End of training
Browse files
README.md
CHANGED
@@ -16,13 +16,10 @@ model-index:
|
|
16 |
dataset:
|
17 |
name: PolyAI/minds14
|
18 |
type: PolyAI/minds14
|
19 |
-
config: en-US
|
20 |
-
split: train
|
21 |
-
args: en-US
|
22 |
metrics:
|
23 |
- name: Wer
|
24 |
type: wer
|
25 |
-
value:
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -32,9 +29,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
32 |
|
33 |
This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the PolyAI/minds14 dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
-
- Loss:
|
36 |
-
- Wer Ortho:
|
37 |
-
- Wer:
|
38 |
|
39 |
## Model description
|
40 |
|
@@ -60,26 +57,19 @@ The following hyperparameters were used during training:
|
|
60 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
61 |
- lr_scheduler_type: constant_with_warmup
|
62 |
- lr_scheduler_warmup_steps: 50
|
63 |
-
- training_steps:
|
64 |
- mixed_precision_training: Native AMP
|
65 |
|
66 |
### Training results
|
67 |
|
68 |
-
| Training Loss | Epoch
|
69 |
-
|
70 |
-
| 0.
|
71 |
-
| 0.0002 | 35.71 | 1000 | 0.8518 | 37.8779 | 37.9412 |
|
72 |
-
| 0.0001 | 53.57 | 1500 | 0.8928 | 38.6181 | 38.6471 |
|
73 |
-
| 0.0001 | 71.43 | 2000 | 0.9228 | 38.2480 | 38.2353 |
|
74 |
-
| 0.0 | 89.29 | 2500 | 0.9525 | 38.3097 | 38.2941 |
|
75 |
-
| 0.0 | 107.14 | 3000 | 0.9759 | 38.9883 | 38.9412 |
|
76 |
-
| 0.0 | 125.0 | 3500 | 0.9967 | 38.9266 | 38.8824 |
|
77 |
-
| 0.0 | 142.86 | 4000 | 1.0162 | 39.6669 | 39.3529 |
|
78 |
|
79 |
|
80 |
### Framework versions
|
81 |
|
82 |
- Transformers 4.35.2
|
83 |
- Pytorch 2.1.0+cu121
|
84 |
-
- Datasets 2.
|
85 |
- Tokenizers 0.15.0
|
|
|
16 |
dataset:
|
17 |
name: PolyAI/minds14
|
18 |
type: PolyAI/minds14
|
|
|
|
|
|
|
19 |
metrics:
|
20 |
- name: Wer
|
21 |
type: wer
|
22 |
+
value: 33.64544319600499
|
23 |
---
|
24 |
|
25 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
29 |
|
30 |
This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the PolyAI/minds14 dataset.
|
31 |
It achieves the following results on the evaluation set:
|
32 |
+
- Loss: 0.7041
|
33 |
+
- Wer Ortho: 33.8532
|
34 |
+
- Wer: 33.6454
|
35 |
|
36 |
## Model description
|
37 |
|
|
|
57 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
58 |
- lr_scheduler_type: constant_with_warmup
|
59 |
- lr_scheduler_warmup_steps: 50
|
60 |
+
- training_steps: 500
|
61 |
- mixed_precision_training: Native AMP
|
62 |
|
63 |
### Training results
|
64 |
|
65 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|
66 |
+
|:-------------:|:-----:|:----:|:---------------:|:---------:|:-------:|
|
67 |
+
| 0.0015 | 17.86 | 500 | 0.7041 | 33.8532 | 33.6454 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
|
69 |
|
70 |
### Framework versions
|
71 |
|
72 |
- Transformers 4.35.2
|
73 |
- Pytorch 2.1.0+cu121
|
74 |
+
- Datasets 2.16.0
|
75 |
- Tokenizers 0.15.0
|