Model save

README.md

@@ -1,12 +1,7 @@
 ---
-base_model: openai/whisper-medium
-datasets:
-- Marcusxx/CngFSt10_5
-language:
-- ko
 license: apache-2.0
+base_model: openai/whisper-medium
 tags:
-- hf-asr-leaderboard
 - generated_from_trainer
 model-index:
 - name: CngFSt10_5_sec_model
@@ -18,10 +13,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # CngFSt10_5_sec_model
 
-This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Marcusxx/CngFSt10_5 dataset.
+This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0068
-- Cer: 22.2054
+- Loss: 0.0061
+- Cer: 15.8432
 
 ## Model description
 
@@ -46,7 +41,7 @@ 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
-- lr_scheduler_warmup_steps: 500
+- lr_scheduler_warmup_steps: 250
 - training_steps: 3000
 - mixed_precision_training: Native AMP
 
@@ -54,9 +49,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch   | Step | Validation Loss | Cer     |
 |:-------------:|:-------:|:----:|:---------------:|:-------:|
-| 0.0263        | 3.6765  | 1000 | 0.0404          | 52.6497 |
-| 0.0009        | 7.3529  | 2000 | 0.0106          | 20.7149 |
-| 0.0003        | 11.0294 | 3000 | 0.0068          | 22.2054 |
+| 0.0207        | 3.6765  | 1000 | 0.0354          | 36.5029 |
+| 0.0007        | 7.3529  | 2000 | 0.0076          | 15.6500 |
+| 0.0003        | 11.0294 | 3000 | 0.0061          | 15.8432 |
 
 
 ### Framework versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8ba792c8c1a71cf251bc81a4272c4ee598fb6e8b1eaf3499bbe2b0f219d497c0
+oid sha256:6e4e1c5a0bc50315b665bf84a3e8da41cba3b8f8d487e5bcf85eea4b9eeff18a
 size 3055544304