There is no `train.txt` file in the `/srv/audio-classification-tf/models/2020-07-24_010059/dataset_files` directory."

[amgsharma]

The steps I took to arrive at this message are as follows: 1)

docker search deephdc docker run -ti -p 5000:5000 -p 6006:6006 -p 8888:8888 deephdc/deep-oc-audio-classification-tf

Put your images in the./data/audios folder. If you have your data somewhere else you can use that location by setting the dataset_directory parameter in the training args. Please use a standard audio format (like .mp3 or .wav).

I placed my audio files in /data/train_data/sound_1 and /data/train_data/sound_2

The response body looks like:

{
  "date": "2020-07-24 01:17:10.078958",
  "args": {
    "base_directory": "\"/Users/user/audio-classification-tf/data/dataset_files\"",
    "dataset_directory": "\"/Users/user/audio-classification-tf/data/train_data\"",
    "num_classes": "2",
    "files_to_PCM": "true",
    "compute_embeddings": "true",
    "mode": "\"normal\"",
    "initial_lr": "0.001",
    "batch_size": "500",
    "epochs": "15",
    "ckpt_freq": "null",
    "lr_schedule_mode": "\"step\"",
    "lr_step_decay": "0.1",
    "lr_step_schedule": "[0.7, 0.9]",
    "l2_reg": "0.0001",
    "use_validation": "false",
    "use_early_stopping": "false",
    "use_multiprocessing": "true",
    "use_tensorboard": "true",
    "use_remote": "false"
  },
  "uuid": "e5e2d012c4754cd491f44e43834579ef",
  "status": "error",
  "message": "Invalid value for the split_name parameter: there is no `train.txt` file in the `/Users/user/audio-classification-tf/data/dataset_files/models/2020-07-24_011710/dataset_files` directory."

and well, that makes sense because the dataset_files directory is in the base_dir path above.

[IgnacioHeredia]

You have to leave base_dir with it's default value (or use /Users/user/audio-classification-tf without the final data/dataset_files) because this has to point to the main folder, not the dataset folder.