Update app.py

app.py

@@ -65,6 +65,8 @@ from ultimate_rvc.web.config.component import (
     DropdownConfig,
     RadioConfig,
     SliderConfig,
+    NumberConfig,
+    TextboxConfig,
 )
 from ultimate_rvc.web.config.tab import (
     SongGenerationConfig,
@@ -89,6 +91,847 @@ type StrPath = str | PathLike[str]
 type Json = Mapping[str, Json] | Sequence[Json] | str | int | float | bool | None
 
 
+
+class BaseTabConfig(BaseModel):
+    """
+    Base model defining common component configuration settings for
+    UI tabs.
+
+    Attributes
+    ----------
+    embedder_model : DropdownConfig
+        Configuration settings for an embedder model dropdown component.
+    custom_embedder_model : DropdownConfig
+        Configuration settings for a custom embedder model dropdown
+        component.
+
+    """
+
+    embedder_model: DropdownConfig = DropdownConfig(
+        label="Embedder model",
+        info="The model to use for generating speaker embeddings.",
+        value=EmbedderModel.CONTENTVEC,
+        choices=list(EmbedderModel),
+        exclude_value=True,
+    )
+    custom_embedder_model: DropdownConfig = DropdownConfig(
+        label="Custom embedder model",
+        info="Select a custom embedder model from the dropdown.",
+        value=None,
+        visible=False,
+        render=False,
+        exclude_value=True,
+    )
+
+
+class GenerationConfig(BaseTabConfig):
+    """
+    Common component configuration settings for generation tabs.
+
+    voice_model : DropdownConfig
+        Configuration settings for a voice model dropdown component.
+    f0_methods : DropdownConfig
+        Configuration settings for a pitch extraction algorithms
+        dropdown component.
+    index_rate : SliderConfig
+        Configuration settings for an index rate slider component.
+    rms_mix_rate : SliderConfig
+        Configuration settings for a RMS mix rate slider component.
+    protect_rate : SliderConfig
+        Configuration settings for a protect rate slider component.
+    split_voice : CheckboxConfig
+        Configuration settings for a split voice checkbox component.
+    autotune_voice: CheckboxConfig
+        Configuration settings for an autotune voice checkbox component.
+    autotune_strength: SliderConfig
+        Configuration settings for an autotune strength slider
+        component.
+    sid : NumberConfig
+        Configuration settings for a speaker ID number component.
+    output_sr : DropdownConfig
+        Configuration settings for an output sample rate dropdown
+        component.
+    output_format : DropdownConfig
+        Configuration settings for an output format dropdown
+        component.
+    output_name : TextboxConfig
+        Configuration settings for an output name textbox component.
+
+    See Also
+    --------
+    BaseTabConfig
+        Parent model defining common component configuration settings
+        for UI tabs.
+
+    """
+
+    voice_model: DropdownConfig = DropdownConfig(
+        label="Voice model",
+        info="Select a model to use for voice conversion.",
+        value=None,
+        render=False,
+        exclude_value=True,
+    )
+    f0_methods: DropdownConfig = DropdownConfig(
+        label="Pitch extraction algorithm(s)",
+        info=(
+            "If more than one method is selected, then the median of the pitch values"
+            " extracted by each method is used. RMVPE is recommended for most cases and"
+            " is the default when no method is selected."
+        ),
+        value=[F0Method.RMVPE],
+        choices=list(F0Method),
+        multiselect=True,
+    )
+    index_rate: SliderConfig = SliderConfig(
+        label="Index rate",
+        info=(
+            "Increase to bias the conversion towards the accent of the voice model."
+            " Decrease to potentially reduce artifacts coming from the voice"
+            " model.<br><br><br>"
+        ),
+        value=0.3,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    rms_mix_rate: SliderConfig = SliderConfig(
+        label="RMS mix rate",
+        info=(
+            "How much to mimic the loudness (0) of the input voice or a fixed loudness"
+            " (1). A value of 1 is recommended for most cases.<br><br>"
+        ),
+        value=1.0,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    protect_rate: SliderConfig = SliderConfig(
+        label="Protect rate",
+        info=(
+            "Controls the extent to which consonants and breathing sounds are protected"
+            " from artifacts. A higher value offers more protection but may worsen the"
+            " indexing effect.<br><br>"
+        ),
+        value=0.33,
+        minimum=0.0,
+        maximum=0.5,
+    )
+
+    hop_length: SliderConfig = SliderConfig.hop_length(
+        label="Hop length",
+        info=(
+            "How often the CREPE-based pitch extraction method checks for pitch changes"
+            " measured in milliseconds. Lower values lead to longer conversion times"
+            " and a higher risk of voice cracks, but better pitch accuracy."
+        ),
+        visible=True,
+    )
+
+    split_voice: CheckboxConfig = CheckboxConfig(
+        label="Split input voice",
+        info=(
+            "Whether to split the input voice track into smaller segments before"
+            " converting it. This can improve output quality for longer voice tracks."
+        ),
+        value=False,
+    )
+    autotune_voice: CheckboxConfig = CheckboxConfig(
+        label="Autotune converted voice",
+        info="Whether to apply autotune to the converted voice.<br><br>",
+        value=False,
+        exclude_value=True,
+    )
+    autotune_strength: SliderConfig = SliderConfig(
+        label="Autotune intensity",
+        info=(
+            "Higher values result in stronger snapping to the chromatic grid and"
+            " artifacting."
+        ),
+        value=1.0,
+        minimum=0.0,
+        maximum=1.0,
+        visible=False,
+    )
+    sid: NumberConfig = NumberConfig(
+        label="Speaker ID",
+        info="Speaker ID for multi-speaker-models.",
+        value=0,
+        precision=0,
+    )
+    output_sr: DropdownConfig = DropdownConfig(
+        label="Output sample rate",
+        info="The sample rate of the mixed output track.",
+        value=SampleRate.HZ_44100,
+        choices=list(SampleRate),
+    )
+    output_format: DropdownConfig = DropdownConfig(
+        label="Output format",
+        info="The audio format of the mixed output track.",
+        value=AudioExt.MP3,
+        choices=list(AudioExt),
+    )
+    output_name: TextboxConfig = TextboxConfig(
+        label="Output name",
+        info="If no name is provided, a suitable name will be generated automatically.",
+        value=None,
+        placeholder="Ultimate RVC output",
+        exclude_value=True,
+    )
+
+
+class SongGenerationConfig(GenerationConfig):
+    """
+    Common component configuration settings for song generation tabs.
+
+    Attributes
+    ----------
+    source_type : DropdownConfig
+        Configuration settings for a source type dropdown component.
+    source : TextboxConfig
+        Configuration settings for an input source textbox component.
+    cached_song : DropdownConfig
+        Configuration settings for a cached song dropdown component.
+    clean_strength : SliderConfig
+        Configuration settings for a clean strength slider component.
+    clean_voice : CheckboxConfig
+        Configuration settings for a clean voice checkbox component.
+    room_size : SliderConfig
+        Configuration settings for a room size slider component.
+    wet_level : SliderConfig
+        Configuration settings for a wetness level slider component.
+    dry_level : SliderConfig
+        Configuration settings for a dryness level slider component.
+    damping : SliderConfig
+        Configuration settings for a damping level slider component.
+    main_gain : SliderConfig
+        Configuration settings for a main gain slider component.
+    inst_gain : SliderConfig
+        Configuration settings for an instrumentals gain slider
+        component.
+    backup_gain : SliderConfig
+        Configuration settings for a backup vocals gain slider
+        component.
+
+    See Also
+    --------
+    GenerationConfig
+        Parent model defining common component configuration settings
+        for song generation tabs.
+
+    """
+
+    source_type: DropdownConfig = DropdownConfig(
+        label="Source type",
+        info="The type of source to retrieve a song from.",
+        value=SongSourceType.LOCAL_FILE,
+        choices=list(SongSourceType),
+        type="index",
+        exclude_value=True,
+    )
+    source: TextboxConfig = TextboxConfig(
+        label="Source",
+        info="Link to a song on YouTube or the full path of a local audio file.",
+        value=None,
+        exclude_value=True,
+    )
+    cached_song: DropdownConfig = DropdownConfig(
+        label="Source",
+        info="Select a song from the list of cached songs.",
+        value=None,
+        visible=False,
+        render=False,
+        exclude_value=True,
+    )
+    clean_voice: CheckboxConfig = CheckboxConfig(
+        label="Clean converted voice",
+        info=(
+            "Whether to clean the converted voice using noise reduction"
+            " algorithms.<br><br>"
+        ),
+        value=False,
+        exclude_value=True,
+    )
+    clean_strength: SliderConfig = SliderConfig.clean_strength(visible=False)
+    room_size: SliderConfig = SliderConfig(
+        label="Room size",
+        info=(
+            "Size of the room which reverb effect simulates. Increase for longer reverb"
+            " time."
+        ),
+        value=0.15,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    wet_level: SliderConfig = SliderConfig(
+        label="Wetness level",
+        info="Loudness of converted vocals with reverb effect applied.",
+        value=0.2,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    dry_level: SliderConfig = SliderConfig(
+        label="Dryness level",
+        info="Loudness of converted vocals without reverb effect applied.",
+        value=0.8,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    damping: SliderConfig = SliderConfig(
+        label="Damping level",
+        info="Absorption of high frequencies in reverb effect.",
+        value=0.7,
+        minimum=0.0,
+        maximum=1.0,
+    )
+    main_gain: SliderConfig = SliderConfig.gain(
+        label="Main gain",
+        info="The gain to apply to the main vocals.",
+    )
+    inst_gain: SliderConfig = SliderConfig.gain(
+        label="Instrumentals gain",
+        info="The gain to apply to the instrumentals.",
+    )
+    backup_gain: SliderConfig = SliderConfig.gain(
+        label="Backup gain",
+        info="The gain to apply to the backup vocals.",
+    )
+
+
+class SpeechGenerationConfig(GenerationConfig):
+    """
+    Common component configuration settings for speech generation tabs.
+
+    Attributes
+    ----------
+    source_type : DropdownConfig
+        Configuration settings for a source type dropdown component.
+    source : TextboxConfig
+        Configuration settings for an input source textbox component.
+    edge_tts_voice : DropdownConfig
+        Configuration settings for an Edge TTS voice dropdown
+        component.
+    n_octaves : SliderConfig
+        Configuration settings for an octave pitch shift slider
+        component.
+    n_semitones : SliderConfig
+        Configuration settings for a semitone pitch shift slider
+        component.
+    tts_pitch_shift : SliderConfig
+        Configuration settings for a TTS pitch shift slider
+        component.
+    tts_speed_change : SliderConfig
+        Configuration settings for a TTS speed change slider
+        component.
+    tts_volume_change : SliderConfig
+        Configuration settings for a TTS volume change slider
+        component.
+    clean_voice : CheckboxConfig
+        Configuration settings for a clean voice checkbox
+        component.
+    clean_strength : SliderConfig
+        Configuration settings for a clean strength slider
+        component.
+    output_gain : GainSliderConfig
+        Configuration settings for an output gain slider component.
+
+    See Also
+    --------
+    GenerationConfig
+        Parent model defining common component configuration settings
+        for generation tabs.
+
+    """
+
+    source_type: DropdownConfig = DropdownConfig(
+        label="Source type",
+        info="The type of source to generate speech from.",
+        value=SpeechSourceType.TEXT,
+        choices=list(SpeechSourceType),
+        type="index",
+        exclude_value=True,
+    )
+    source: TextboxConfig = TextboxConfig(
+        label="Source",
+        info="Text to generate speech from",
+        value=None,
+        exclude_value=True,
+    )
+    edge_tts_voice: DropdownConfig = DropdownConfig(
+        label="Edge TTS voice",
+        info="Select a voice to use for text to speech conversion.",
+        value=None,
+        render=False,
+        exclude_value=True,
+    )
+    n_octaves: SliderConfig = SliderConfig.octave_shift(
+        label="Octave shift",
+        info=(
+            "The number of octaves to pitch-shift the converted speech by. Use 1 for"
+            " male-to-female and -1 for vice-versa."
+        ),
+    )
+    n_semitones: SliderConfig = SliderConfig.semitone_shift(
+        label="Semitone shift",
+        info="The number of semi-tones to pitch-shift the converted speech by.",
+    )
+    tts_pitch_shift: SliderConfig = SliderConfig(
+        label="Edge TTS pitch shift",
+        info=(
+            "The number of hertz to shift the pitch of the speech generated by Edge"
+            " TTS."
+        ),
+        value=0,
+        minimum=-100,
+        maximum=100,
+        step=1,
+    )
+    tts_speed_change: SliderConfig = SliderConfig(
+        label="TTS speed change",
+        info="The percentual change to the speed of the speech generated by Edge TTS.",
+        value=0,
+        minimum=-50,
+        maximum=100,
+        step=1,
+    )
+    tts_volume_change: SliderConfig = SliderConfig(
+        label="TTS volume change",
+        info="The percentual change to the volume of the speech generated by Edge TTS.",
+        value=0,
+        minimum=-100,
+        maximum=100,
+        step=1,
+    )
+    clean_voice: CheckboxConfig = CheckboxConfig(
+        label="Clean converted voice",
+        info=(
+            "Whether to clean the converted voice using noise reduction"
+            " algorithms.<br><br>"
+        ),
+        value=True,
+        exclude_value=True,
+    )
+    clean_strength: SliderConfig = SliderConfig.clean_strength(visible=True)
+    output_gain: SliderConfig = SliderConfig.gain(
+        label="Output gain",
+        info="The gain to apply to the converted speech.<br><br>",
+    )
+
+
+class TrainingConfig(BaseTabConfig):
+    """
+    Common component configuration settings for training tabs.
+
+    Attributes
+    ----------
+    dataset_type : DropdownConfig
+        Configuration settings for a dataset type dropdown component.
+    dataset : DropdownConfig
+        Configuration settings for a dataset dropdown component.
+    dataset_name : TextboxConfig
+        Configuration settings for a dataset name textbox component.
+    preprocess_model : DropdownConfig
+        Configuration settings for a model name dropdown component
+        for audio preprocessing.
+    sample_rate : DropdownConfig
+        Configuration settings for a sample rate dropdown component.
+    filter_audio : CheckboxConfig
+        Configuration settings for a filter audio checkbox component.
+    clean_audio : CheckboxConfig
+        Configuration settings for a clean audio checkbox component.
+    clean_strength : SliderConfig
+        Configuration settings for a clean strength slider component.
+    split_method : DropdownConfig
+        Configuration settings for an audio splitting method dropdown
+        component.
+    chunk_len : SliderConfig
+        Configuration settings for a chunk length slider component.
+    overlap_len : SliderConfig
+        Configuration settings for an overlap length slider component.
+    preprocess_cores : SliderConfig
+        Configuration settings for a CPU cores slider component for
+        preprocessing.
+    extract_model : DropdownConfig
+        Configuration settings for a model name dropdown component for
+        feature extraction.
+    f0_method : DropdownConfig
+        Configuration settings for an F0 method dropdown component.
+    hop_length : SliderConfig
+        Configuration settings for a hop length slider component.
+    include_mutes : SliderConfig
+        Configuration settings for an include mutes slider component.
+    extract_cores : SliderConfig
+        Configuration settings for a CPU cores slider component for
+        feature extraction.
+    extraction_acceleration : HardwareAccelerationConfig
+        Configuration settings for a hardware acceleration component for
+        feature extraction.
+    extraction_gpus : DropdownConfig
+        Configuration settings for a GPU dropdown compoennt for feature
+        extraction.
+    train_model : DropdownConfig
+        Configuration settings for a model name dropdown component for
+        training.
+    num_epochs : SliderConfig
+        Configuration settings for a number of epochs slider component.
+    batch_size : SliderConfig
+        Configuration settings for a batch size slider component.
+    detect_overtraining : CheckboxConfig
+        Configuration settings for a detect overtraining checkbox
+        component.
+    overtraining_threshold : SliderConfig
+        Configuration settings for an overtraining threshold slider
+        component.
+    vocoder : DropdownConfig
+        Configuration settings for a vocoder dropdown component.
+    index_algorithm : DropdownConfig
+        Configuration settings for an index algorithm dropdown
+        component.
+    pretrained_type : DropdownConfig
+        Configuration settings for a pretrained model type dropdown
+        component.
+    custom_pretrained_model : DropdownConfig
+        Configuration settings for a custom pretrained model dropdown
+        component.
+    save_interval : SliderConfig
+        Configuration settings for a save-interval slider component.
+    save_all_checkpoints : CheckboxConfig
+        Configuration settings for a save-all-checkpoints checkbox
+        component.
+    save_all_weights : CheckboxConfig
+        Configuration settings for a save-all-weights checkbox
+        component.
+    clear_saved_data : CheckboxConfig
+        Configuration settings for a clear-saved-data checkbox
+        component.
+    upload_model : CheckboxConfig
+        Configuration settings for an upload voice model checkbox
+        component.
+    upload_name : TextboxConfig
+        Configuration settings for an upload name textbox component.
+    training_acceleration : HardwareAccelerationConfig
+        Configuration settings for a hardware acceleration component for
+        training.
+    training_gpus : DropdownConfig
+        Configuration settings for a GPU dropdown component for
+        training.
+    preload_dataset : CheckboxConfig
+        Configuration settings for a preload dataset checkbox component.
+    reduce_memory_usage : CheckboxConfig
+        Configuration settings for a reduce-memory-usage checkbox
+        component.
+
+    See Also
+    --------
+    BaseTabConfig
+        Parent model defining common component configuration settings
+        for UI tabs.
+
+    """
+
+    dataset_type: DropdownConfig = DropdownConfig(
+        label="Dataset type",
+        info="Select the type of dataset to preprocess.",
+        value=DatasetType.NEW_DATASET,
+        choices=list(DatasetType),
+        exclude_value=True,
+    )
+    dataset: DropdownConfig = DropdownConfig(
+        label="Dataset path",
+        info=(
+            "The path to an existing dataset. Either select a path to a previously"
+            " created dataset or provide a path to an external dataset."
+        ),
+        value=None,
+        allow_custom_value=True,
+        visible=False,
+        render=False,
+        exclude_value=True,
+    )
+    dataset_name: TextboxConfig = TextboxConfig(
+        label="Dataset name",
+        info=(
+            "The name of the new dataset. If the dataset already exists, the provided"
+            " audio files will be added to it."
+        ),
+        value="My dataset",
+        exclude_value=True,
+    )
+    preprocess_model: DropdownConfig = DropdownConfig(
+        label="Model name",
+        info=(
+            "Name of the model to preprocess the given dataset for. Either select an"
+            " existing model from the dropdown or provide the name of a new model."
+        ),
+        value="My model",
+        allow_custom_value=True,
+        render=False,
+        exclude_value=True,
+    )
+    sample_rate: DropdownConfig = DropdownConfig(
+        label="Sample rate",
+        info="Target sample rate for the audio files in the provided dataset.",
+        value=TrainingSampleRate.HZ_40K,
+        choices=list(TrainingSampleRate),
+    )
+    filter_audio: CheckboxConfig = CheckboxConfig(
+        label="Filter audio",
+        info=(
+            "Whether to remove low-frequency sounds from the audio files in the"
+            " provided dataset by applying a high-pass butterworth filter.<br><br>"
+        ),
+        value=True,
+    )
+    clean_audio: CheckboxConfig = CheckboxConfig(
+        label="Clean audio",
+        info=(
+            "Whether to clean the audio files in the provided dataset using noise"
+            " reduction algorithms.<br><br><br>"
+        ),
+        value=False,
+        exclude_value=True,
+    )
+    clean_strength: SliderConfig = SliderConfig.clean_strength(visible=False)
+    split_method: DropdownConfig = DropdownConfig(
+        label="Audio splitting method",
+        info=(
+            "The method to use for splitting the audio files in the provided dataset."
+            " Use the `Skip` method to skip splitting if the audio files are already"
+            " split. Use the `Simple` method if excessive silence has already been"
+            " removed from the audio files. Use the `Automatic` method for automatic"
+            " silence detection and splitting around it."
+        ),
+        value=AudioSplitMethod.AUTOMATIC,
+        choices=list(AudioSplitMethod),
+        exclude_value=True,
+    )
+    chunk_len: SliderConfig = SliderConfig(
+        label="Chunk length",
+        info="Length of split audio chunks.",
+        value=3.0,
+        minimum=0.5,
+        maximum=5.0,
+        step=0.1,
+        visible=False,
+    )
+    overlap_len: SliderConfig = SliderConfig(
+        label="Overlap length",
+        info="Length of overlap between split audio chunks.",
+        value=0.3,
+        minimum=0.0,
+        maximum=0.4,
+        step=0.1,
+        visible=False,
+    )
+    preprocess_cores: SliderConfig = SliderConfig.cpu_cores()
+
+    extract_model: DropdownConfig = DropdownConfig(
+        label="Model name",
+        info=(
+            "Name of the model with an associated preprocessed dataset to extract"
+            " training features from. When a new dataset is preprocessed, its"
+            " associated model is selected by default."
+        ),
+        value=None,
+        render=False,
+        exclude_value=True,
+    )
+    f0_method: DropdownConfig = DropdownConfig(
+        label="F0 method",
+        info="The method to use for extracting pitch features.",
+        value=TrainingF0Method.RMVPE,
+        choices=list(TrainingF0Method),
+        exclude_value=True,
+    )
+
+    hop_length: SliderConfig = SliderConfig.hop_length(
+        label="Hop length",
+        info="The hop length to use for extracting pitch features.<br><br>",
+        visible=False,
+    )
+    include_mutes: SliderConfig = SliderConfig(
+        label="Include mutes",
+        info=(
+            "The number of mute audio files to include in the generated training file"
+            " list. Adding silent files enables the training model to handle pure"
+            " silence in inferred audio files. If the preprocessed audio dataset"
+            " already contains segments of pure silence, set this to 0."
+        ),
+        value=0,
+        minimum=0,
+        maximum=10,
+        step=1,
+    )
+    extraction_cores: SliderConfig = SliderConfig.cpu_cores()
+    extraction_acceleration: DropdownConfig = DropdownConfig.hardware_acceleration()
+    extraction_gpus: DropdownConfig = DropdownConfig.gpu()
+
+    train_model: DropdownConfig = DropdownConfig(
+        label="Model name",
+        info=(
+            "Name of the model to train. When training features are extracted for a new"
+            " model, its name is selected by default."
+        ),
+        value=None,
+        render=False,
+        exclude_value=True,
+    )
+    num_epochs: SliderConfig = SliderConfig(
+        label="Number of epochs",
+        info=(
+            "The number of epochs to train the voice model. A higher number can improve"
+            " voice model performance but may lead to overtraining."
+        ),
+        value=500,
+        minimum=1,
+        maximum=5000,
+        step=1,
+    )
+    batch_size: SliderConfig = SliderConfig(
+        label="Batch size",
+        info=(
+            "The number of samples in each training batch. It is advisable to align"
+            " this value with the available VRAM of your GPU."
+        ),
+        value=16,
+        minimum=1,
+        maximum=128,
+        step=1,
+    )
+    detect_overtraining: CheckboxConfig = CheckboxConfig(
+        label="Detect overtraining",
+        info=(
+            "Whether to detect overtraining to prevent the voice model from learning"
+            " the training data too well and losing the ability to generalize to new"
+            " data."
+        ),
+        value=True,
+        exclude_value=True,
+    )
+    overtraining_threshold: SliderConfig = SliderConfig(
+        label="Overtraining threshold",
+        info=(
+            "The maximum number of epochs to continue training without any observed"
+            " improvement in voice model performance."
+        ),
+        value=500,
+        minimum=1,
+        maximum=1000,
+        visible=False,
+    )
+    vocoder: DropdownConfig = DropdownConfig(
+        label="Vocoder",
+        info=(
+            "The vocoder to use for audio synthesis during training. HiFi-GAN provides"
+            " basic audio fidelity, while RefineGAN provides the highest audio"
+            " fidelity."
+        ),
+        value=Vocoder.HIFI_GAN,
+        choices=list(Vocoder),
+    )
+    index_algorithm: DropdownConfig = DropdownConfig(
+        label="Index algorithm",
+        info=(
+            "The method to use for generating an index file for the trained voice"
+            " model. `KMeans` is particularly useful for large datasets."
+        ),
+        value=IndexAlgorithm.AUTO,
+        choices=list(IndexAlgorithm),
+    )
+    pretrained_type: DropdownConfig = DropdownConfig(
+        label="Pretrained model type",
+        info=(
+            "The type of pretrained model to finetune the voice model on. `None` will"
+            " train the voice model from scratch, while `Default` will use a pretrained"
+            " model tailored to the specific voice model architecture. `Custom` will"
+            " use a custom pretrained that you provide."
+        ),
+        value=PretrainedType.DEFAULT,
+        choices=list(PretrainedType),
+        exclude_value=True,
+    )
+    custom_pretrained_model: DropdownConfig = DropdownConfig(
+        label="Custom pretrained model",
+        info="Select a custom pretrained model to finetune from the dropdown.",
+        value=None,
+        visible=False,
+        render=False,
+        exclude_value=True,
+    )
+    save_interval: SliderConfig = SliderConfig(
+        label="Save interval",
+        info=(
+            "The epoch interval at which to to save voice model weights and"
+            " checkpoints. The best model weights are always saved regardless of this"
+            " setting."
+        ),
+        value=10,
+        minimum=1,
+        maximum=100,
+        step=1,
+    )
+    save_all_checkpoints: CheckboxConfig = CheckboxConfig(
+        label="Save all checkpoints",
+        info=(
+            "Whether to save a unique checkpoint at each save interval. If not enabled,"
+            " only the latest checkpoint will be saved at each interval."
+        ),
+        value=True,
+    )
+    save_all_weights: CheckboxConfig = CheckboxConfig(
+        label="Save all weights",
+        info=(
+            "Whether to save unique voice model weights at each save interval. If not"
+            " enabled, only the best voice model weights will be saved."
+        ),
+        value=True,
+    )
+    clear_saved_data: CheckboxConfig = CheckboxConfig(
+        label="Clear saved data",
+        info=(
+            "Whether to delete any existing training data associated with the voice"
+            " model before training commences. Enable this setting only if you are"
+            " training a new voice model from scratch or restarting training."
+        ),
+        value=False,
+    )
+    upload_model: CheckboxConfig = CheckboxConfig(
+        label="Upload voice model",
+        info=(
+            "Whether to automatically upload the trained voice model so that it can be"
+            " used for generation tasks within the Ultimate RVC app."
+        ),
+        value=False,
+        exclude_value=True,
+    )
+    upload_name: TextboxConfig = TextboxConfig(
+        label="Upload name",
+        info="The name to give the uploaded voice model.",
+        value=None,
+        visible=False,
+        exclude_value=True,
+    )
+    training_acceleration: DropdownConfig = DropdownConfig.hardware_acceleration()
+    training_gpus: DropdownConfig = DropdownConfig.gpu()
+    preload_dataset: CheckboxConfig = CheckboxConfig(
+        label="Preload dataset",
+        info=(
+            "Whether to preload all training data into GPU memory. This can improve"
+            " training speed but requires a lot of VRAM.<br><br>"
+        ),
+        value=True,
+    )
+    reduce_memory_usage: CheckboxConfig = CheckboxConfig(
+        label="Reduce memory usage",
+        info=(
+            "Whether to reduce VRAM usage at the cost of slower training speed by"
+            " enabling activation checkpointing. This is useful for GPUs with limited"
+            " memory (e.g., <6GB VRAM) or when training with a batch size larger than"
+            " what your GPU can normally accommodate."
+        ),
+        value=False,
+    )
+
+
+
 class SegmentSize(IntEnum):
     """Enumeration of segment sizes for audio separation."""