End of training

README.md

@@ -3,10 +3,12 @@ license: mit
 base_model: microsoft/git-base
 tags:
 - generated_from_trainer
+metrics:
+- wer
+- rouge
 model-index:
 - name: git-base-env
   results: []
-pipeline_tag: image-to-text
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,8 +18,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/git-base](https://huggingface.co/microsoft/git-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 4.0922
-- Wer Score: 16.2039
+- Loss: 8.4059
+- Wer: 59.0853
+- Rouge1: 1.86
+- Rouge2: 0.57
+- Rougel: 1.63
+- Rougelsum: 1.63
 
 ## Model description
 
@@ -36,21 +42,23 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 5e-05
+- learning_rate: 5e-07
 - train_batch_size: 1
 - eval_batch_size: 1
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 1
+- num_epochs: 3
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Wer Score |
-|:-------------:|:-----:|:----:|:---------------:|:---------:|
-| 8.6797        | 0.3   | 30   | 6.9086          | 46.3882   |
-| 6.0221        | 0.61  | 60   | 5.0189          | 17.6447   |
-| 4.5819        | 0.91  | 90   | 4.0922          | 16.2039   |
+| Training Loss | Epoch | Step | Validation Loss | Wer     | Rouge1 | Rouge2 | Rougel | Rougelsum |
+|:-------------:|:-----:|:----:|:---------------:|:-------:|:------:|:------:|:------:|:---------:|
+| 8.6175        | 0.6   | 30   | 8.4884          | 59.0612 | 1.84   | 0.5    | 1.56   | 1.56      |
+| 8.5757        | 1.2   | 60   | 8.4512          | 58.9443 | 1.81   | 0.53   | 1.57   | 1.57      |
+| 8.541         | 1.8   | 90   | 8.4258          | 59.2653 | 1.86   | 0.56   | 1.62   | 1.62      |
+| 8.5196        | 2.4   | 120  | 8.4114          | 58.9926 | 1.84   | 0.57   | 1.62   | 1.62      |
+| 8.5091        | 3.0   | 150  | 8.4059          | 59.0853 | 1.86   | 0.57   | 1.63   | 1.63      |
 
 
 ### Framework versions
@@ -58,4 +66,4 @@ The following hyperparameters were used during training:
 - Transformers 4.34.0
 - Pytorch 2.0.1+cu118
 - Datasets 2.14.5
-- Tokenizers 0.14.0
+- Tokenizers 0.14.1

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa621223a81c9e914e3e1945fc31fe7d52ca722bdcaa5165c1e6eba404911f42
+oid sha256:d800d6d8cd07558d0d467f48d17ae16cde75faf67dabc59c77bcbf2ad206d512
 size 706584273

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3679bee832094f58e650803df21786c544b30d2a797fd5a47d537de6e023e9c1
+oid sha256:911607e0512ff0a1471e3a02ab86e58868a12ada2fdd3f3a320d088d33aea6b8
 size 4027