samples": 5380,
+    "eval_samples_per_second": 93.229,
+    "eval_steps_per_second": 11.662,
+    "perplexity": 1.704902023958078,
+    "total_flos": 3.4826972570320896e+16,
+    "train_loss": 0.31742249039632825,
+    "train_runtime": 6981.9193,
+    "train_samples": 21429,
+    "train_samples_per_second": 61.384,
+    "train_steps_per_second": 7.674
+}

egy_training_log.txt

@@ -296,3 +296,5 @@ INFO:root:Epoch 6.0: Train Loss = 0.1557, Eval Loss = 0.6072628498077393
 INFO:absl:Using default tokenizer.
 INFO:root:Epoch 7.0: Train Loss = 0.1308, Eval Loss = 0.6262651681900024
 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": 0.42251895839624176,
+    "eval_loss": 0.5335076451301575,
+    "eval_rouge1": 0.6607626579317042,
+    "eval_rouge2": 0.4163462041346451,
+    "eval_rougeL": 0.6589800378443118,
+    "eval_runtime": 57.7071,
+    "eval_samples": 5380,
+    "eval_samples_per_second": 93.229,
+    "eval_steps_per_second": 11.662,
+    "perplexity": 1.704902023958078
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 7.0,
+    "total_flos": 3.4826972570320896e+16,
+    "train_loss": 0.31742249039632825,
+    "train_runtime": 6981.9193,
+    "train_samples": 21429,
+    "train_samples_per_second": 61.384,
+    "train_steps_per_second": 7.674
+}

trainer_state.json

@@ -0,0 +1,184 @@
+{
+  "best_metric": 0.5335076451301575,
+  "best_model_checkpoint": "/home/iais_marenpielka/Bouthaina/res_nw_dj_03/checkpoint-5358",
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 18753,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.8902978897094727,
+      "learning_rate": 4.7947437829691034e-05,
+      "loss": 0.8829,
+      "step": 2679
+    },
+    {
+      "epoch": 1.0,
+      "eval_bleu": 0.40331287333676363,
+      "eval_loss": 0.5615599751472473,
+      "eval_rouge1": 0.6238739270954698,
+      "eval_rouge2": 0.3695879500967286,
+      "eval_rougeL": 0.6222094852350151,
+      "eval_runtime": 224.9918,
+      "eval_samples_per_second": 23.912,
+      "eval_steps_per_second": 2.991,
+      "step": 2679
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3653810024261475,
+      "learning_rate": 4.542388847023361e-05,
+      "loss": 0.4337,
+      "step": 5358
+    },
+    {
+      "epoch": 2.0,
+      "eval_bleu": 0.42251895839624176,
+      "eval_loss": 0.5335076451301575,
+      "eval_rouge1": 0.6607626579317042,
+      "eval_rouge2": 0.4163462041346451,
+      "eval_rougeL": 0.6589800378443118,
+      "eval_runtime": 163.0544,
+      "eval_samples_per_second": 32.995,
+      "eval_steps_per_second": 4.127,
+      "step": 5358
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.8291383981704712,
+      "learning_rate": 4.290033911077619e-05,
+      "loss": 0.2951,
+      "step": 8037
+    },
+    {
+      "epoch": 3.0,
+      "eval_bleu": 0.4308510587099172,
+      "eval_loss": 0.5472562909126282,
+      "eval_rouge1": 0.6702713862651033,
+      "eval_rouge2": 0.4320237158236569,
+      "eval_rougeL": 0.6688182372945262,
+      "eval_runtime": 162.9515,
+      "eval_samples_per_second": 33.016,
+      "eval_steps_per_second": 4.13,
+      "step": 8037
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 2.728090524673462,
+      "learning_rate": 4.0376789751318766e-05,
+      "loss": 0.2054,
+      "step": 10716
+    },
+    {
+      "epoch": 4.0,
+      "eval_bleu": 0.4276970379095206,
+      "eval_loss": 0.5758230686187744,
+      "eval_rouge1": 0.6763953460411876,
+      "eval_rouge2": 0.43998373859116113,
+      "eval_rougeL": 0.6748324672744146,
+      "eval_runtime": 40.4598,
+      "eval_samples_per_second": 132.971,
+      "eval_steps_per_second": 16.634,
+      "step": 10716
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 1.6252306699752808,
+      "learning_rate": 3.785324039186134e-05,
+      "loss": 0.1557,
+      "step": 13395
+    },
+    {
+      "epoch": 5.0,
+      "eval_bleu": 0.43093263057964937,
+      "eval_loss": 0.6072628498077393,
+      "eval_rouge1": 0.6728345471617314,
+      "eval_rouge2": 0.4398083411751642,
+      "eval_rougeL": 0.6716233259289648,
+      "eval_runtime": 76.1433,
+      "eval_samples_per_second": 70.656,
+      "eval_steps_per_second": 8.839,
+      "step": 13395
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 1.810899019241333,
+      "learning_rate": 3.532969103240392e-05,
+      "loss": 0.1308,
+      "step": 16074
+    },
+    {
+      "epoch": 6.0,
+      "eval_bleu": 0.4314626379100602,
+      "eval_loss": 0.6262651681900024,
+      "eval_rouge1": 0.6767217787616953,
+      "eval_rouge2": 0.44252126541905024,
+      "eval_rougeL": 0.675198694040748,
+      "eval_runtime": 42.2219,
+      "eval_samples_per_second": 127.422,
+      "eval_steps_per_second": 15.94,
+      "step": 16074
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 1.121442198753357,
+      "learning_rate": 3.28061416729465e-05,
+      "loss": 0.1183,
+      "step": 18753
+    },
+    {
+      "epoch": 7.0,
+      "eval_bleu": 0.42802429549225074,
+      "eval_loss": 0.6490646600723267,
+      "eval_rouge1": 0.6730562948760737,
+      "eval_rouge2": 0.43834978582096723,
+      "eval_rougeL": 0.6715407272843534,
+      "eval_runtime": 174.5563,
+      "eval_samples_per_second": 30.821,
+      "eval_steps_per_second": 3.855,
+      "step": 18753
+    },
+    {
+      "epoch": 7.0,
+      "step": 18753,
+      "total_flos": 3.4826972570320896e+16,
+      "train_loss": 0.31742249039632825,
+      "train_runtime": 6981.9193,
+      "train_samples_per_second": 61.384,
+      "train_steps_per_second": 7.674
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 53580,
+  "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": 3.4826972570320896e+16,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}