Hugging Face's logo

jckim
/
checkpoints

PEFT
TensorBoard
Safetensors
multilingual
Generated from Trainer
Model card Files
xet
Metrics Community
jckim commited on
Commit
eddf0fb
·
verified ·
1 Parent(s): d4bdba3

gglabs/stt-turbo-multilingual-v0.0.10

Browse files
Files changed (4) hide show
  1. README.md +21 -11
  2. adapter_model.safetensors +1 -1
  3. runs/May23_11-15-26_ubuntu/events.out.tfevents.1747998928.ubuntu.2088417.0 +3 -0
  4. training_args.bin +1 -1
README.md CHANGED
@@ -20,8 +20,8 @@ should probably proofread and complete it, then remove this comment. -->
20
 
21
  This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the custom_multilingual dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.3849
24
- - Wer: 25.9259
25
 
26
  ## Model description
27
 
@@ -40,22 +40,32 @@ More information needed
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
- - learning_rate: 0.001
44
- - train_batch_size: 4
45
- - eval_batch_size: 4
46
  - seed: 42
 
 
47
  - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
48
  - lr_scheduler_type: linear
49
- - lr_scheduler_warmup_steps: 5
50
- - training_steps: 20
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
- | Training Loss | Epoch | Step | Validation Loss | Wer |
56
- |:-------------:|:------:|:----:|:---------------:|:-------:|
57
- | 0.6954 | 0.1754 | 10 | 0.4176 | 33.3333 |
58
- | 0.2358 | 0.3509 | 20 | 0.3849 | 25.9259 |
 
 
 
 
 
 
 
 
59
 
60
 
61
  ### Framework versions
 
20
 
21
  This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the custom_multilingual dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.3980
24
+ - Wer: 15.3116
25
 
26
  ## Model description
27
 
 
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
+ - learning_rate: 0.0005
44
+ - train_batch_size: 16
45
+ - eval_batch_size: 32
46
  - seed: 42
47
+ - gradient_accumulation_steps: 4
48
+ - total_train_batch_size: 64
49
  - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
50
  - lr_scheduler_type: linear
51
+ - lr_scheduler_warmup_steps: 200
52
+ - training_steps: 5000
53
  - mixed_precision_training: Native AMP
54
 
55
  ### Training results
56
 
57
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
58
+ |:-------------:|:-------:|:----:|:---------------:|:-------:|
59
+ | 0.3199 | 1.0971 | 500 | 0.3644 | 18.0913 |
60
+ | 0.2659 | 2.1942 | 1000 | 0.3579 | 17.1415 |
61
+ | 0.2321 | 3.2913 | 1500 | 0.3579 | 16.6319 |
62
+ | 0.2101 | 4.3884 | 2000 | 0.3635 | 16.3539 |
63
+ | 0.1901 | 5.4855 | 2500 | 0.3677 | 15.7470 |
64
+ | 0.1673 | 6.5826 | 3000 | 0.3748 | 15.4413 |
65
+ | 0.146 | 7.6796 | 3500 | 0.3813 | 15.4691 |
66
+ | 0.13 | 8.7767 | 4000 | 0.3879 | 15.6266 |
67
+ | 0.1206 | 9.8738 | 4500 | 0.3950 | 15.4459 |
68
+ | 0.1091 | 10.9709 | 5000 | 0.3980 | 15.3116 |
69
 
70
 
71
  ### Framework versions
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9e250308d65adb63715682609aad8380931d6a6e3eed668fc2cf884641d6895d
3
  size 78677456
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30c0d6d7a16b36b1af07c24d22195c702e8f39cb3baf65c0cdde8080ffa2f2d3
3
  size 78677456
runs/May23_11-15-26_ubuntu/events.out.tfevents.1747998928.ubuntu.2088417.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:103209f726da1817ae0c4cf7a1ba70022df77533b9f8beb9958eb0d1d9f078c8
3
+ size 20470
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cbf1d1b72d537e72b57798bbbbc8dd3efa369137a34584d722d696526f8e4e06
3
  size 5496
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8373683a2465563929a973e2ec6b94a5075a0414524cfa447c6b53e234a32892
3
  size 5496