Pushing of the new best model checkpoint

README.md

@@ -4,20 +4,13 @@ license: mit
 base_model: dbmdz/bert-base-turkish-cased
 tags:
 - generated_from_trainer
-- turkish
 metrics:
 - accuracy
 - precision
 - recall
-- f1
 model-index:
 - name: turkish-sentiment
   results: []
-datasets:
-- winvoker/turkish-sentiment-analysis-dataset
-language:
-- tr
-pipeline_tag: text-classification
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -25,14 +18,14 @@ should probably proofread and complete it, then remove this comment. -->
 
 # turkish-sentiment
 
-This model is a fine-tuned version of [dbmdz/bert-base-turkish-cased](https://huggingface.co/dbmdz/bert-base-turkish-cased) on winvoker/turkish-sentiment-analysis-dataset dataset.
+This model is a fine-tuned version of [dbmdz/bert-base-turkish-cased](https://huggingface.co/dbmdz/bert-base-turkish-cased) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0957
-- Accuracy: 0.9657
-- F1 Macro: 0.9398
-- F1 Weighted: 0.9653
-- Precision: 0.9651
-- Recall: 0.9657
+- Loss: 0.1019
+- Accuracy: 0.9644
+- F1 Macro: 0.9380
+- F1 Weighted: 0.9640
+- Precision: 0.9638
+- Recall: 0.9644
 
 ## Model description
 
@@ -48,8 +41,6 @@ More information needed
 
 ## Training procedure
 
-trained on the full dataset for 1000 steps(apx 20 mins on a single gpu). 
-
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
@@ -59,18 +50,14 @@ 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: cosine
-- lr_scheduler_warmup_steps: 600
-- training_steps: 1000
+- lr_scheduler_warmup_steps: 400
+- training_steps: 200
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted | Precision | Recall |
 |:-------------:|:------:|:----:|:---------------:|:--------:|:--------:|:-----------:|:---------:|:------:|
-| 0.6035        | 0.0290 | 200  | 0.1912          | 0.9305   | 0.8739   | 0.9278      | 0.9274    | 0.9305 |
-| 0.1402        | 0.0581 | 400  | 0.1224          | 0.9561   | 0.9181   | 0.9539      | 0.9557    | 0.9561 |
-| 0.1245        | 0.0871 | 600  | 0.1227          | 0.9582   | 0.9217   | 0.9561      | 0.9578    | 0.9582 |
-| 0.1091        | 0.1162 | 800  | 0.1066          | 0.9632   | 0.9364   | 0.9630      | 0.9629    | 0.9632 |
-| 0.1078        | 0.1452 | 1000 | 0.0957          | 0.9657   | 0.9398   | 0.9653      | 0.9651    | 0.9657 |
+| 0.1044        | 0.0290 | 200  | 0.1019          | 0.9644   | 0.9380   | 0.9640      | 0.9638    | 0.9644 |
 
 
 ### Framework versions
@@ -78,4 +65,4 @@ The following hyperparameters were used during training:
 - Transformers 4.48.0.dev0
 - Pytorch 2.4.1+cu121
 - Datasets 3.1.0
-- Tokenizers 0.21.0
+- Tokenizers 0.21.0

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa7d1d1839e52dac02a60f0f98cca24387c7ff819c8afdde121240579fc92608
+oid sha256:7629d2ca8246c25d7c17ea0f835f96432adc9294907609bfdc3907cf3eb270b7
 size 442502140

tokenizer.json

@@ -1,7 +1,19 @@
 {
   "version": "1.0",
-  "truncation": null,
-  "padding": null,
+  "truncation": {
+    "direction": "Right",
+    "max_length": 512,
+    "strategy": "LongestFirst",
+    "stride": 0
+  },
+  "padding": {
+    "strategy": "BatchLongest",
+    "direction": "Right",
+    "pad_to_multiple_of": null,
+    "pad_id": 0,
+    "pad_type_id": 0,
+    "pad_token": "[PAD]"
+  },
   "added_tokens": [
     {
       "id": 0,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4e51841518efb60a1f07899ff0d8588a95e3157f16f69252edf7960f8e9bd5a
+oid sha256:819b4dee70186c832cdd15328f3bd5724c46c781eeaff0971aaf144b6481ccd0
 size 5432