End of training
Browse files
README.md
CHANGED
@@ -7,34 +7,34 @@ tags:
|
|
7 |
- hf-asr-leaderboard
|
8 |
- generated_from_trainer
|
9 |
datasets:
|
10 |
-
- Sonal0205/
|
11 |
metrics:
|
12 |
- wer
|
13 |
model-index:
|
14 |
-
- name: Whisper Medium te - telugu
|
15 |
results:
|
16 |
- task:
|
17 |
name: Automatic Speech Recognition
|
18 |
type: automatic-speech-recognition
|
19 |
dataset:
|
20 |
name: Telugu ASR
|
21 |
-
type: Sonal0205/
|
22 |
args: 'config: te, split: test'
|
23 |
metrics:
|
24 |
- name: Wer
|
25 |
type: wer
|
26 |
-
value:
|
27 |
---
|
28 |
|
29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
30 |
should probably proofread and complete it, then remove this comment. -->
|
31 |
|
32 |
-
# Whisper Medium te - telugu
|
33 |
|
34 |
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Telugu ASR dataset.
|
35 |
It achieves the following results on the evaluation set:
|
36 |
-
- Loss: 0.
|
37 |
-
- Wer:
|
38 |
|
39 |
## Model description
|
40 |
|
@@ -66,15 +66,15 @@ The following hyperparameters were used during training:
|
|
66 |
|
67 |
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
68 |
|:-------------:|:------:|:----:|:---------------:|:-------:|
|
69 |
-
| 0.
|
70 |
-
| 0.
|
71 |
-
| 0.
|
72 |
-
| 0.
|
73 |
|
74 |
|
75 |
### Framework versions
|
76 |
|
77 |
- Transformers 4.42.0.dev0
|
78 |
-
- Pytorch 2.4.0.
|
79 |
- Datasets 2.19.1
|
80 |
- Tokenizers 0.19.1
|
|
|
7 |
- hf-asr-leaderboard
|
8 |
- generated_from_trainer
|
9 |
datasets:
|
10 |
+
- Sonal0205/telugu_asr_recent
|
11 |
metrics:
|
12 |
- wer
|
13 |
model-index:
|
14 |
+
- name: Whisper Medium te - telugu new
|
15 |
results:
|
16 |
- task:
|
17 |
name: Automatic Speech Recognition
|
18 |
type: automatic-speech-recognition
|
19 |
dataset:
|
20 |
name: Telugu ASR
|
21 |
+
type: Sonal0205/telugu_asr_recent
|
22 |
args: 'config: te, split: test'
|
23 |
metrics:
|
24 |
- name: Wer
|
25 |
type: wer
|
26 |
+
value: 23.513674197384066
|
27 |
---
|
28 |
|
29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
30 |
should probably proofread and complete it, then remove this comment. -->
|
31 |
|
32 |
+
# Whisper Medium te - telugu new
|
33 |
|
34 |
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Telugu ASR dataset.
|
35 |
It achieves the following results on the evaluation set:
|
36 |
+
- Loss: 0.0925
|
37 |
+
- Wer: 23.5137
|
38 |
|
39 |
## Model description
|
40 |
|
|
|
66 |
|
67 |
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
68 |
|:-------------:|:------:|:----:|:---------------:|:-------:|
|
69 |
+
| 0.0452 | 1.0582 | 1000 | 0.0992 | 42.1522 |
|
70 |
+
| 0.0148 | 2.1164 | 2000 | 0.0884 | 27.2889 |
|
71 |
+
| 0.0065 | 3.1746 | 3000 | 0.0924 | 24.7919 |
|
72 |
+
| 0.0013 | 4.2328 | 4000 | 0.0925 | 23.5137 |
|
73 |
|
74 |
|
75 |
### Framework versions
|
76 |
|
77 |
- Transformers 4.42.0.dev0
|
78 |
+
- Pytorch 2.4.0.dev20240606+cu118
|
79 |
- Datasets 2.19.1
|
80 |
- Tokenizers 0.19.1
|