Model save
Browse files
README.md
CHANGED
|
@@ -3,23 +3,11 @@ license: apache-2.0
|
|
| 3 |
base_model: openai/whisper-base
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
-
datasets:
|
| 7 |
-
- razhan/common_voice_ckb_16
|
| 8 |
metrics:
|
| 9 |
- wer
|
| 10 |
model-index:
|
| 11 |
- name: whisper-base-ckb
|
| 12 |
-
results:
|
| 13 |
-
- task:
|
| 14 |
-
name: Automatic Speech Recognition
|
| 15 |
-
type: automatic-speech-recognition
|
| 16 |
-
dataset:
|
| 17 |
-
name: razhan/common_voice_ckb_16
|
| 18 |
-
type: razhan/common_voice_ckb_16
|
| 19 |
-
metrics:
|
| 20 |
-
- name: Wer
|
| 21 |
-
type: wer
|
| 22 |
-
value: 0.2917510463075469
|
| 23 |
---
|
| 24 |
|
| 25 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -27,10 +15,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 27 |
|
| 28 |
# whisper-base-ckb
|
| 29 |
|
| 30 |
-
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on
|
| 31 |
It achieves the following results on the evaluation set:
|
| 32 |
-
- Loss: 0.
|
| 33 |
-
- Wer: 0.
|
| 34 |
|
| 35 |
## Model description
|
| 36 |
|
|
@@ -60,7 +48,7 @@ The following hyperparameters were used during training:
|
|
| 60 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 61 |
- lr_scheduler_type: linear
|
| 62 |
- lr_scheduler_warmup_steps: 200
|
| 63 |
-
- training_steps:
|
| 64 |
- mixed_precision_training: Native AMP
|
| 65 |
|
| 66 |
### Training results
|
|
@@ -77,6 +65,19 @@ The following hyperparameters were used during training:
|
|
| 77 |
| 0.1015 | 8.7 | 800 | 0.1489 | 0.3059 |
|
| 78 |
| 0.0968 | 9.78 | 900 | 0.1440 | 0.2954 |
|
| 79 |
| 0.0939 | 10.87 | 1000 | 0.1420 | 0.2918 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
|
| 82 |
### Framework versions
|
|
|
|
| 3 |
base_model: openai/whisper-base
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
|
|
|
| 6 |
metrics:
|
| 7 |
- wer
|
| 8 |
model-index:
|
| 9 |
- name: whisper-base-ckb
|
| 10 |
+
results: []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
---
|
| 12 |
|
| 13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 15 |
|
| 16 |
# whisper-base-ckb
|
| 17 |
|
| 18 |
+
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0641
|
| 21 |
+
- Wer: 0.1262
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: linear
|
| 50 |
- lr_scheduler_warmup_steps: 200
|
| 51 |
+
- training_steps: 2300
|
| 52 |
- mixed_precision_training: Native AMP
|
| 53 |
|
| 54 |
### Training results
|
|
|
|
| 65 |
| 0.1015 | 8.7 | 800 | 0.1489 | 0.3059 |
|
| 66 |
| 0.0968 | 9.78 | 900 | 0.1440 | 0.2954 |
|
| 67 |
| 0.0939 | 10.87 | 1000 | 0.1420 | 0.2918 |
|
| 68 |
+
| 0.0919 | 11.96 | 1100 | 0.1315 | 0.2742 |
|
| 69 |
+
| 0.0839 | 13.04 | 1200 | 0.1217 | 0.2597 |
|
| 70 |
+
| 0.0713 | 14.13 | 1300 | 0.1132 | 0.2371 |
|
| 71 |
+
| 0.0687 | 15.22 | 1400 | 0.1091 | 0.2372 |
|
| 72 |
+
| 0.0647 | 16.3 | 1500 | 0.1022 | 0.2173 |
|
| 73 |
+
| 0.059 | 17.39 | 1600 | 0.0967 | 0.2043 |
|
| 74 |
+
| 0.0539 | 18.48 | 1700 | 0.0897 | 0.1929 |
|
| 75 |
+
| 0.0518 | 19.57 | 1800 | 0.0827 | 0.1718 |
|
| 76 |
+
| 0.0495 | 20.65 | 1900 | 0.0787 | 0.1667 |
|
| 77 |
+
| 0.0444 | 21.74 | 2000 | 0.0718 | 0.1469 |
|
| 78 |
+
| 0.0392 | 22.83 | 2100 | 0.0671 | 0.1368 |
|
| 79 |
+
| 0.0335 | 23.91 | 2200 | 0.0645 | 0.1263 |
|
| 80 |
+
| 0.0292 | 25.0 | 2300 | 0.0641 | 0.1262 |
|
| 81 |
|
| 82 |
|
| 83 |
### Framework versions
|