Training complete

README.md

@@ -1,7 +1,10 @@
 ---
+base_model: vinai/bartpho-syllable
 tags:
 - vietnamese-model
 - generated_from_trainer
+metrics:
+- sacrebleu
 model-index:
 - name: nlp-vietnamese
   results: []
@@ -12,9 +15,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # nlp-vietnamese
 
-This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+This model is a fine-tuned version of [vinai/bartpho-syllable](https://huggingface.co/vinai/bartpho-syllable) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.1229
+- Loss: 0.1109
+- Sacrebleu: 18.8039
 
 ## Model description
 
@@ -34,11 +38,11 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 2e-05
-- train_batch_size: 8
-- eval_batch_size: 8
+- train_batch_size: 12
+- eval_batch_size: 12
 - seed: 42
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 32
+- total_train_batch_size: 48
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 5
@@ -46,13 +50,13 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss |
-|:-------------:|:-----:|:----:|:---------------:|
-| 2.0814        | 1.0   | 125  | 1.7464          |
-| 1.3908        | 2.0   | 250  | 1.3544          |
-| 1.312         | 3.0   | 375  | 1.2367          |
-| 1.1243        | 4.0   | 500  | 1.1627          |
-| 1.0922        | 5.0   | 625  | 1.1229          |
+| Training Loss | Epoch | Step | Validation Loss | Sacrebleu |
+|:-------------:|:-----:|:----:|:---------------:|:---------:|
+| No log        | 0.992 | 62   | 0.7852          | 5.8526    |
+| No log        | 2.0   | 125  | 0.2320          | 12.5980   |
+| No log        | 2.992 | 187  | 0.1587          | 15.8119   |
+| No log        | 4.0   | 250  | 0.1222          | 17.9026   |
+| No log        | 4.96  | 310  | 0.1109          | 18.8039   |
 
 
 ### Framework versions

generation_config.json

@@ -1,5 +1,8 @@
 {
-  "_from_model_config": true,
-  "pad_token_id": 0,
+  "bos_token_id": 0,
+  "decoder_start_token_id": 2,
+  "eos_token_id": 2,
+  "forced_eos_token_id": 2,
+  "pad_token_id": 1,
   "transformers_version": "4.40.2"
 }