End of training

README.md

@@ -17,10 +17,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # test
 
-This model is a fine-tuned version of [meta-llama/Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Llama-3.2-1B-Instruct](https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct) on the wsc dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3492
-- Num Input Tokens Seen: 46944
+- Loss: 0.3459
+- Num Input Tokens Seen: 49376
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.34708982706069946,
-    "eval_runtime": 0.6975,
-    "eval_samples_per_second": 80.282,
-    "eval_steps_per_second": 20.071,
+    "eval_loss": 0.34589245915412903,
+    "eval_runtime": 0.8481,
+    "eval_samples_per_second": 66.033,
+    "eval_steps_per_second": 16.508,
     "num_input_tokens_seen": 49376,
-    "total_flos": 497055112495104.0,
-    "train_loss": 0.5283625726699829,
-    "train_runtime": 476.3079,
-    "train_samples_per_second": 1.046,
-    "train_steps_per_second": 0.262
+    "total_flos": 497127920369664.0,
+    "train_loss": 1.1438678817749024,
+    "train_runtime": 264.1495,
+    "train_samples_per_second": 1.885,
+    "train_steps_per_second": 0.473
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.34708982706069946,
-    "eval_runtime": 0.6975,
-    "eval_samples_per_second": 80.282,
-    "eval_steps_per_second": 20.071,
+    "eval_loss": 0.34589245915412903,
+    "eval_runtime": 0.8481,
+    "eval_samples_per_second": 66.033,
+    "eval_steps_per_second": 16.508,
     "num_input_tokens_seen": 49376
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "num_input_tokens_seen": 49376,
-    "total_flos": 497055112495104.0,
-    "train_loss": 0.5283625726699829,
-    "train_runtime": 476.3079,
-    "train_samples_per_second": 1.046,
-    "train_steps_per_second": 0.262
+    "total_flos": 497127920369664.0,
+    "train_loss": 1.1438678817749024,
+    "train_runtime": 264.1495,
+    "train_samples_per_second": 1.885,
+    "train_steps_per_second": 0.473
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 84,
-  "best_metric": 0.34708982706069946,
-  "best_model_checkpoint": "saves/test/checkpoint-84",
+  "best_global_step": 112,
+  "best_metric": 0.34589245915412903,
+  "best_model_checkpoint": "saves/test/checkpoint-112",
   "epoch": 1.0,
   "eval_steps": 7,
   "global_step": 125,
@@ -11,354 +11,354 @@
   "log_history": [
     {
       "epoch": 0.04,
-      "grad_norm": 301.9017333984375,
+      "grad_norm": 556.8270263671875,
       "learning_rate": 1.5384615384615387e-05,
-      "loss": 1.0409,
+      "loss": 10.9709,
       "num_input_tokens_seen": 2144,
       "step": 5
     },
     {
       "epoch": 0.056,
-      "eval_loss": 0.3513108789920807,
-      "eval_runtime": 0.6546,
-      "eval_samples_per_second": 85.55,
-      "eval_steps_per_second": 21.387,
+      "eval_loss": 6.5227251052856445,
+      "eval_runtime": 0.7016,
+      "eval_samples_per_second": 79.812,
+      "eval_steps_per_second": 19.953,
       "num_input_tokens_seen": 2880,
       "step": 7
     },
     {
       "epoch": 0.08,
-      "grad_norm": 37.0427360534668,
+      "grad_norm": 166.4779815673828,
       "learning_rate": 3.461538461538462e-05,
-      "loss": 0.4086,
+      "loss": 6.4075,
       "num_input_tokens_seen": 4128,
       "step": 10
     },
     {
       "epoch": 0.112,
-      "eval_loss": 1.1120651960372925,
-      "eval_runtime": 0.6513,
-      "eval_samples_per_second": 85.976,
-      "eval_steps_per_second": 21.494,
+      "eval_loss": 1.382468581199646,
+      "eval_runtime": 0.7422,
+      "eval_samples_per_second": 75.454,
+      "eval_steps_per_second": 18.864,
       "num_input_tokens_seen": 5920,
       "step": 14
     },
     {
       "epoch": 0.12,
-      "grad_norm": 14.121573448181152,
+      "grad_norm": 137.21327209472656,
       "learning_rate": 4.999016565957633e-05,
-      "loss": 0.8807,
+      "loss": 2.5338,
       "num_input_tokens_seen": 6240,
       "step": 15
     },
     {
       "epoch": 0.16,
-      "grad_norm": 6.5915303230285645,
+      "grad_norm": 15.133822441101074,
       "learning_rate": 4.96467754629559e-05,
-      "loss": 0.9267,
+      "loss": 0.5326,
       "num_input_tokens_seen": 8096,
       "step": 20
     },
     {
       "epoch": 0.168,
-      "eval_loss": 0.35109928250312805,
-      "eval_runtime": 0.7163,
-      "eval_samples_per_second": 78.179,
-      "eval_steps_per_second": 19.545,
+      "eval_loss": 0.4987373352050781,
+      "eval_runtime": 0.7195,
+      "eval_samples_per_second": 77.827,
+      "eval_steps_per_second": 19.457,
       "num_input_tokens_seen": 8416,
       "step": 21
     },
     {
       "epoch": 0.2,
-      "grad_norm": 19.82288360595703,
+      "grad_norm": 18.45602798461914,
       "learning_rate": 4.881937806807241e-05,
-      "loss": 0.7142,
+      "loss": 0.4144,
       "num_input_tokens_seen": 10112,
       "step": 25
     },
     {
       "epoch": 0.224,
-      "eval_loss": 0.38589543104171753,
-      "eval_runtime": 0.7026,
-      "eval_samples_per_second": 79.699,
-      "eval_steps_per_second": 19.925,
+      "eval_loss": 0.4531269073486328,
+      "eval_runtime": 0.7843,
+      "eval_samples_per_second": 71.406,
+      "eval_steps_per_second": 17.851,
       "num_input_tokens_seen": 11264,
       "step": 28
     },
     {
       "epoch": 0.24,
-      "grad_norm": 14.33414363861084,
+      "grad_norm": 42.76151657104492,
       "learning_rate": 4.752422169756048e-05,
-      "loss": 0.385,
+      "loss": 0.4563,
       "num_input_tokens_seen": 12032,
       "step": 30
     },
     {
       "epoch": 0.28,
-      "grad_norm": 21.1870174407959,
+      "grad_norm": 4.994872570037842,
       "learning_rate": 4.5786740307563636e-05,
-      "loss": 0.5983,
+      "loss": 0.4802,
       "num_input_tokens_seen": 13824,
       "step": 35
     },
     {
       "epoch": 0.28,
-      "eval_loss": 0.658510684967041,
-      "eval_runtime": 0.6682,
-      "eval_samples_per_second": 83.804,
-      "eval_steps_per_second": 20.951,
+      "eval_loss": 0.36931881308555603,
+      "eval_runtime": 0.8719,
+      "eval_samples_per_second": 64.225,
+      "eval_steps_per_second": 16.056,
       "num_input_tokens_seen": 13824,
       "step": 35
     },
     {
       "epoch": 0.32,
-      "grad_norm": 6.050720691680908,
+      "grad_norm": 8.647699356079102,
       "learning_rate": 4.364105412207914e-05,
-      "loss": 0.394,
+      "loss": 0.3809,
       "num_input_tokens_seen": 15840,
       "step": 40
     },
     {
       "epoch": 0.336,
-      "eval_loss": 0.41257768869400024,
-      "eval_runtime": 0.6969,
-      "eval_samples_per_second": 80.356,
-      "eval_steps_per_second": 20.089,
+      "eval_loss": 0.3872639238834381,
+      "eval_runtime": 0.7387,
+      "eval_samples_per_second": 75.812,
+      "eval_steps_per_second": 18.953,
       "num_input_tokens_seen": 16672,
       "step": 42
     },
     {
       "epoch": 0.36,
-      "grad_norm": 29.777023315429688,
+      "grad_norm": 12.718330383300781,
       "learning_rate": 4.1129299588552193e-05,
-      "loss": 0.4533,
+      "loss": 0.3844,
       "num_input_tokens_seen": 17920,
       "step": 45
     },
     {
       "epoch": 0.392,
-      "eval_loss": 1.1762187480926514,
-      "eval_runtime": 0.6684,
-      "eval_samples_per_second": 83.778,
-      "eval_steps_per_second": 20.945,
+      "eval_loss": 0.3777945637702942,
+      "eval_runtime": 0.7665,
+      "eval_samples_per_second": 73.06,
+      "eval_steps_per_second": 18.265,
       "num_input_tokens_seen": 19296,
       "step": 49
     },
     {
       "epoch": 0.4,
-      "grad_norm": 0.5394620895385742,
+      "grad_norm": 12.110234260559082,
       "learning_rate": 3.830080191288342e-05,
-      "loss": 0.1866,
+      "loss": 0.2817,
       "num_input_tokens_seen": 19712,
       "step": 50
     },
     {
       "epoch": 0.44,
-      "grad_norm": 19.856895446777344,
+      "grad_norm": 10.657136917114258,
       "learning_rate": 3.521110642339991e-05,
-      "loss": 1.3512,
+      "loss": 0.3831,
       "num_input_tokens_seen": 21952,
       "step": 55
     },
     {
       "epoch": 0.448,
-      "eval_loss": 0.806473433971405,
-      "eval_runtime": 0.6813,
-      "eval_samples_per_second": 82.196,
-      "eval_steps_per_second": 20.549,
+      "eval_loss": 0.4436803460121155,
+      "eval_runtime": 0.7702,
+      "eval_samples_per_second": 72.71,
+      "eval_steps_per_second": 18.178,
       "num_input_tokens_seen": 22432,
       "step": 56
     },
     {
       "epoch": 0.48,
-      "grad_norm": 3.317122220993042,
+      "grad_norm": 10.894862174987793,
       "learning_rate": 3.1920887785621235e-05,
-      "loss": 0.7948,
+      "loss": 0.5576,
       "num_input_tokens_seen": 24160,
       "step": 60
     },
     {
       "epoch": 0.504,
-      "eval_loss": 1.0268325805664062,
-      "eval_runtime": 0.679,
-      "eval_samples_per_second": 82.475,
-      "eval_steps_per_second": 20.619,
+      "eval_loss": 0.35032057762145996,
+      "eval_runtime": 0.9256,
+      "eval_samples_per_second": 60.502,
+      "eval_steps_per_second": 15.125,
       "num_input_tokens_seen": 25504,
       "step": 63
     },
     {
       "epoch": 0.52,
-      "grad_norm": 29.714481353759766,
+      "grad_norm": 7.415125370025635,
       "learning_rate": 2.849475848838749e-05,
-      "loss": 0.6941,
+      "loss": 0.4013,
       "num_input_tokens_seen": 26112,
       "step": 65
     },
     {
       "epoch": 0.56,
-      "grad_norm": 16.07105827331543,
+      "grad_norm": 9.572220802307129,
       "learning_rate": 2.5e-05,
-      "loss": 0.3463,
+      "loss": 0.3242,
       "num_input_tokens_seen": 28064,
       "step": 70
     },
     {
       "epoch": 0.56,
-      "eval_loss": 0.35280704498291016,
-      "eval_runtime": 0.6586,
-      "eval_samples_per_second": 85.032,
-      "eval_steps_per_second": 21.258,
+      "eval_loss": 0.37164703011512756,
+      "eval_runtime": 0.816,
+      "eval_samples_per_second": 68.627,
+      "eval_steps_per_second": 17.157,
       "num_input_tokens_seen": 28064,
       "step": 70
     },
     {
       "epoch": 0.6,
-      "grad_norm": 10.748852729797363,
+      "grad_norm": 11.036535263061523,
       "learning_rate": 2.1505241511612522e-05,
-      "loss": 0.3652,
+      "loss": 0.3963,
       "num_input_tokens_seen": 29824,
       "step": 75
     },
     {
       "epoch": 0.616,
-      "eval_loss": 0.3505268096923828,
-      "eval_runtime": 0.6898,
-      "eval_samples_per_second": 81.188,
-      "eval_steps_per_second": 20.297,
+      "eval_loss": 0.3748786747455597,
+      "eval_runtime": 0.7992,
+      "eval_samples_per_second": 70.066,
+      "eval_steps_per_second": 17.516,
       "num_input_tokens_seen": 30720,
       "step": 77
     },
     {
       "epoch": 0.64,
-      "grad_norm": 8.937973976135254,
+      "grad_norm": 2.2476918697357178,
       "learning_rate": 1.8079112214378768e-05,
-      "loss": 0.3476,
+      "loss": 0.3946,
       "num_input_tokens_seen": 31904,
       "step": 80
     },
     {
       "epoch": 0.672,
-      "eval_loss": 0.34708982706069946,
-      "eval_runtime": 0.7099,
-      "eval_samples_per_second": 78.887,
-      "eval_steps_per_second": 19.722,
+      "eval_loss": 0.3603578209877014,
+      "eval_runtime": 0.7982,
+      "eval_samples_per_second": 70.16,
+      "eval_steps_per_second": 17.54,
       "num_input_tokens_seen": 33504,
       "step": 84
     },
     {
       "epoch": 0.68,
-      "grad_norm": 9.469837188720703,
+      "grad_norm": 2.605731248855591,
       "learning_rate": 1.4788893576600099e-05,
-      "loss": 0.3491,
+      "loss": 0.3496,
       "num_input_tokens_seen": 33984,
       "step": 85
     },
     {
       "epoch": 0.72,
-      "grad_norm": 4.574488639831543,
+      "grad_norm": 2.532665967941284,
       "learning_rate": 1.1699198087116589e-05,
-      "loss": 0.3395,
+      "loss": 0.337,
       "num_input_tokens_seen": 35776,
       "step": 90
     },
     {
       "epoch": 0.728,
-      "eval_loss": 0.3647787868976593,
-      "eval_runtime": 0.7422,
-      "eval_samples_per_second": 75.449,
-      "eval_steps_per_second": 18.862,
+      "eval_loss": 0.35710158944129944,
+      "eval_runtime": 0.8602,
+      "eval_samples_per_second": 65.102,
+      "eval_steps_per_second": 16.276,
       "num_input_tokens_seen": 36128,
       "step": 91
     },
     {
       "epoch": 0.76,
-      "grad_norm": 4.598474502563477,
+      "grad_norm": 2.5953240394592285,
       "learning_rate": 8.870700411447816e-06,
-      "loss": 0.4569,
+      "loss": 0.4315,
       "num_input_tokens_seen": 37472,
       "step": 95
     },
     {
       "epoch": 0.784,
-      "eval_loss": 0.3610990643501282,
-      "eval_runtime": 0.683,
-      "eval_samples_per_second": 81.991,
-      "eval_steps_per_second": 20.498,
+      "eval_loss": 0.3520326614379883,
+      "eval_runtime": 0.7911,
+      "eval_samples_per_second": 70.792,
+      "eval_steps_per_second": 17.698,
       "num_input_tokens_seen": 38592,
       "step": 98
     },
     {
       "epoch": 0.8,
-      "grad_norm": 3.7893221378326416,
+      "grad_norm": 2.179095506668091,
       "learning_rate": 6.358945877920861e-06,
-      "loss": 0.3916,
+      "loss": 0.38,
       "num_input_tokens_seen": 39328,
       "step": 100
     },
     {
       "epoch": 0.84,
-      "grad_norm": 1.4534764289855957,
+      "grad_norm": 5.5090203285217285,
       "learning_rate": 4.213259692436367e-06,
-      "loss": 0.3191,
+      "loss": 0.371,
       "num_input_tokens_seen": 41280,
       "step": 105
     },
     {
       "epoch": 0.84,
-      "eval_loss": 0.43045976758003235,
-      "eval_runtime": 0.7093,
-      "eval_samples_per_second": 78.948,
-      "eval_steps_per_second": 19.737,
+      "eval_loss": 0.34758228063583374,
+      "eval_runtime": 0.8072,
+      "eval_samples_per_second": 69.373,
+      "eval_steps_per_second": 17.343,
       "num_input_tokens_seen": 41280,
       "step": 105
     },
     {
       "epoch": 0.88,
-      "grad_norm": 9.676743507385254,
+      "grad_norm": 1.8210620880126953,
       "learning_rate": 2.475778302439524e-06,
-      "loss": 0.3951,
+      "loss": 0.364,
       "num_input_tokens_seen": 43552,
       "step": 110
     },
     {
       "epoch": 0.896,
-      "eval_loss": 0.4486384391784668,
-      "eval_runtime": 0.6661,
-      "eval_samples_per_second": 84.076,
-      "eval_steps_per_second": 21.019,
+      "eval_loss": 0.34589245915412903,
+      "eval_runtime": 0.822,
+      "eval_samples_per_second": 68.123,
+      "eval_steps_per_second": 17.031,
       "num_input_tokens_seen": 44160,
       "step": 112
     },
     {
       "epoch": 0.92,
-      "grad_norm": 0.255204439163208,
+      "grad_norm": 5.644977569580078,
       "learning_rate": 1.180621931927592e-06,
-      "loss": 0.3107,
+      "loss": 0.3554,
       "num_input_tokens_seen": 45216,
       "step": 115
     },
     {
       "epoch": 0.952,
-      "eval_loss": 0.44127246737480164,
-      "eval_runtime": 0.6928,
-      "eval_samples_per_second": 80.827,
-      "eval_steps_per_second": 20.207,
+      "eval_loss": 0.3492301404476166,
+      "eval_runtime": 0.8757,
+      "eval_samples_per_second": 63.945,
+      "eval_steps_per_second": 15.986,
       "num_input_tokens_seen": 46944,
       "step": 119
     },
     {
       "epoch": 0.96,
-      "grad_norm": 9.6677827835083,
+      "grad_norm": 1.6824895143508911,
       "learning_rate": 3.5322453704410286e-07,
-      "loss": 0.2809,
+      "loss": 0.3494,
       "num_input_tokens_seen": 47360,
       "step": 120
     },
     {
       "epoch": 1.0,
-      "grad_norm": 9.322759628295898,
+      "grad_norm": 10.473832130432129,
       "learning_rate": 9.834340423678368e-09,
-      "loss": 0.4785,
+      "loss": 0.3588,
       "num_input_tokens_seen": 49376,
       "step": 125
     },
@@ -366,11 +366,11 @@
       "epoch": 1.0,
       "num_input_tokens_seen": 49376,
       "step": 125,
-      "total_flos": 497055112495104.0,
-      "train_loss": 0.5283625726699829,
-      "train_runtime": 476.3079,
-      "train_samples_per_second": 1.046,
-      "train_steps_per_second": 0.262
+      "total_flos": 497127920369664.0,
+      "train_loss": 1.1438678817749024,
+      "train_runtime": 264.1495,
+      "train_samples_per_second": 1.885,
+      "train_steps_per_second": 0.473
     }
   ],
   "logging_steps": 5,
@@ -390,7 +390,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 497055112495104.0,
+  "total_flos": 497127920369664.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null