End of training

README.md

@@ -3,6 +3,8 @@ license: mit
 base_model: unicamp-dl/ptt5-large-t5-vocab
 tags:
 - generated_from_trainer
+datasets:
+- tiagoblima/qg_squad_v1_pt
 model-index:
 - name: t5_large-qg-aap
   results: []
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # t5_large-qg-aap
 
-This model is a fine-tuned version of [unicamp-dl/ptt5-large-t5-vocab](https://huggingface.co/unicamp-dl/ptt5-large-t5-vocab) on an unknown dataset.
+This model is a fine-tuned version of [unicamp-dl/ptt5-large-t5-vocab](https://huggingface.co/unicamp-dl/ptt5-large-t5-vocab) on the tiagoblima/qg_squad_v1_pt dataset.
 It achieves the following results on the evaluation set:
 - Loss: 5.5901
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 5.0,
-    "eval_loss": 7.420807361602783,
-    "eval_runtime": 587.2121,
+    "eval_loss": 5.590092658996582,
+    "eval_runtime": 526.161,
     "eval_samples": 8869,
-    "eval_samples_per_second": 15.104,
-    "eval_steps_per_second": 7.553,
-    "train_loss": 6.77048750395822,
-    "train_runtime": 16676.4775,
+    "eval_samples_per_second": 16.856,
+    "eval_steps_per_second": 4.215,
+    "train_loss": 5.410919136576133,
+    "train_runtime": 16658.5174,
     "train_samples": 51704,
-    "train_samples_per_second": 15.502,
-    "train_steps_per_second": 0.121
+    "train_samples_per_second": 15.519,
+    "train_steps_per_second": 0.243
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 5.0,
-    "eval_loss": 7.420807361602783,
-    "eval_runtime": 587.2121,
+    "eval_loss": 5.590092658996582,
+    "eval_runtime": 526.161,
     "eval_samples": 8869,
-    "eval_samples_per_second": 15.104,
-    "eval_steps_per_second": 7.553
+    "eval_samples_per_second": 16.856,
+    "eval_steps_per_second": 4.215
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 5.0,
-    "train_loss": 6.77048750395822,
-    "train_runtime": 16676.4775,
+    "train_loss": 5.410919136576133,
+    "train_runtime": 16658.5174,
     "train_samples": 51704,
-    "train_samples_per_second": 15.502,
-    "train_steps_per_second": 0.121
+    "train_samples_per_second": 15.519,
+    "train_steps_per_second": 0.243
 }

trainer_state.json

@@ -1,185 +1,305 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 5.590092658996582,
+  "best_model_checkpoint": "/temp/t5_large-qg-aap/checkpoint-4040",
   "epoch": 5.0,
   "eval_steps": 500,
-  "global_step": 2020,
+  "global_step": 4040,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.25,
-      "learning_rate": 0.0028514851485148514,
-      "loss": 8.3611,
+      "epoch": 0.12,
+      "learning_rate": 0.004876237623762376,
+      "loss": 8.2771,
       "step": 100
     },
     {
-      "epoch": 0.5,
-      "learning_rate": 0.002702970297029703,
-      "loss": 8.0607,
+      "epoch": 0.25,
+      "learning_rate": 0.004752475247524752,
+      "loss": 7.8175,
       "step": 200
     },
     {
-      "epoch": 0.74,
-      "learning_rate": 0.0025544554455445546,
-      "loss": 7.7784,
+      "epoch": 0.37,
+      "learning_rate": 0.004628712871287129,
+      "loss": 7.4476,
       "step": 300
     },
     {
-      "epoch": 0.99,
-      "learning_rate": 0.0024059405940594063,
-      "loss": 7.5413,
+      "epoch": 0.5,
+      "learning_rate": 0.0045049504950495055,
+      "loss": 7.0782,
       "step": 400
     },
+    {
+      "epoch": 0.62,
+      "learning_rate": 0.004381188118811882,
+      "loss": 6.755,
+      "step": 500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 0.004257425742574258,
+      "loss": 6.5292,
+      "step": 600
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 0.004133663366336634,
+      "loss": 6.3201,
+      "step": 700
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 0.00400990099009901,
+      "loss": 6.15,
+      "step": 800
+    },
     {
       "epoch": 1.0,
-      "eval_loss": 8.850217819213867,
-      "eval_runtime": 584.1219,
-      "eval_samples_per_second": 15.183,
-      "eval_steps_per_second": 7.593,
-      "step": 404
+      "eval_loss": 7.336104869842529,
+      "eval_runtime": 526.3079,
+      "eval_samples_per_second": 16.851,
+      "eval_steps_per_second": 4.214,
+      "step": 808
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.0038861386138613866,
+      "loss": 6.0156,
+      "step": 900
     },
     {
       "epoch": 1.24,
-      "learning_rate": 0.0022574257425742577,
-      "loss": 7.3312,
-      "step": 500
+      "learning_rate": 0.0037623762376237627,
+      "loss": 5.8846,
+      "step": 1000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 0.0036386138613861384,
+      "loss": 5.7566,
+      "step": 1100
     },
     {
       "epoch": 1.49,
-      "learning_rate": 0.0021089108910891086,
-      "loss": 7.1317,
-      "step": 600
+      "learning_rate": 0.0035148514851485145,
+      "loss": 5.6981,
+      "step": 1200
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 0.0033910891089108915,
+      "loss": 5.6151,
+      "step": 1300
     },
     {
       "epoch": 1.73,
-      "learning_rate": 0.0019603960396039604,
-      "loss": 6.9568,
-      "step": 700
+      "learning_rate": 0.0032673267326732676,
+      "loss": 5.4776,
+      "step": 1400
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 0.0031435643564356438,
+      "loss": 5.3873,
+      "step": 1500
     },
     {
       "epoch": 1.98,
-      "learning_rate": 0.0018118811881188118,
-      "loss": 6.7965,
-      "step": 800
+      "learning_rate": 0.00301980198019802,
+      "loss": 5.3335,
+      "step": 1600
     },
     {
       "epoch": 2.0,
-      "eval_loss": 8.118351936340332,
-      "eval_runtime": 586.5561,
-      "eval_samples_per_second": 15.12,
-      "eval_steps_per_second": 7.561,
-      "step": 808
+      "eval_loss": 6.409176349639893,
+      "eval_runtime": 525.7688,
+      "eval_samples_per_second": 16.869,
+      "eval_steps_per_second": 4.219,
+      "step": 1616
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 0.002896039603960396,
+      "loss": 5.2361,
+      "step": 1700
     },
     {
       "epoch": 2.23,
-      "learning_rate": 0.0016633663366336635,
-      "loss": 6.6621,
-      "step": 900
+      "learning_rate": 0.0027722772277227726,
+      "loss": 5.1746,
+      "step": 1800
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 0.0026485148514851487,
+      "loss": 5.1069,
+      "step": 1900
     },
     {
       "epoch": 2.48,
-      "learning_rate": 0.001514851485148515,
-      "loss": 6.5571,
-      "step": 1000
+      "learning_rate": 0.002524752475247525,
+      "loss": 5.07,
+      "step": 2000
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 0.002400990099009901,
+      "loss": 5.0201,
+      "step": 2100
     },
     {
       "epoch": 2.72,
-      "learning_rate": 0.0013663366336633665,
-      "loss": 6.4736,
-      "step": 1100
+      "learning_rate": 0.0022772277227722775,
+      "loss": 4.9693,
+      "step": 2200
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 0.0021534653465346536,
+      "loss": 4.9579,
+      "step": 2300
     },
     {
       "epoch": 2.97,
-      "learning_rate": 0.001217821782178218,
-      "loss": 6.3963,
-      "step": 1200
+      "learning_rate": 0.0020297029702970298,
+      "loss": 4.8807,
+      "step": 2400
     },
     {
       "epoch": 3.0,
-      "eval_loss": 7.695041179656982,
-      "eval_runtime": 586.9195,
-      "eval_samples_per_second": 15.111,
-      "eval_steps_per_second": 7.556,
-      "step": 1212
+      "eval_loss": 5.913175582885742,
+      "eval_runtime": 526.104,
+      "eval_samples_per_second": 16.858,
+      "eval_steps_per_second": 4.216,
+      "step": 2424
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 0.0019059405940594061,
+      "loss": 4.8449,
+      "step": 2500
     },
     {
       "epoch": 3.22,
-      "learning_rate": 0.0010693069306930692,
-      "loss": 6.3235,
-      "step": 1300
+      "learning_rate": 0.001782178217821782,
+      "loss": 4.7936,
+      "step": 2600
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 0.0016584158415841586,
+      "loss": 4.7996,
+      "step": 2700
     },
     {
       "epoch": 3.47,
-      "learning_rate": 0.0009207920792079207,
-      "loss": 6.283,
-      "step": 1400
+      "learning_rate": 0.0015346534653465347,
+      "loss": 4.7631,
+      "step": 2800
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 0.0014108910891089108,
+      "loss": 4.7178,
+      "step": 2900
     },
     {
       "epoch": 3.71,
-      "learning_rate": 0.0007722772277227723,
-      "loss": 6.2224,
-      "step": 1500
+      "learning_rate": 0.0012871287128712872,
+      "loss": 4.6978,
+      "step": 3000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 0.0011633663366336635,
+      "loss": 4.6717,
+      "step": 3100
     },
     {
       "epoch": 3.96,
-      "learning_rate": 0.0006237623762376238,
-      "loss": 6.1664,
-      "step": 1600
+      "learning_rate": 0.0010396039603960397,
+      "loss": 4.6492,
+      "step": 3200
     },
     {
       "epoch": 4.0,
-      "eval_loss": 7.485514163970947,
-      "eval_runtime": 587.006,
-      "eval_samples_per_second": 15.109,
-      "eval_steps_per_second": 7.555,
-      "step": 1616
+      "eval_loss": 5.6656341552734375,
+      "eval_runtime": 526.1669,
+      "eval_samples_per_second": 16.856,
+      "eval_steps_per_second": 4.215,
+      "step": 3232
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 0.0009158415841584158,
+      "loss": 4.6611,
+      "step": 3300
     },
     {
       "epoch": 4.21,
-      "learning_rate": 0.0004752475247524753,
-      "loss": 6.1536,
-      "step": 1700
+      "learning_rate": 0.0007920792079207921,
+      "loss": 4.6159,
+      "step": 3400
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 0.0006683168316831684,
+      "loss": 4.6386,
+      "step": 3500
     },
     {
       "epoch": 4.46,
-      "learning_rate": 0.00032673267326732675,
-      "loss": 6.1402,
-      "step": 1800
+      "learning_rate": 0.0005445544554455446,
+      "loss": 4.5997,
+      "step": 3600
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 0.00042079207920792084,
+      "loss": 4.5819,
+      "step": 3700
     },
     {
       "epoch": 4.7,
-      "learning_rate": 0.0001782178217821782,
-      "loss": 6.0973,
-      "step": 1900
+      "learning_rate": 0.000297029702970297,
+      "loss": 4.5785,
+      "step": 3800
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 0.00017326732673267329,
+      "loss": 4.588,
+      "step": 3900
     },
     {
       "epoch": 4.95,
-      "learning_rate": 2.9702970297029706e-05,
-      "loss": 6.1028,
-      "step": 2000
+      "learning_rate": 4.950495049504951e-05,
+      "loss": 4.591,
+      "step": 4000
     },
     {
       "epoch": 5.0,
-      "eval_loss": 7.420807361602783,
-      "eval_runtime": 587.2049,
-      "eval_samples_per_second": 15.104,
-      "eval_steps_per_second": 7.553,
-      "step": 2020
+      "eval_loss": 5.590092658996582,
+      "eval_runtime": 526.0558,
+      "eval_samples_per_second": 16.859,
+      "eval_steps_per_second": 4.216,
+      "step": 4040
     },
     {
       "epoch": 5.0,
-      "step": 2020,
+      "step": 4040,
       "total_flos": 4.4078196477394944e+17,
-      "train_loss": 6.77048750395822,
-      "train_runtime": 16676.4775,
-      "train_samples_per_second": 15.502,
-      "train_steps_per_second": 0.121
+      "train_loss": 5.410919136576133,
+      "train_runtime": 16658.5174,
+      "train_samples_per_second": 15.519,
+      "train_steps_per_second": 0.243
     }
   ],
   "logging_steps": 100,
-  "max_steps": 2020,
+  "max_steps": 4040,
   "num_train_epochs": 5,
   "save_steps": 500,
   "total_flos": 4.4078196477394944e+17,