|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- precision |
|
- recall |
|
- f1 |
|
model-index: |
|
- name: multilingual_speech_to_intent_wav2vec |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# multilingual_speech_to_intent_wav2vec |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.5542 |
|
- Accuracy: 0.7430 |
|
- Precision: 0.8060 |
|
- Recall: 0.7430 |
|
- F1: 0.7456 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 0.0003 |
|
- train_batch_size: 32 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 100 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:| |
|
| 2.3588 | 1.0 | 219 | 1.4144 | 0.5916 | 0.6385 | 0.5916 | 0.5322 | |
|
| 0.8825 | 2.0 | 438 | 0.7289 | 0.8195 | 0.8635 | 0.8195 | 0.8243 | |
|
| 0.7836 | 3.0 | 657 | 0.6739 | 0.8514 | 0.8648 | 0.8514 | 0.8513 | |
|
| 0.7345 | 4.0 | 876 | 0.4483 | 0.9080 | 0.9189 | 0.9080 | 0.9071 | |
|
| 0.7204 | 5.0 | 1095 | 0.5039 | 0.8882 | 0.9059 | 0.8882 | 0.8915 | |
|
| 0.5355 | 6.0 | 1314 | 0.5051 | 0.8967 | 0.9049 | 0.8967 | 0.8971 | |
|
| 0.5939 | 7.0 | 1533 | 0.3162 | 0.9314 | 0.9387 | 0.9314 | 0.9322 | |
|
| 0.5311 | 8.0 | 1752 | 0.3218 | 0.9292 | 0.9318 | 0.9292 | 0.9292 | |
|
| 0.5098 | 9.0 | 1971 | 0.5819 | 0.8804 | 0.8858 | 0.8804 | 0.8809 | |
|
| 0.508 | 10.0 | 2190 | 0.5930 | 0.8804 | 0.8843 | 0.8804 | 0.8792 | |
|
| 0.4672 | 11.0 | 2409 | 0.3127 | 0.9229 | 0.9251 | 0.9229 | 0.9222 | |
|
| 0.4619 | 12.0 | 2628 | 0.3761 | 0.9193 | 0.9227 | 0.9193 | 0.9193 | |
|
| 0.4668 | 13.0 | 2847 | 0.6386 | 0.8740 | 0.8800 | 0.8740 | 0.8726 | |
|
| 0.444 | 14.0 | 3066 | 0.4134 | 0.9073 | 0.9133 | 0.9073 | 0.9079 | |
|
| 0.4059 | 15.0 | 3285 | 0.3106 | 0.9349 | 0.9370 | 0.9349 | 0.9347 | |
|
| 0.3857 | 16.0 | 3504 | 0.3639 | 0.9222 | 0.9296 | 0.9222 | 0.9217 | |
|
| 0.432 | 17.0 | 3723 | 0.5168 | 0.8896 | 0.8977 | 0.8896 | 0.8885 | |
|
| 0.3909 | 18.0 | 3942 | 1.0967 | 0.8004 | 0.8269 | 0.8004 | 0.8022 | |
|
| 0.4341 | 19.0 | 4161 | 0.7655 | 0.8556 | 0.8624 | 0.8556 | 0.8554 | |
|
| 0.3673 | 20.0 | 4380 | 0.2394 | 0.9505 | 0.9525 | 0.9505 | 0.9505 | |
|
| 0.3784 | 21.0 | 4599 | 0.4200 | 0.9207 | 0.9228 | 0.9207 | 0.9202 | |
|
| 0.4064 | 22.0 | 4818 | 0.5932 | 0.8818 | 0.8876 | 0.8818 | 0.8820 | |
|
| 0.3825 | 23.0 | 5037 | 0.9998 | 0.8493 | 0.8616 | 0.8493 | 0.8484 | |
|
| 0.3485 | 24.0 | 5256 | 1.1882 | 0.7877 | 0.8071 | 0.7877 | 0.7888 | |
|
| 0.3242 | 25.0 | 5475 | 0.5562 | 0.9073 | 0.9118 | 0.9073 | 0.9076 | |
|
| 0.3526 | 26.0 | 5694 | 0.6743 | 0.8832 | 0.8927 | 0.8832 | 0.8825 | |
|
| 0.3573 | 27.0 | 5913 | 0.3483 | 0.9271 | 0.9313 | 0.9271 | 0.9272 | |
|
| 0.3381 | 28.0 | 6132 | 1.1346 | 0.8018 | 0.8152 | 0.8018 | 0.8017 | |
|
| 0.3243 | 29.0 | 6351 | 0.9003 | 0.8316 | 0.8439 | 0.8316 | 0.8315 | |
|
| 0.3045 | 30.0 | 6570 | 0.9181 | 0.8493 | 0.8570 | 0.8493 | 0.8482 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.43.3 |
|
- Pytorch 2.1.0+cu118 |
|
- Datasets 2.20.0 |
|
- Tokenizers 0.19.1 |
|
|
