Samuael/ethipic-sec2sec-tiginya

README.md

@@ -2,9 +2,6 @@
 library_name: transformers
 tags:
 - generated_from_trainer
-metrics:
-- wer
-- bleu
 model-index:
 - name: ethipic-sec2sec-tigre
   results: []
@@ -17,10 +14,15 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2052
-- Wer: 0.1880
-- Cer: 0.1223
-- Bleu: 76.9650
+- eval_loss: 0.1009
+- eval_wer: 0.0416
+- eval_cer: 0.0113
+- eval_bleu: 91.6015
+- eval_runtime: 30.3787
+- eval_samples_per_second: 9.842
+- eval_steps_per_second: 0.099
+- epoch: 4.0
+- step: 51000
 
 ## Model description
 
@@ -45,20 +47,9 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 5
+- num_epochs: 10
 - mixed_precision_training: Native AMP
 
-### Training results
-
-| Training Loss | Epoch | Step | Validation Loss | Wer    | Cer    | Bleu    |
-|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:-------:|
-| 0.1039        | 1.0   | 391  | 0.1998          | 0.1848 | 0.1126 | 76.8906 |
-| 0.1197        | 2.0   | 782  | 0.1984          | 0.1025 | 0.0368 | 83.6572 |
-| 0.0705        | 3.0   | 1173 | 0.2049          | 0.1312 | 0.0613 | 81.4318 |
-| 0.0735        | 4.0   | 1564 | 0.2007          | 0.1803 | 0.1081 | 77.3169 |
-| 0.0545        | 5.0   | 1955 | 0.2052          | 0.1880 | 0.1223 | 76.9650 |
-
-
 ### Framework versions
 
 - Transformers 4.44.2

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b9441d53be5bfca7778e5c971ec3a1122cf16972a56b639c005090aa75f68c3
+oid sha256:fce730e13d4366c1abc82c1bdfa9e541d103c184a0a813cff0a4e24c61d7d5d6
 size 240738368