[Discussion] I am getting an error when trying to train a custom model #905
Labels
question
Further information is requested
Comments
|
Check if your training data contains an inconsistent number of channels. It looks like one of the samples is stereo, while another one is mono. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I am getting the following error when running the spleeter train command:
I am not sure what I am doing wron here is my config json:
{ "train_csv": "training_data.csv", "validation_csv": "validation_data.csv", "model_dir": "chdb_model", "mix_name": "combined", "instrument_list": ["vocals", "drums", "rhythm", "guitar", "song"], "sample_rate":44100, "frame_length":4096, "frame_step":1024, "T":512, "F":1024, "n_channels":2, "n_chunks_per_song":40, "separation_exponent":2, "mask_extension":"zeros", "learning_rate": 1e-4, "batch_size":4, "training_cache":"cache/training", "validation_cache":"cache/validation", "train_max_steps": 200000, "throttle_secs":1800, "random_seed":3, "save_checkpoints_steps":1000, "save_summary_steps":5, "model":{ "type":"unet.unet", "params":{ "conv_activation":"ELU", "deconv_activation":"ELU" } } }It does look like it is able to load a lot of the audio but I am not sure what it is doing when it fails out.
The text was updated successfully, but these errors were encountered: