Rafaelrosendo1/Large-v3

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [Rafaelrosendo1/whisper-large-pt](https://huggingface.co/Rafaelrosendo1/whisper-large-pt) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2592
-- Wer: 23.4359
+- Loss: 0.2627
+- Wer: 23.2194
 
 ## Model description
 
@@ -37,23 +37,24 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 2e-05
+- learning_rate: 1e-07
 - train_batch_size: 32
 - eval_batch_size: 8
 - seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 64
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- training_steps: 4000
+- lr_scheduler_warmup_steps: 1000
+- training_steps: 6000
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Wer     |
 |:-------------:|:-----:|:----:|:---------------:|:-------:|
-| 0.0906        | 1.19  | 1000 | 0.2326          | 24.8668 |
-| 0.0369        | 2.38  | 2000 | 0.2219          | 23.9078 |
-| 0.0136        | 3.57  | 3000 | 0.2426          | 22.2621 |
-| 0.0013        | 4.76  | 4000 | 0.2592          | 23.4359 |
+| 0.0038        | 4.76  | 2000 | 0.2440          | 22.8608 |
+| 0.0021        | 9.51  | 4000 | 0.2568          | 23.1314 |
+| 0.0023        | 14.27 | 6000 | 0.2627          | 23.2194 |
 
 
 ### Framework versions

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:31c7229f17ecad1f982d396d0b6976842a20697b9b1e498be4c4b0fc5fc985b0
+oid sha256:f0e77daef6e5b1ea689a3013d02265002e27c7cae1af169c90c6e36843ea593c
 size 6173656065

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:732f1257b93d1d4a9bebc119e93e69ad60f18d3b9188bde177ad82fd12926a29
+oid sha256:9738d9d96a4bcfc9a7feaec576bde30bb54710e10dc41b219ec274dd44e43832
 size 4664