diff options
| author | Pherkel | 2023-09-11 21:52:42 +0200 |
|---|---|---|
| committer | Pherkel | 2023-09-11 21:52:42 +0200 |
| commit | 58b30927bd870604a4077a8af9ec3cad7b0be21c (patch) | |
| tree | 7dd492fa8f14ff61c88545448972022ead324c31 /config.train.yaml | |
| parent | 9ca17d8a83369257f4cc42c963e25baf35a28f8f (diff) | |
changed config to yaml!
Diffstat (limited to 'config.train.yaml')
| -rw-r--r-- | config.train.yaml | 28 |
1 files changed, 28 insertions, 0 deletions
diff --git a/config.train.yaml b/config.train.yaml new file mode 100644 index 0000000..c82439d --- /dev/null +++ b/config.train.yaml @@ -0,0 +1,28 @@ +model: + n_cnn_layers: 3 + n_rnn_layers: 5 + rnn_dim: 512 + n_feats: 128 # number of mel features + stride: 2 + dropout: 0.25 # recommended to be around 0.4 for smaller datasets, 0.1 for really large datasets + +training: + learning_rate: 5e-4 + batch_size: 8 # recommended to maximum number that fits on the GPU (batch size of 32 fits on a 12GB GPU) + epochs: 3 + eval_every_n: 3 # evaluate every n epochs + num_workers: 8 # number of workers for dataloader + +dataset: + download: True + dataset_root_path: "YOUR/PATH" # files will be downloaded into this dir + language_name: "mls_german_opus" + limited_supervision: False # set to True if you want to use limited supervision + dataset_percentage: 1.0 # percentage of dataset to use (1.0 = 100%) + +tokenizer: + tokenizer_path: "data/tokenizers/char_tokenizer_german.yaml" + +checkpoints: + model_load_path: "YOUR/PATH" # path to load model from + model_save_path: "YOUR/PATH" # path to save model to
\ No newline at end of file |