Update README.md
Browse files
README.md
CHANGED
@@ -12,4 +12,33 @@ datasets:
|
|
12 |
model-index:
|
13 |
- name: uzbek_stt
|
14 |
results: []
|
15 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
model-index:
|
13 |
- name: uzbek_stt
|
14 |
results: []
|
15 |
+
---
|
16 |
+
|
17 |
+
|
18 |
+
## Oyqiz jamoasi a'zolari tomonidan qilingan STT ning eng yaxshi versiyasi!
|
19 |
+
### Foziljon To'lqinov, Shaxboz Zohidov, Abduraxim Jabborov, Yahyoxon Rahimov, Mahmud Jumanazarov
|
20 |
+
|
21 |
+
|
22 |
+
|
23 |
+
Bu model [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) va MOZILLA-FOUNDATION/COMMON_VOICE_10_0 - UZ dataseti bilan 100ta epoxta o'qitilgan.
|
24 |
+
O'qitish natijalari:
|
25 |
+
- Xatolik: 0.1963
|
26 |
+
- So'z xatoligi: 0.2102
|
27 |
+
|
28 |
+
|
29 |
+
### O'qitish giperparameterlari
|
30 |
+
|
31 |
+
O'qitish uchun ishlatilgan giperparameterlar:
|
32 |
+
- learning_rate: 0.00003
|
33 |
+
- train_batch_size: 8
|
34 |
+
- eval_batch_size: 8
|
35 |
+
- seed: 42
|
36 |
+
- distributed_type: multi-GPU
|
37 |
+
- num_devices: 2
|
38 |
+
- total_train_batch_size: 16
|
39 |
+
- total_eval_batch_size: 16
|
40 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
41 |
+
- lr_scheduler_type: linear
|
42 |
+
- lr_scheduler_warmup_steps: 500
|
43 |
+
- num_epochs: 100.0
|
44 |
+
- mixed_precision_training: Native AMP
|