End of training

README.md

@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/rembert](https://huggingface.co/google/rembert) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6902
+- Loss: 0.7024
 - Accuracy: 0.1263
 - 1-f1: 0.2242
 - 1-recall: 1.0
@@ -48,6 +48,7 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.06
 - num_epochs: 20
 - mixed_precision_training: Native AMP
 
@@ -55,9 +56,24 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy | 1-f1   | 1-recall | 1-precision | Balanced Acc |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:--------:|:-----------:|:------------:|
-| 0.6913        | 1.0   | 50   | 0.6902          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
-| 0.7327        | 2.0   | 100  | 0.6979          | 0.8737   | 0.0    | 0.0      | 0.0         | 0.5          |
-| 0.7579        | 3.0   | 150  | 0.6902          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.6697        | 1.0   | 50   | 0.6939          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.6672        | 2.0   | 100  | 0.7424          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.8553        | 3.0   | 150  | 0.8350          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.6019        | 4.0   | 200  | 0.7917          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.5024        | 5.0   | 250  | 0.6939          | 0.1414   | 0.2273 | 1.0      | 0.1282      | 0.5087       |
+| 0.6982        | 6.0   | 300  | 0.7261          | 0.1364   | 0.2262 | 1.0      | 0.1276      | 0.5058       |
+| 0.8583        | 7.0   | 350  | 0.7103          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
+| 0.5487        | 8.0   | 400  | 0.7149          | 0.1364   | 0.2262 | 1.0      | 0.1276      | 0.5058       |
+| 0.5309        | 9.0   | 450  | 0.7218          | 0.1313   | 0.2252 | 1.0      | 0.1269      | 0.5029       |
+| 0.5436        | 10.0  | 500  | 0.6868          | 0.1919   | 0.2381 | 1.0      | 0.1351      | 0.5376       |
+| 0.5617        | 11.0  | 550  | 0.6876          | 0.1919   | 0.2381 | 1.0      | 0.1351      | 0.5376       |
+| 0.8479        | 12.0  | 600  | 0.6812          | 0.1919   | 0.2381 | 1.0      | 0.1351      | 0.5376       |
+| 1.4176        | 13.0  | 650  | 0.6805          | 0.1818   | 0.2358 | 1.0      | 0.1337      | 0.5318       |
+| 0.3742        | 14.0  | 700  | 0.6920          | 0.1566   | 0.2304 | 1.0      | 0.1302      | 0.5173       |
+| 0.7671        | 15.0  | 750  | 0.6905          | 0.1465   | 0.2283 | 1.0      | 0.1289      | 0.5116       |
+| 0.994         | 16.0  | 800  | 0.6967          | 0.1313   | 0.2252 | 1.0      | 0.1269      | 0.5029       |
+| 0.3561        | 17.0  | 850  | 0.6962          | 0.1414   | 0.2273 | 1.0      | 0.1282      | 0.5087       |
+| 0.5991        | 18.0  | 900  | 0.7024          | 0.1263   | 0.2242 | 1.0      | 0.1263      | 0.5          |
 
 
 ### Framework versions

config.json

@@ -3,14 +3,14 @@
   "architectures": [
     "RemBertForSequenceClassification"
   ],
-  "attention_probs_dropout_prob": 0,
+  "attention_probs_dropout_prob": 0.2,
   "bos_token_id": 312,
   "classifier_dropout_prob": 0.1,
   "embedding_dropout_prob": 0,
   "embedding_size": 256,
   "eos_token_id": 313,
   "hidden_act": "gelu",
-  "hidden_dropout_prob": 0,
+  "hidden_dropout_prob": 0.2,
   "hidden_size": 1152,
   "initializer_range": 0.02,
   "input_embedding_size": 256,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:da0d2d0d22e2fbc9fa9a316210dd6b9f1a7989abfbc54878a93bf79c3ab3bd8d
+oid sha256:7a6331c4cfd52314358d78ab79ff8bb22d5a0cdd5c383bf8ca5288b8a253d60f
 size 2303755328

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:65f7fa90913bf0c7930f3712e85365f6fdb32a20694afa786d91ec1938e454a5
+oid sha256:badb20288754208f9d61f66e8c7d11560c2a60220427735ce14d6ab5fe50ceda
 size 5432