End of training

README.md

@@ -18,12 +18,12 @@ 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: 0.6761
-- Accuracy: 0.9226
-- 1-f1: 0.0
-- 1-recall: 0.0
-- 1-precision: 0.0
-- Balanced Acc: 0.5
+- Loss: 0.7020
+- Accuracy: 0.6533
+- 1-f1: 0.032
+- 1-recall: 0.0741
+- 1-precision: 0.0204
+- Balanced Acc: 0.3880
 
 ## Model description
 
@@ -48,18 +48,18 @@ 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
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Accuracy | 1-f1 | 1-recall | 1-precision | Balanced Acc |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|:----:|:--------:|:-----------:|:------------:|
-| 0.5666        | 1.0   | 88   | 0.6764          | 0.9226   | 0.0  | 0.0      | 0.0         | 0.5          |
-| 0.5654        | 2.0   | 176  | 0.6810          | 0.9226   | 0.0  | 0.0      | 0.0         | 0.5          |
-| 0.6907        | 3.0   | 264  | 0.6758          | 0.9226   | 0.0  | 0.0      | 0.0         | 0.5          |
-| 0.4527        | 4.0   | 352  | 0.6837          | 0.9226   | 0.0  | 0.0      | 0.0         | 0.5          |
-| 0.6271        | 5.0   | 440  | 0.6761          | 0.9226   | 0.0  | 0.0      | 0.0         | 0.5          |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy | 1-f1   | 1-recall | 1-precision | Balanced Acc |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:--------:|:-----------:|:------------:|
+| 0.5237        | 1.0   | 88   | 0.7195          | 0.1060   | 0.1475 | 1.0      | 0.0796      | 0.5155       |
+| 0.433         | 2.0   | 176  | 0.6923          | 0.5244   | 0.0778 | 0.2593   | 0.0458      | 0.4029       |
+| 0.6704        | 3.0   | 264  | 0.7048          | 0.1175   | 0.1492 | 1.0      | 0.0806      | 0.5217       |
+| 0.2894        | 4.0   | 352  | 0.7020          | 0.6533   | 0.032  | 0.0741   | 0.0204      | 0.3880       |
 
 
 ### 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 sha256:948c4c0ccad2a83fcb7f277605d98e4596121406cee7aa98b6b568671535e049
+oid sha256:5a2853fa0473f21c697afb1adab5d9159aba479a2b8c4b666c70eadc67bdfe76
 size 2303755328

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:04ab655a84a0c5d53b2568cef11f4fc3fb0fbed3d4e9cc746ee12ad4f0634b9f
+oid sha256:06a953771df378e332d38a93fe9cc5e6adacf7d3be0f5f27b2b12b91c8a0efe9
 size 5432