Commit
•
42761e8
1
Parent(s):
6542d42
update model card README.md
Browse files
README.md
CHANGED
|
@@ -7,9 +7,24 @@ tags:
|
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- mozilla-foundation/common_voice_11_0
|
|
|
|
|
|
|
| 10 |
model-index:
|
| 11 |
- name: Whisper Large v2 Hindi
|
| 12 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
---
|
| 14 |
|
| 15 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -18,6 +33,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
# Whisper Large v2 Hindi
|
| 19 |
|
| 20 |
This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on the Common Voice 11.0 dataset.
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -37,17 +55,28 @@ More information needed
|
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
- learning_rate: 1e-05
|
| 40 |
-
- train_batch_size:
|
| 41 |
-
- eval_batch_size:
|
| 42 |
- seed: 42
|
| 43 |
- gradient_accumulation_steps: 2
|
| 44 |
-
- total_train_batch_size:
|
| 45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
- lr_scheduler_warmup_steps: 50
|
| 48 |
-
- training_steps:
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
### Framework versions
|
| 52 |
|
| 53 |
- Transformers 4.26.0.dev0
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- mozilla-foundation/common_voice_11_0
|
| 10 |
+
metrics:
|
| 11 |
+
- wer
|
| 12 |
model-index:
|
| 13 |
- name: Whisper Large v2 Hindi
|
| 14 |
+
results:
|
| 15 |
+
- task:
|
| 16 |
+
name: Automatic Speech Recognition
|
| 17 |
+
type: automatic-speech-recognition
|
| 18 |
+
dataset:
|
| 19 |
+
name: Common Voice 11.0
|
| 20 |
+
type: mozilla-foundation/common_voice_11_0
|
| 21 |
+
config: hi
|
| 22 |
+
split: test
|
| 23 |
+
args: hi
|
| 24 |
+
metrics:
|
| 25 |
+
- name: Wer
|
| 26 |
+
type: wer
|
| 27 |
+
value: 12.424051358477588
|
| 28 |
---
|
| 29 |
|
| 30 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 33 |
# Whisper Large v2 Hindi
|
| 34 |
|
| 35 |
This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on the Common Voice 11.0 dataset.
|
| 36 |
+
It achieves the following results on the evaluation set:
|
| 37 |
+
- Loss: 0.1915
|
| 38 |
+
- Wer: 12.4241
|
| 39 |
|
| 40 |
## Model description
|
| 41 |
|
|
|
|
| 55 |
|
| 56 |
The following hyperparameters were used during training:
|
| 57 |
- learning_rate: 1e-05
|
| 58 |
+
- train_batch_size: 8
|
| 59 |
+
- eval_batch_size: 8
|
| 60 |
- seed: 42
|
| 61 |
- gradient_accumulation_steps: 2
|
| 62 |
+
- total_train_batch_size: 16
|
| 63 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 64 |
- lr_scheduler_type: linear
|
| 65 |
- lr_scheduler_warmup_steps: 50
|
| 66 |
+
- training_steps: 500
|
| 67 |
- mixed_precision_training: Native AMP
|
| 68 |
|
| 69 |
+
### Training results
|
| 70 |
+
|
| 71 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 72 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|
|
| 73 |
+
| 0.0936 | 0.37 | 100 | 0.2300 | 16.2100 |
|
| 74 |
+
| 0.09 | 0.73 | 200 | 0.2117 | 14.4876 |
|
| 75 |
+
| 0.0415 | 1.1 | 300 | 0.2048 | 13.0832 |
|
| 76 |
+
| 0.0372 | 1.47 | 400 | 0.1951 | 12.5559 |
|
| 77 |
+
| 0.0307 | 1.84 | 500 | 0.1915 | 12.4241 |
|
| 78 |
+
|
| 79 |
+
|
| 80 |
### Framework versions
|
| 81 |
|
| 82 |
- Transformers 4.26.0.dev0
|