update model card README.md
Browse files
README.md
CHANGED
@@ -1,12 +1,7 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
tags:
|
4 |
-
- automatic-speech-recognition
|
5 |
-
- google/fleurs
|
6 |
- generated_from_trainer
|
7 |
-
- hf-asr-leaderboard
|
8 |
-
- ps
|
9 |
-
- Pashto
|
10 |
datasets:
|
11 |
- fleurs
|
12 |
metrics:
|
@@ -18,15 +13,15 @@ model-index:
|
|
18 |
name: Automatic Speech Recognition
|
19 |
type: automatic-speech-recognition
|
20 |
dataset:
|
21 |
-
name:
|
22 |
type: fleurs
|
23 |
config: ps_af
|
24 |
split: test
|
25 |
-
args:
|
26 |
metrics:
|
27 |
- name: Wer
|
28 |
type: wer
|
29 |
-
value: 0.
|
30 |
---
|
31 |
|
32 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -34,11 +29,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
34 |
|
35 |
# facebook/wav2vec2-xls-r-300m
|
36 |
|
37 |
-
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the
|
38 |
It achieves the following results on the evaluation set:
|
39 |
-
- Loss: 0.
|
40 |
-
- Wer: 0.
|
41 |
-
- Cer: 0.
|
42 |
|
43 |
## Model description
|
44 |
|
@@ -66,25 +61,26 @@ The following hyperparameters were used during training:
|
|
66 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
67 |
- lr_scheduler_type: linear
|
68 |
- lr_scheduler_warmup_steps: 2000
|
69 |
-
-
|
70 |
- mixed_precision_training: Native AMP
|
71 |
|
72 |
### Training results
|
73 |
|
74 |
-
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
75 |
-
|
76 |
-
| 5.0767 | 6.33 | 500 | 4.8783 | 1.0 |
|
77 |
-
| 3.1156 | 12.66 | 1000 | 3.0990 | 1.0 |
|
78 |
-
| 1.3506 | 18.99 | 1500 | 1.1056 | 0.7031 |
|
79 |
-
| 0.9997 | 25.32 | 2000 | 0.
|
80 |
-
| 0.7838 | 31.65 | 2500 | 0.
|
81 |
-
| 0.6665 | 37.97 | 3000 | 0.
|
82 |
-
| 0.6265 | 44.3 | 3500 | 0.
|
|
|
83 |
|
84 |
|
85 |
### Framework versions
|
86 |
|
87 |
- Transformers 4.26.0.dev0
|
88 |
-
- Pytorch 1.13.
|
89 |
- Datasets 2.7.1.dev0
|
90 |
- Tokenizers 0.13.2
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
tags:
|
|
|
|
|
4 |
- generated_from_trainer
|
|
|
|
|
|
|
5 |
datasets:
|
6 |
- fleurs
|
7 |
metrics:
|
|
|
13 |
name: Automatic Speech Recognition
|
14 |
type: automatic-speech-recognition
|
15 |
dataset:
|
16 |
+
name: fleurs
|
17 |
type: fleurs
|
18 |
config: ps_af
|
19 |
split: test
|
20 |
+
args: ps_af
|
21 |
metrics:
|
22 |
- name: Wer
|
23 |
type: wer
|
24 |
+
value: 0.5156036834924966
|
25 |
---
|
26 |
|
27 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
29 |
|
30 |
# facebook/wav2vec2-xls-r-300m
|
31 |
|
32 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the fleurs dataset.
|
33 |
It achieves the following results on the evaluation set:
|
34 |
+
- Loss: 0.9162
|
35 |
+
- Wer: 0.5156
|
36 |
+
- Cer: 0.1969
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
61 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
62 |
- lr_scheduler_type: linear
|
63 |
- lr_scheduler_warmup_steps: 2000
|
64 |
+
- training_steps: 4000
|
65 |
- mixed_precision_training: Native AMP
|
66 |
|
67 |
### Training results
|
68 |
|
69 |
+
| Training Loss | Epoch | Step | Cer | Validation Loss | Wer |
|
70 |
+
|:-------------:|:-----:|:----:|:------:|:---------------:|:------:|
|
71 |
+
| 5.0767 | 6.33 | 500 | 1.0 | 4.8783 | 1.0 |
|
72 |
+
| 3.1156 | 12.66 | 1000 | 1.0 | 3.0990 | 1.0 |
|
73 |
+
| 1.3506 | 18.99 | 1500 | 0.2889 | 1.1056 | 0.7031 |
|
74 |
+
| 0.9997 | 25.32 | 2000 | 0.2301 | 0.9191 | 0.5944 |
|
75 |
+
| 0.7838 | 31.65 | 2500 | 0.2152 | 0.8952 | 0.5556 |
|
76 |
+
| 0.6665 | 37.97 | 3000 | 0.2017 | 0.8908 | 0.5252 |
|
77 |
+
| 0.6265 | 44.3 | 3500 | 0.1954 | 0.9063 | 0.5133 |
|
78 |
+
| 0.5935 | 50.63 | 4000 | 0.9162 | 0.5156 | 0.1969 |
|
79 |
|
80 |
|
81 |
### Framework versions
|
82 |
|
83 |
- Transformers 4.26.0.dev0
|
84 |
+
- Pytorch 1.13.1+cu117
|
85 |
- Datasets 2.7.1.dev0
|
86 |
- Tokenizers 0.13.2
|