End of training
Browse files
README.md
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: openai/whisper-small
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
datasets:
|
| 8 |
+
- balbus-classifier
|
| 9 |
+
metrics:
|
| 10 |
+
- accuracy
|
| 11 |
+
model-index:
|
| 12 |
+
- name: miosipof/whisper-small-ft-balbus-sep28k-v1.2
|
| 13 |
+
results:
|
| 14 |
+
- task:
|
| 15 |
+
name: Audio Classification
|
| 16 |
+
type: audio-classification
|
| 17 |
+
dataset:
|
| 18 |
+
name: Apple dataset
|
| 19 |
+
type: balbus-classifier
|
| 20 |
+
config: default
|
| 21 |
+
split: train
|
| 22 |
+
args: default
|
| 23 |
+
metrics:
|
| 24 |
+
- name: Accuracy
|
| 25 |
+
type: accuracy
|
| 26 |
+
value: 0.8082105922908179
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 30 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 31 |
+
|
| 32 |
+
# miosipof/whisper-small-ft-balbus-sep28k-v1.2
|
| 33 |
+
|
| 34 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Apple dataset dataset.
|
| 35 |
+
It achieves the following results on the evaluation set:
|
| 36 |
+
- Loss: 0.1107
|
| 37 |
+
- Accuracy: 0.8082
|
| 38 |
+
|
| 39 |
+
## Model description
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Intended uses & limitations
|
| 44 |
+
|
| 45 |
+
More information needed
|
| 46 |
+
|
| 47 |
+
## Training and evaluation data
|
| 48 |
+
|
| 49 |
+
More information needed
|
| 50 |
+
|
| 51 |
+
## Training procedure
|
| 52 |
+
|
| 53 |
+
### Training hyperparameters
|
| 54 |
+
|
| 55 |
+
The following hyperparameters were used during training:
|
| 56 |
+
- learning_rate: 5e-06
|
| 57 |
+
- train_batch_size: 32
|
| 58 |
+
- eval_batch_size: 16
|
| 59 |
+
- seed: 42
|
| 60 |
+
- gradient_accumulation_steps: 2
|
| 61 |
+
- total_train_batch_size: 64
|
| 62 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 63 |
+
- lr_scheduler_type: linear
|
| 64 |
+
- lr_scheduler_warmup_ratio: 0.5
|
| 65 |
+
- training_steps: 500
|
| 66 |
+
- mixed_precision_training: Native AMP
|
| 67 |
+
|
| 68 |
+
### Training results
|
| 69 |
+
|
| 70 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 71 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|
|
| 72 |
+
| 0.1709 | 0.1253 | 50 | 0.1690 | 0.5650 |
|
| 73 |
+
| 0.1669 | 0.2506 | 100 | 0.1630 | 0.6369 |
|
| 74 |
+
| 0.1542 | 0.3759 | 150 | 0.1439 | 0.7131 |
|
| 75 |
+
| 0.1283 | 0.5013 | 200 | 0.1214 | 0.7802 |
|
| 76 |
+
| 0.1158 | 0.6266 | 250 | 0.1171 | 0.7935 |
|
| 77 |
+
| 0.1059 | 0.7519 | 300 | 0.1131 | 0.7985 |
|
| 78 |
+
| 0.1142 | 0.8772 | 350 | 0.1102 | 0.8081 |
|
| 79 |
+
| 0.104 | 1.0025 | 400 | 0.1112 | 0.8068 |
|
| 80 |
+
| 0.0924 | 1.1278 | 450 | 0.1114 | 0.8087 |
|
| 81 |
+
| 0.0959 | 1.2531 | 500 | 0.1107 | 0.8082 |
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
### Framework versions
|
| 85 |
+
|
| 86 |
+
- Transformers 4.45.2
|
| 87 |
+
- Pytorch 2.2.0
|
| 88 |
+
- Datasets 3.2.0
|
| 89 |
+
- Tokenizers 0.20.3
|