Hugging Face's logo

Freki89
/
whisper-medium-lora

PEFT
TensorBoard
Safetensors
Generated from Trainer
Model card Files Metrics Community
Freki89 commited on
Commit
30c8030
•
1 Parent(s): b185f41

Model save

Browse files
Files changed (3) hide show
  1. README.md +34 -40
  2. adapter_model.safetensors +1 -1
  3. runs/Apr27_08-18-28_129-213-80-45/events.out.tfevents.1714205908.129-213-80-45.9391.0 +2 -2
README.md CHANGED
@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.4144
20
 
21
  ## Model description
22
 
@@ -36,8 +36,8 @@ More information needed
36
 
37
  The following hyperparameters were used during training:
38
  - learning_rate: 1e-05
39
- - train_batch_size: 4
40
- - eval_batch_size: 4
41
  - seed: 42
42
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
  - lr_scheduler_type: linear
@@ -47,44 +47,38 @@ The following hyperparameters were used during training:
47
 
48
  ### Training results
49
 
50
- | Training Loss | Epoch | Step | Validation Loss |
51
- |:-------------:|:-----:|:----:|:---------------:|
52
- | 1.6978 | 0.17 | 250 | 1.4847 |
53
- | 1.2131 | 0.33 | 500 | 1.0400 |
54
- | 0.8693 | 0.5 | 750 | 0.7492 |
55
- | 0.5225 | 0.66 | 1000 | 0.4714 |
56
- | 0.4522 | 0.83 | 1250 | 0.4557 |
57
- | 0.454 | 0.99 | 1500 | 0.4465 |
58
- | 0.4393 | 1.16 | 1750 | 0.4410 |
59
- | 0.4084 | 1.33 | 2000 | 0.4369 |
60
- | 0.4046 | 1.49 | 2250 | 0.4338 |
61
- | 0.4088 | 1.66 | 2500 | 0.4308 |
62
- | 0.3949 | 1.82 | 2750 | 0.4281 |
63
- | 0.3647 | 1.99 | 3000 | 0.4261 |
64
- | 0.3762 | 2.16 | 3250 | 0.4248 |
65
- | 0.4099 | 2.32 | 3500 | 0.4226 |
66
- | 0.3753 | 2.49 | 3750 | 0.4216 |
67
- | 0.3845 | 2.65 | 4000 | 0.4200 |
68
- | 0.3408 | 2.82 | 4250 | 0.4189 |
69
- | 0.4163 | 2.98 | 4500 | 0.4186 |
70
- | 0.3382 | 3.15 | 4750 | 0.4177 |
71
- | 0.4096 | 3.32 | 5000 | 0.4171 |
72
- | 0.3698 | 3.48 | 5250 | 0.4162 |
73
- | 0.3885 | 3.65 | 5500 | 0.4157 |
74
- | 0.3334 | 3.81 | 5750 | 0.4152 |
75
- | 0.3882 | 3.98 | 6000 | 0.4148 |
76
- | 0.3534 | 4.14 | 6250 | 0.4141 |
77
- | 0.3838 | 4.31 | 6500 | 0.4144 |
78
- | 0.3564 | 4.48 | 6750 | 0.4139 |
79
- | 0.3689 | 4.64 | 7000 | 0.4140 |
80
- | 0.3389 | 4.81 | 7250 | 0.4143 |
81
- | 0.3892 | 4.97 | 7500 | 0.4144 |
82
 
83
 
84
  ### Framework versions
85
 
86
- - PEFT 0.8.2
87
- - Transformers 4.38.2
88
- - Pytorch 2.1.2+cu121
89
- - Datasets 2.16.1
90
- - Tokenizers 0.15.1
 
16
 
17
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.3493
20
 
21
  ## Model description
22
 
 
36
 
37
  The following hyperparameters were used during training:
38
  - learning_rate: 1e-05
39
+ - train_batch_size: 16
40
+ - eval_batch_size: 16
41
  - seed: 42
42
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
  - lr_scheduler_type: linear
 
47
 
48
  ### Training results
49
 
50
+ | Training Loss | Epoch | Step | Validation Loss |
51
+ |:-------------:|:------:|:----:|:---------------:|
52
+ | 1.636 | 0.2051 | 250 | 1.3245 |
53
+ | 1.0308 | 0.4102 | 500 | 0.8528 |
54
+ | 0.5665 | 0.6153 | 750 | 0.4152 |
55
+ | 0.419 | 0.8203 | 1000 | 0.3934 |
56
+ | 0.3843 | 1.0254 | 1250 | 0.3830 |
57
+ | 0.3811 | 1.2305 | 1500 | 0.3760 |
58
+ | 0.3602 | 1.4356 | 1750 | 0.3708 |
59
+ | 0.3609 | 1.6407 | 2000 | 0.3665 |
60
+ | 0.3541 | 1.8458 | 2250 | 0.3631 |
61
+ | 0.3682 | 2.0509 | 2500 | 0.3610 |
62
+ | 0.3527 | 2.2559 | 2750 | 0.3587 |
63
+ | 0.3468 | 2.4610 | 3000 | 0.3574 |
64
+ | 0.354 | 2.6661 | 3250 | 0.3559 |
65
+ | 0.3392 | 2.8712 | 3500 | 0.3546 |
66
+ | 0.3507 | 3.0763 | 3750 | 0.3533 |
67
+ | 0.3343 | 3.2814 | 4000 | 0.3526 |
68
+ | 0.3439 | 3.4865 | 4250 | 0.3520 |
69
+ | 0.3491 | 3.6916 | 4500 | 0.3512 |
70
+ | 0.3486 | 3.8966 | 4750 | 0.3505 |
71
+ | 0.3544 | 4.1017 | 5000 | 0.3502 |
72
+ | 0.3352 | 4.3068 | 5250 | 0.3498 |
73
+ | 0.3429 | 4.5119 | 5500 | 0.3492 |
74
+ | 0.3451 | 4.7170 | 5750 | 0.3494 |
75
+ | 0.324 | 4.9221 | 6000 | 0.3493 |
 
 
 
 
 
 
76
 
77
 
78
  ### Framework versions
79
 
80
+ - PEFT 0.10.0
81
+ - Transformers 4.40.1
82
+ - Pytorch 2.0.1
83
+ - Datasets 2.19.0
84
+ - Tokenizers 0.19.1
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fa4eb30d32f33d7cca08297abad0abdf025a6c54b9411ce214588904f181667e
3
  size 18915424
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:062f76a4355f142a2a3b02e9d36359c6ef9a5e812f6d8fb6fb3ac6e3b72ded56
3
  size 18915424
runs/Apr27_08-18-28_129-213-80-45/events.out.tfevents.1714205908.129-213-80-45.9391.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a31b0e50b21e59ab5d1e3ec976baa5ecfbd4857e5857843c54a2f5f3d67a2e04
3
- size 18577
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50b81d2cf33ce5bb8c9ec46148d321e659b58777d5df9797d89393e1a692a146
3
+ size 18931