Model save

README.md

@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7132
+- Loss: 0.5317
 
 ## Model description
 
@@ -34,24 +34,30 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-05
-- train_batch_size: 4
-- eval_batch_size: 8
+- learning_rate: 0.0001
+- train_batch_size: 16
+- eval_batch_size: 16
 - seed: 42
 - gradient_accumulation_steps: 16
-- total_train_batch_size: 64
+- total_train_batch_size: 256
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 30
-- num_epochs: 3
+- lr_scheduler_warmup_steps: 100
+- num_epochs: 9
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.8502        | 0.99  | 56   | 0.8263          |
-| 0.7905        | 1.98  | 112  | 0.7338          |
-| 0.701         | 2.98  | 168  | 0.7132          |
+| 3.9039        | 0.99  | 14   | 2.4214          |
+| 0.944         | 1.98  | 28   | 0.8868          |
+| 0.7808        | 2.97  | 42   | 0.7562          |
+| 0.7087        | 3.96  | 56   | 0.7181          |
+| 0.6645        | 4.96  | 70   | 0.6481          |
+| 0.6183        | 5.95  | 84   | 0.6046          |
+| 0.5837        | 6.94  | 98   | 0.5691          |
+| 0.556         | 8.0   | 113  | 0.5426          |
+| 0.5493        | 8.92  | 126  | 0.5317          |
 
 
 ### Framework versions

config.json

@@ -71,5 +71,5 @@
   "torch_dtype": "float32",
   "transformers_version": "4.34.1",
   "use_cache": true,
-  "vocab_size": 56506
+  "vocab_size": 75113
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d98c48737eae6142ef6dd51c167bda2b56f3a6ebea3a9fbb1b43c5eb7833b091
-size 577168541
+oid sha256:d3d70cf8712689ffe77aa29b031b3d23a212a6ea576c9643fe938d8c51160534
+size 634403677

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:460d629b245e065bcc4d5e06bda958509cd2445c46a272c2715f32878f5e9dba
-size 4091
+oid sha256:3d30de82f7ff22c5fd6e1fb92863b1356f834ccd0ecbf28fb2220cfbb752c37f
+size 4155