Model save

README.md

@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4813
+- Loss: 0.4477
 
 ## Model description
 
@@ -36,21 +36,47 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-05
-- train_batch_size: 24
-- eval_batch_size: 24
+- train_batch_size: 32
+- eval_batch_size: 32
 - seed: 42
 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
-- num_epochs: 3
+- num_epochs: 50
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch | Step  | Validation Loss |
-|:-------------:|:-----:|:-----:|:---------------:|
-| 0.1713        | 1.0   | 11089 | 0.4982          |
-| 0.1659        | 2.0   | 22178 | 0.4854          |
-| 0.1676        | 3.0   | 33267 | 0.4813          |
+| Training Loss | Epoch | Step   | Validation Loss |
+|:-------------:|:-----:|:------:|:---------------:|
+| 0.1287        | 1.0   | 8317   | 0.5023          |
+| 0.1249        | 2.0   | 16634  | 0.4836          |
+| 0.1214        | 3.0   | 24951  | 0.4746          |
+| 0.1225        | 4.0   | 33268  | 0.4691          |
+| 0.1212        | 5.0   | 41585  | 0.4653          |
+| 0.1192        | 6.0   | 49902  | 0.4625          |
+| 0.1203        | 7.0   | 58219  | 0.4595          |
+| 0.1182        | 8.0   | 66536  | 0.4579          |
+| 0.117         | 9.0   | 74853  | 0.4565          |
+| 0.1184        | 10.0  | 83170  | 0.4543          |
+| 0.1165        | 11.0  | 91487  | 0.4539          |
+| 0.116         | 12.0  | 99804  | 0.4528          |
+| 0.117         | 13.0  | 108121 | 0.4515          |
+| 0.1161        | 14.0  | 116438 | 0.4512          |
+| 0.1149        | 15.0  | 124755 | 0.4510          |
+| 0.1157        | 16.0  | 133072 | 0.4503          |
+| 0.1162        | 17.0  | 141389 | 0.4497          |
+| 0.1155        | 18.0  | 149706 | 0.4497          |
+| 0.115         | 19.0  | 158023 | 0.4496          |
+| 0.1156        | 20.0  | 166340 | 0.4488          |
+| 0.1156        | 21.0  | 174657 | 0.4492          |
+| 0.114         | 22.0  | 182974 | 0.4483          |
+| 0.1122        | 23.0  | 191291 | 0.4488          |
+| 0.1134        | 24.0  | 199608 | 0.4480          |
+| 0.114         | 25.0  | 207925 | 0.4476          |
+| 0.1159        | 26.0  | 216242 | 0.4481          |
+| 0.115         | 27.0  | 224559 | 0.4475          |
+| 0.1145        | 28.0  | 232876 | 0.4475          |
+| 0.1141        | 29.0  | 241193 | 0.4477          |
 
 
 ### Framework versions

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b7cc961de0e24ebd647cbc1cf0b4ccb92c019b75bdba429e454fb91b80565d4a
+oid sha256:9f836ab9f3dc683b822a1b8706d59f59fdf4acefa7431eec9529ab86d1ba4966
 size 309974336

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 3.0,
+    "epoch": 29.0,
     "eval_loss": 0.48132315278053284,
     "eval_runtime": 378.7177,
     "eval_samples_per_second": 150.323,
     "eval_steps_per_second": 6.266,
     "perplexity": 1.6182141308134652,
-    "total_flos": 5.233249244912026e+16,
-    "train_loss": 0.17665502242223255,
-    "train_runtime": 3282.9562,
-    "train_samples_per_second": 243.185,
-    "train_steps_per_second": 10.133
+    "total_flos": 5.058807603414958e+17,
+    "train_loss": 0.11824046083851339,
+    "train_runtime": 25872.5126,
+    "train_samples_per_second": 514.295,
+    "train_steps_per_second": 16.073
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 5.233249244912026e+16,
-    "train_loss": 0.17665502242223255,
-    "train_runtime": 3282.9562,
-    "train_samples_per_second": 243.185,
-    "train_steps_per_second": 10.133
+    "epoch": 29.0,
+    "total_flos": 5.058807603414958e+17,
+    "train_loss": 0.11824046083851339,
+    "train_runtime": 25872.5126,
+    "train_samples_per_second": 514.295,
+    "train_steps_per_second": 16.073
 }