End of training
Browse files
README.md
CHANGED
|
@@ -5,9 +5,24 @@ tags:
|
|
| 5 |
- generated_from_trainer
|
| 6 |
datasets:
|
| 7 |
- marsyas/gtzan
|
|
|
|
|
|
|
| 8 |
model-index:
|
| 9 |
- name: distilhubert-finetuned-gtzan
|
| 10 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
---
|
| 12 |
|
| 13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -16,6 +31,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
# distilhubert-finetuned-gtzan
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
@@ -35,19 +53,35 @@ More information needed
|
|
| 35 |
|
| 36 |
The following hyperparameters were used during training:
|
| 37 |
- learning_rate: 5e-05
|
| 38 |
-
- train_batch_size:
|
| 39 |
-
- eval_batch_size:
|
| 40 |
- seed: 42
|
| 41 |
-
- gradient_accumulation_steps:
|
| 42 |
- total_train_batch_size: 8
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_ratio: 0.1
|
| 46 |
- num_epochs: 10
|
| 47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
### Framework versions
|
| 49 |
|
| 50 |
-
- Transformers 4.33.
|
| 51 |
- Pytorch 2.0.1+cu118
|
| 52 |
-
- Datasets 2.14.
|
| 53 |
- Tokenizers 0.13.3
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
datasets:
|
| 7 |
- marsyas/gtzan
|
| 8 |
+
metrics:
|
| 9 |
+
- accuracy
|
| 10 |
model-index:
|
| 11 |
- name: distilhubert-finetuned-gtzan
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Audio Classification
|
| 15 |
+
type: audio-classification
|
| 16 |
+
dataset:
|
| 17 |
+
name: GTZAN
|
| 18 |
+
type: marsyas/gtzan
|
| 19 |
+
config: all
|
| 20 |
+
split: train
|
| 21 |
+
args: all
|
| 22 |
+
metrics:
|
| 23 |
+
- name: Accuracy
|
| 24 |
+
type: accuracy
|
| 25 |
+
value: 0.84
|
| 26 |
---
|
| 27 |
|
| 28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 31 |
# distilhubert-finetuned-gtzan
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
|
| 34 |
+
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.5875
|
| 36 |
+
- Accuracy: 0.84
|
| 37 |
|
| 38 |
## Model description
|
| 39 |
|
|
|
|
| 53 |
|
| 54 |
The following hyperparameters were used during training:
|
| 55 |
- learning_rate: 5e-05
|
| 56 |
+
- train_batch_size: 4
|
| 57 |
+
- eval_batch_size: 4
|
| 58 |
- seed: 42
|
| 59 |
+
- gradient_accumulation_steps: 2
|
| 60 |
- total_train_batch_size: 8
|
| 61 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 62 |
- lr_scheduler_type: linear
|
| 63 |
- lr_scheduler_warmup_ratio: 0.1
|
| 64 |
- num_epochs: 10
|
| 65 |
|
| 66 |
+
### Training results
|
| 67 |
+
|
| 68 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 69 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 70 |
+
| 1.9664 | 1.0 | 112 | 1.7811 | 0.51 |
|
| 71 |
+
| 1.2687 | 2.0 | 225 | 1.2183 | 0.73 |
|
| 72 |
+
| 0.8758 | 3.0 | 337 | 0.9457 | 0.72 |
|
| 73 |
+
| 0.7324 | 4.0 | 450 | 0.9182 | 0.76 |
|
| 74 |
+
| 0.4299 | 5.0 | 562 | 0.6771 | 0.79 |
|
| 75 |
+
| 0.3001 | 6.0 | 675 | 0.6645 | 0.78 |
|
| 76 |
+
| 0.22 | 7.0 | 787 | 0.5920 | 0.82 |
|
| 77 |
+
| 0.2417 | 8.0 | 900 | 0.6002 | 0.82 |
|
| 78 |
+
| 0.1849 | 9.0 | 1012 | 0.6047 | 0.83 |
|
| 79 |
+
| 0.1259 | 9.96 | 1120 | 0.5875 | 0.84 |
|
| 80 |
+
|
| 81 |
+
|
| 82 |
### Framework versions
|
| 83 |
|
| 84 |
+
- Transformers 4.33.1
|
| 85 |
- Pytorch 2.0.1+cu118
|
| 86 |
+
- Datasets 2.14.5
|
| 87 |
- Tokenizers 0.13.3
|