Index

A

activation (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) add_blank_token() (speeq.data.tokenizers.BaseTokenizer method) add_eos_token (speeq.config.ASRDataConfig attribute) add_eos_token() (speeq.data.tokenizers.BaseTokenizer method) add_oov_token() (speeq.data.tokenizers.BaseTokenizer method) add_pad_token() (speeq.data.tokenizers.BaseTokenizer method) add_pos_enc() (in module speeq.utils.utils) add_sos_token (speeq.config.ASRDataConfig attribute)

add_sos_token() (speeq.data.tokenizers.BaseTokenizer method) add_token() (speeq.data.tokenizers.BaseTokenizer method) AddAndNorm (class in speeq.models.layers) address (speeq.config.DistConfig attribute) ASRDataConfig (class in speeq.config) att_kernel_size (speeq.models.templates.SpeechTransformerTemp attribute) att_out_channels (speeq.models.templates.SpeechTransformerTemp attribute) AudioLoader (class in speeq.data.processes)

B

backend (speeq.config.DistConfig attribute) backward_pass() (speeq.trainers.trainers.BaseDistTrainer method) (speeq.trainers.trainers.BaseTrainer method) BaseDistTrainer (class in speeq.trainers.trainers) BaseSchedulerTemplate (class in speeq.trainers.templates) BaseTemplate (class in speeq.models.templates) BaseTokenizer (class in speeq.data.tokenizers) BaseTrainer (class in speeq.trainers.trainers) BasicAttSeq2SeqRNN (class in speeq.models.seq2seq) BasicAttSeq2SeqRNNTemp (class in speeq.models.templates) batch_detokenizer() (speeq.data.tokenizers.BaseTokenizer method) batch_size (speeq.config.TrainerConfig attribute)

batch_tokenizer() (speeq.data.tokenizers.BaseTokenizer method) BERT (class in speeq.models.ctc) BERTTemp (class in speeq.models.templates) bidirectional (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) blank (speeq.trainers.criterions.CTCLoss attribute) block_repetition (speeq.models.templates.QuartzNetTemp attribute) blocks_kernel_size (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute)

C

calc_data_len() (in module speeq.utils.utils) CausalVGGBlock (class in speeq.models.layers) channel_inc (speeq.models.templates.JasperTemp attribute) channels_size (speeq.models.templates.QuartzNetTemp attribute) CharTokenizer (class in speeq.data.tokenizers) clear() (in module speeq.utils.utils) clear_screen (speeq.config.TrainerConfig attribute) Conformer (class in speeq.models.ctc) ConformerBlock (class in speeq.models.layers) ConformerConvModule (class in speeq.models.layers) ConformerCTCTemp (class in speeq.models.templates) ConformerEncoder (class in speeq.models.encoders) ConformerFeedForward (class in speeq.models.layers) ConformerPreNet (class in speeq.models.layers) ConformerRelativeMHSA (class in speeq.models.layers) ConformerTransducer (class in speeq.models.transducers) ConformerTransducerTemp (class in speeq.models.templates) ConsistentAttenuator (class in speeq.data.augmenters)

ContextNet (class in speeq.models.transducers) ContextNetBlock (class in speeq.models.layers) ContextNetConvLayer (class in speeq.models.layers) ContextNetEncoder (class in speeq.models.encoders) ContextNetResidual (class in speeq.models.layers) ContextNetTemp (class in speeq.models.templates) Conv1DLayers (class in speeq.models.layers) ConvPredModule (class in speeq.models.layers) CReLu (class in speeq.models.activations) criterion (speeq.config.TrainerConfig attribute) criterion_args (speeq.config.TrainerConfig attribute) CrossEntropyLoss (class in speeq.trainers.criterions) CSVDataset (class in speeq.data.loaders) CTCLoss (class in speeq.trainers.criterions) CTCModel (class in speeq.models.ctc) CTCModelBuilderTemp (class in speeq.models.templates) CTCPredictor (class in speeq.predictors) CTCSkeleton (class in speeq.models.skeletons) CTCTrainer (class in speeq.trainers.trainers)

D

d_model (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) (speeq.trainers.templates.NoamSchedulerTemp attribute) dec_left_size (speeq.models.templates.TransformerTransducerTemp attribute) dec_num_layers (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) dec_right_size (speeq.models.templates.TransformerTransducerTemp attribute) decay_rate (speeq.trainers.templates.SqueezeformerNoamSchedulerTemp attribute)

decoder (speeq.models.templates.Seq2SeqBuilderTemp attribute) (speeq.models.templates.TransducerBuilderTemp attribute) DeepSpeechV1 (class in speeq.models.ctc) DeepSpeechV1Encoder (class in speeq.models.encoders) DeepSpeechV1Temp (class in speeq.models.templates) DeepSpeechV2 (class in speeq.models.ctc) DeepSpeechV2Encoder (class in speeq.models.encoders) DeepSpeechV2Temp (class in speeq.models.templates) device (speeq.config.TrainerConfig attribute) dist_config (speeq.config.TrainerConfig attribute) DistConfig (class in speeq.config) DistCTCTrainer (class in speeq.trainers.trainers) DistSeq2SeqTrainer (class in speeq.trainers.trainers) DistTransducerTrainer (class in speeq.trainers.trainers) DynamicPadder (class in speeq.data.padders)

E

emb_dim (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) embed() (speeq.models.ctc.BERT method) enc_left_size (speeq.models.templates.TransformerTransducerTemp attribute) enc_num_layers (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute)

enc_right_size (speeq.models.templates.TransformerTransducerTemp attribute) encoder (speeq.models.templates.CTCModelBuilderTemp attribute) (speeq.models.templates.Seq2SeqBuilderTemp attribute) (speeq.models.templates.TransducerBuilderTemp attribute) epilog_channel_size (speeq.models.templates.QuartzNetTemp attribute) epilog_kernel_size (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) epochs (speeq.config.TrainerConfig attribute) execute() (speeq.data.processors.OrderedProcessor method) (speeq.data.processors.SpeechProcessor method) (speeq.data.processors.StochasticProcessor method)

F

feat_size (speeq.models.templates.CTCModelBuilderTemp attribute) (speeq.models.templates.TransducerBuilderTemp attribute) FeatExtractor (class in speeq.data.processes) FeatStacker (class in speeq.data.processes) FeedForwardModule (class in speeq.models.layers) ff_expansion_factor (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) ff_size (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) fit() (speeq.trainers.trainers.BaseDistTrainer method) (speeq.trainers.trainers.BaseTrainer method) forward() (speeq.models.activations.CReLu method) (speeq.models.activations.Sigmax method) (speeq.models.ctc.BERT method) (speeq.models.ctc.CTCModel method) (speeq.models.decoders.GlobAttRNNDecoder method) (speeq.models.decoders.LocationAwareAttDecoder method) (speeq.models.decoders.SpeechTransformerDecoder method) (speeq.models.decoders.TransducerRNNDecoder method) (speeq.models.decoders.TransformerDecoder method) (speeq.models.decoders.TransformerTransducerDecoder method) (speeq.models.encoders.ConformerEncoder method) (speeq.models.encoders.ContextNetEncoder method) (speeq.models.encoders.DeepSpeechV1Encoder method) (speeq.models.encoders.DeepSpeechV2Encoder method) (speeq.models.encoders.JasperEncoder method) (speeq.models.encoders.PyramidRNNEncoder method) (speeq.models.encoders.RNNEncoder method) (speeq.models.encoders.SpeechTransformerEncoder method) (speeq.models.encoders.SqueezeformerEncoder method) (speeq.models.encoders.TransformerTransducerEncoder method) (speeq.models.encoders.VGGTransformerEncoder method) (speeq.models.encoders.Wav2LetterEncoder method) (speeq.models.layers.AddAndNorm method) (speeq.models.layers.CausalVGGBlock method) (speeq.models.layers.ConformerBlock method) (speeq.models.layers.ConformerConvModule method) (speeq.models.layers.ConformerFeedForward method) (speeq.models.layers.ConformerPreNet method) (speeq.models.layers.ConformerRelativeMHSA method) (speeq.models.layers.ContextNetBlock method) (speeq.models.layers.ContextNetConvLayer method) (speeq.models.layers.ContextNetResidual method) (speeq.models.layers.Conv1DLayers method) (speeq.models.layers.ConvPredModule method) (speeq.models.layers.FeedForwardModule method) (speeq.models.layers.GlobalMulAttention method) (speeq.models.layers.GroupsShuffle method) (speeq.models.layers.JasperBlock method) (speeq.models.layers.JasperBlocks method) (speeq.models.layers.JasperResidual method) (speeq.models.layers.JasperSubBlock method) (speeq.models.layers.LocAwareGlobalAddAttention method) (speeq.models.layers.MaskedMultiHeadAtt method) (speeq.models.layers.MultiHeadAtt method) (speeq.models.layers.MultiHeadAtt2d method) (speeq.models.layers.PackedRNN method) (speeq.models.layers.PositionalEmbedding method) (speeq.models.layers.PredModule method) (speeq.models.layers.QuartzSubBlock method) (speeq.models.layers.RowConv1D method) (speeq.models.layers.Scaling1d method) (speeq.models.layers.SpeechTransformerDecLayer method) (speeq.models.layers.SpeechTransformerEncLayer method) (speeq.models.layers.SqueezeAndExcit1D method) (speeq.models.layers.SqueezeformerBlock method) (speeq.models.layers.SqueezeformerConvModule method) (speeq.models.layers.SqueezeformerFeedForward method) (speeq.models.layers.SqueezeformerRelativeMHSA method) (speeq.models.layers.TransformerDecLayer method) (speeq.models.layers.TransformerEncLayer method) (speeq.models.layers.TransformerEncLayerWithAttTruncation method) (speeq.models.layers.TransformerTransducerLayer method) (speeq.models.layers.TruncatedRelativeMHSA method) (speeq.models.layers.TruncatedSelfAttention method) (speeq.models.layers.VGGTransformerPreNet method) (speeq.models.seq2seq.BasicAttSeq2SeqRNN method) (speeq.models.seq2seq.SpeechTransformer method) (speeq.models.skeletons.Seq2SeqSkeleton method) (speeq.models.transducers.TransformerTransducer method) (speeq.trainers.criterions.CrossEntropyLoss method) (speeq.trainers.criterions.NLLLoss method) (speeq.trainers.criterions.RNNTLoss method)

forward_pass() (speeq.trainers.trainers.CTCTrainer method) (speeq.trainers.trainers.Seq2SeqTrainer method) (speeq.trainers.trainers.TransducerTrainer method) FrameContextualizer (class in speeq.data.processes) FrequencyMasking (class in speeq.data.augmenters) func() (speeq.data.augmenters.FrequencyMasking method) (speeq.data.augmenters.Reverberation method) (speeq.data.augmenters.TimeMasking method) (speeq.data.augmenters.VariableAttenuator method) (speeq.data.augmenters.VolumeChanger method) (speeq.data.augmenters.WhiteNoiseInjector method) (speeq.data.processes.StochasticProcess method)

G

get_asr_datasets() (in module speeq.data.registry) get_asr_loaders() (in module speeq.data.registry) get_asr_trainer() (in module speeq.trainers.registry) get_batch() (speeq.data.loaders.SpeechTextLoader method) get_criterion() (in module speeq.trainers.registry) get_dict() (speeq.models.templates.BaseTemplate method) (speeq.trainers.templates.BaseSchedulerTemplate method) get_flatten_results() (in module speeq.trainers.criterions) get_key_tag() (in module speeq.utils.utils) get_logger() (in module speeq.utils.loggers) get_looking_ahead_mask() (speeq.models.layers.MaskedMultiHeadAtt method) (speeq.models.layers.TruncatedSelfAttention method) get_lr() (speeq.trainers.schedulers.NoamScheduler method) (speeq.trainers.schedulers.SqueezeformerNoamScheduler method) get_mask_from_lens() (in module speeq.utils.utils) get_model() (in module speeq.models.registry)

get_optimizer() (in module speeq.trainers.registry) get_pad_mask() (in module speeq.utils.utils) get_positional_encoding() (in module speeq.utils.utils) get_speech_padder() (in module speeq.data.registry) get_state_dict() (in module speeq.utils.utils) get_text_list() (in module speeq.utils.utils) get_text_padder() (in module speeq.data.registry) get_tokenizer() (in module speeq.data.registry) get_tokens() (speeq.data.tokenizers.CharTokenizer method) (speeq.data.tokenizers.WordTokenizer method) GlobalMulAttention (class in speeq.models.layers) GlobAttRNNDecoder (class in speeq.models.decoders) grad_acc_steps (speeq.config.TrainerConfig attribute) grad_clip_norm_type (speeq.config.TrainerConfig attribute) grad_clip_thresh (speeq.config.TrainerConfig attribute) groups (speeq.models.templates.QuartzNetTemp attribute) GroupsShuffle (class in speeq.models.layers)

H

h (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute)

has_bnorm() (in module speeq.utils.utils) hidden_size (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute)

I

ids2tokens() (speeq.data.tokenizers.BaseTokenizer method) ignore_index (speeq.trainers.criterions.CrossEntropyLoss attribute) (speeq.trainers.criterions.NLLLoss attribute) ignore_optim_state (speeq.config.TrainerConfig attribute) in_features (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) (speeq.models.templates.Wav2LetterTemp attribute)

init_dist() (speeq.trainers.trainers.BaseDistTrainer method) inline_log() (speeq.trainers.trainers.BaseTrainer method) inv_temperature (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) is_master (speeq.trainers.trainers.BaseDistTrainer property) (speeq.trainers.trainers.BaseTrainer property)

J

Jasper (class in speeq.models.ctc) JasperBlock (class in speeq.models.layers) JasperBlocks (class in speeq.models.layers) JasperEncoder (class in speeq.models.encoders) JasperResidual (class in speeq.models.layers)

JasperSubBlock (class in speeq.models.layers) JasperTemp (class in speeq.models.templates) join_net (speeq.models.templates.TransducerBuilderTemp attribute) joint_size (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute)

K

kernel_size (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute)

kernel_sizes_per_vgg_block (speeq.models.templates.VGGTransformerTransducerTemp attribute)

L

label_smoothing (speeq.trainers.criterions.CrossEntropyLoss attribute) LAS (class in speeq.models.seq2seq) LASTemp (class in speeq.models.templates) launch_training_job() (in module speeq.trainers.trainers) layers_channels_size (speeq.models.templates.Wav2LetterTemp attribute) layers_kernel_size (speeq.models.templates.Wav2LetterTemp attribute) left_size (speeq.models.templates.VGGTransformerTransducerTemp attribute) list_ctc_models() (in module speeq.models.registry) list_seq2seq_models() (in module speeq.models.registry) list_transducer_models() (in module speeq.models.registry) load_csv() (in module speeq.utils.utils) load_json() (in module speeq.utils.utils) load_state_dict() (in module speeq.utils.utils) (speeq.trainers.schedulers.Scheduler method)

load_text() (in module speeq.utils.utils) load_tokenizer() (in module speeq.data.registry) (speeq.data.tokenizers.BaseTokenizer method) load_tokenizer_from_dict() (speeq.data.tokenizers.BaseTokenizer method) LocationAwareAttDecoder (class in speeq.models.decoders) LocAwareGlobalAddAttention (class in speeq.models.layers) log() (speeq.utils.loggers.TBLogger method) log_step() (speeq.utils.loggers.TBLogger method) log_steps_frequency (speeq.config.TrainerConfig attribute) logdir (speeq.config.TrainerConfig attribute) logger (speeq.config.TrainerConfig attribute) lr_peak (speeq.trainers.templates.SqueezeformerNoamSchedulerTemp attribute)

M

MaskedMultiHeadAtt (class in speeq.models.layers) masking_value (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) max_clip_value (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) max_len (speeq.models.templates.BERTTemp attribute) model_path (speeq.config.ModelConfig attribute) ModelConfig (class in speeq.config) module speeq.config speeq.data.augmenters speeq.data.loaders speeq.data.padders speeq.data.processes speeq.data.processors speeq.data.registry speeq.data.tokenizers speeq.models.activations speeq.models.ctc speeq.models.decoders speeq.models.encoders speeq.models.layers speeq.models.registry speeq.models.seq2seq speeq.models.skeletons speeq.models.templates speeq.models.transducers speeq.predictors speeq.trainers.criterions speeq.trainers.registry speeq.trainers.schedulers speeq.trainers.templates speeq.trainers.trainers speeq.utils.loggers speeq.utils.utils

MultiHeadAtt (class in speeq.models.layers) MultiHeadAtt2d (class in speeq.models.layers)

N

n (speeq.models.templates.SqueezeformerCTCTemp attribute) n_channels_per_vgg_block (speeq.models.templates.VGGTransformerTransducerTemp attribute) n_conf_layers (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) n_conv (speeq.models.templates.DeepSpeechV2Temp attribute) n_conv_layers (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.Wav2LetterTemp attribute) n_conv_layers_per_vgg_block (speeq.models.templates.VGGTransformerTransducerTemp attribute) n_dec_layers (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) n_enc_layers (speeq.models.templates.SpeechTransformerTemp attribute) n_gpus (speeq.config.DistConfig attribute) n_layers (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute)

n_linear_layers (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) n_logs (speeq.config.TrainerConfig attribute) n_rnn (speeq.models.templates.DeepSpeechV2Temp attribute) n_sub_layers (speeq.models.templates.ContextNetTemp attribute) n_vgg_blocks (speeq.models.templates.VGGTransformerTransducerTemp attribute) name (speeq.config.TrainerConfig attribute) (speeq.models.templates.BaseTemplate property) (speeq.trainers.templates.BaseSchedulerTemplate property) NLLLoss (class in speeq.trainers.criterions) NoamScheduler (class in speeq.trainers.schedulers) NoamSchedulerTemp (class in speeq.trainers.templates) num_blocks (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) num_sub_blocks (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute)

O

optim_args (speeq.config.TrainerConfig attribute) optimizer (speeq.config.TrainerConfig attribute)

OrderedProcessor (class in speeq.data.processors) out_channels (speeq.models.templates.ContextNetTemp attribute) outdir (speeq.config.TrainerConfig attribute)

P

p_dropout (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.BERTTemp attribute) (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute) (speeq.models.templates.Wav2LetterTemp attribute) PackedGRU (class in speeq.models.layers) PackedLSTM (class in speeq.models.layers) PackedRNN (class in speeq.models.layers) pad() (speeq.data.padders.DynamicPadder method) (speeq.data.padders.StaticPadder method) padding_type (speeq.config.ASRDataConfig attribute) perform_attention() (speeq.models.layers.MultiHeadAtt method) perform_frequency_attention() (speeq.models.layers.MultiHeadAtt2d method) pooling_kernel_size (speeq.models.templates.SqueezeformerCTCTemp attribute) pooling_stride (speeq.models.templates.SqueezeformerCTCTemp attribute) port (speeq.config.DistConfig attribute) PositionalEmbedding (class in speeq.models.layers) post_conv_channels_size (speeq.models.templates.Wav2LetterTemp attribute)

post_conv_kernel_size (speeq.models.templates.Wav2LetterTemp attribute) pre_conv_kernel_size (speeq.models.templates.Wav2LetterTemp attribute) pre_conv_stride (speeq.models.templates.Wav2LetterTemp attribute) pred_activation (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) pred_net (speeq.models.templates.CTCModelBuilderTemp attribute) predict() (speeq.models.ctc.DeepSpeechV1 method) (speeq.models.decoders.GlobAttRNNDecoder method) (speeq.models.decoders.LocationAwareAttDecoder method) (speeq.models.decoders.SpeechTransformerDecoder method) (speeq.models.decoders.TransducerRNNDecoder method) (speeq.models.decoders.TransformerDecoder method) (speeq.models.seq2seq.BasicAttSeq2SeqRNN method) (speeq.models.seq2seq.SpeechTransformer method) (speeq.predictors.CTCPredictor method) (speeq.predictors.Seq2SeqPredictor method) (speeq.predictors.TransducerPredictor method) PredModule (class in speeq.models.layers) prelog_kernel_size (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) prelog_n_channels (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) prelog_stride (speeq.models.templates.JasperTemp attribute) (speeq.models.templates.QuartzNetTemp attribute) preprocess_tokens() (speeq.data.tokenizers.CharTokenizer method) (speeq.data.tokenizers.WordTokenizer method) PyramidRNNEncoder (class in speeq.models.encoders)

Q

QuartzBlock (class in speeq.models.layers) QuartzBlocks (class in speeq.models.layers) QuartzNet (class in speeq.models.ctc)

QuartzNetEncoder (class in speeq.models.encoders) QuartzNetTemp (class in speeq.models.templates) QuartzSubBlock (class in speeq.models.layers)

R

reduction_factor (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.LASTemp attribute) remove_positionals() (in module speeq.trainers.criterions) res_scaling (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) Reverberation (class in speeq.data.augmenters) reverse (speeq.config.ASRDataConfig attribute) right_size (speeq.models.templates.VGGTransformerTransducerTemp attribute) rnn_type (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.DeepSpeechV1Temp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNTTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) (speeq.models.templates.VGGTransformerTransducerTemp attribute)

RNNEncoder (class in speeq.models.encoders) RNNTLoss (class in speeq.trainers.criterions) RNNTransducer (class in speeq.models.transducers) RNNTTemp (class in speeq.models.templates) RNNWithLocationAwareAtt (class in speeq.models.seq2seq) RNNWithLocationAwareAttTemp (class in speeq.models.templates) RowConv1D (class in speeq.models.layers) run() (speeq.data.processes.AudioLoader method) (speeq.data.processes.FeatExtractor method) (speeq.data.processes.FeatStacker method) (speeq.data.processes.FrameContextualizer method) (speeq.data.processes.StochasticProcess method)

S

save_json() (in module speeq.utils.utils) save_state_dict() (in module speeq.utils.utils) save_text() (in module speeq.utils.utils) save_tokenizer() (speeq.data.tokenizers.BaseTokenizer method) Scaling1d (class in speeq.models.layers) Scheduler (class in speeq.trainers.schedulers) scheduler_template (speeq.config.TrainerConfig attribute) sep (speeq.config.ASRDataConfig attribute) Seq2SeqBuilderTemp (class in speeq.models.templates) Seq2SeqPredictor (class in speeq.predictors) Seq2SeqSkeleton (class in speeq.models.skeletons) Seq2SeqTrainer (class in speeq.trainers.trainers) set_state_dict() (in module speeq.utils.utils) set_tokenizer() (speeq.data.tokenizers.BaseTokenizer method) shuffle (speeq.config.ASRDataConfig attribute) Sigmax (class in speeq.models.activations) sort_key (speeq.config.ASRDataConfig attribute) speech_key (speeq.config.ASRDataConfig attribute) speech_pad_max_len (speeq.config.ASRDataConfig attribute) speech_processor (speeq.config.ASRDataConfig attribute) SpeechProcessor (class in speeq.data.processors) SpeechTextDataset (class in speeq.data.loaders) SpeechTextLoader (class in speeq.data.loaders) SpeechTransformer (class in speeq.models.seq2seq) SpeechTransformerDecLayer (class in speeq.models.layers) SpeechTransformerDecoder (class in speeq.models.decoders) SpeechTransformerEncLayer (class in speeq.models.layers) SpeechTransformerEncoder (class in speeq.models.encoders) SpeechTransformerTemp (class in speeq.models.templates) speeq.config module speeq.data.augmenters module speeq.data.loaders module speeq.data.padders module speeq.data.processes module speeq.data.processors module speeq.data.registry module speeq.data.tokenizers module speeq.models.activations module speeq.models.ctc module speeq.models.decoders module speeq.models.encoders module speeq.models.layers module speeq.models.registry module

speeq.models.seq2seq module speeq.models.skeletons module speeq.models.templates module speeq.models.transducers module speeq.predictors module speeq.trainers.criterions module speeq.trainers.registry module speeq.trainers.schedulers module speeq.trainers.templates module speeq.trainers.trainers module speeq.utils.loggers module speeq.utils.utils module SqueezeAndExcit1D (class in speeq.models.layers) Squeezeformer (class in speeq.models.ctc) SqueezeformerBlock (class in speeq.models.layers) SqueezeformerConvModule (class in speeq.models.layers) SqueezeformerCTCTemp (class in speeq.models.templates) SqueezeformerEncoder (class in speeq.models.encoders) SqueezeformerFeedForward (class in speeq.models.layers) SqueezeformerNoamScheduler (class in speeq.trainers.schedulers) SqueezeformerNoamSchedulerTemp (class in speeq.trainers.templates) SqueezeformerRelativeMHSA (class in speeq.models.layers) ss_groups (speeq.models.templates.SqueezeformerCTCTemp attribute) ss_kernel_size (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) ss_n_conv_layers (speeq.models.templates.SqueezeformerCTCTemp attribute) ss_num_conv_layers (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) ss_stride (speeq.models.templates.ConformerCTCTemp attribute) (speeq.models.templates.ConformerTransducerTemp attribute) (speeq.models.templates.SqueezeformerCTCTemp attribute) state_dict() (speeq.trainers.schedulers.NoamScheduler method) (speeq.trainers.schedulers.Scheduler method) (speeq.trainers.schedulers.SqueezeformerNoamScheduler method) StaticPadder (class in speeq.data.padders) step() (speeq.trainers.schedulers.Scheduler method) StochasticProcess (class in speeq.data.processes) StochasticProcessor (class in speeq.data.processors) stride (speeq.models.templates.ContextNetTemp attribute) (speeq.models.templates.DeepSpeechV2Temp attribute) (speeq.models.templates.SpeechTransformerTemp attribute) (speeq.models.templates.TransformerTransducerTemp attribute)

T

t_peak (speeq.trainers.templates.SqueezeformerNoamSchedulerTemp attribute) tau (speeq.models.templates.DeepSpeechV2Temp attribute) TBLogger (class in speeq.utils.loggers) teacher_forcing_rate (speeq.models.templates.BasicAttSeq2SeqRNNTemp attribute) (speeq.models.templates.LASTemp attribute) (speeq.models.templates.RNNWithLocationAwareAttTemp attribute) template (speeq.config.ModelConfig attribute) test() (speeq.trainers.trainers.BaseTrainer method) testing_path (speeq.config.ASRDataConfig attribute) text_key (speeq.config.ASRDataConfig attribute) text_pad_max_len (speeq.config.ASRDataConfig attribute) text_processor (speeq.config.ASRDataConfig attribute) TimeMasking (class in speeq.data.augmenters) tokenize() (speeq.data.tokenizers.BaseTokenizer method) tokenizer_path (speeq.config.ASRDataConfig attribute) tokenizer_type (speeq.config.ASRDataConfig attribute) train() (speeq.trainers.trainers.BaseDistTrainer method) (speeq.trainers.trainers.BaseTrainer method) train_step() (speeq.trainers.trainers.BaseTrainer method) TrainerConfig (class in speeq.config) training (speeq.models.activations.CReLu attribute) (speeq.models.activations.Sigmax attribute) (speeq.models.ctc.BERT attribute) (speeq.models.ctc.Conformer attribute) (speeq.models.ctc.CTCModel attribute) (speeq.models.ctc.DeepSpeechV1 attribute) (speeq.models.ctc.DeepSpeechV2 attribute) (speeq.models.ctc.Jasper attribute) (speeq.models.ctc.QuartzNet attribute) (speeq.models.ctc.Squeezeformer attribute) (speeq.models.ctc.Wav2Letter attribute) (speeq.models.decoders.GlobAttRNNDecoder attribute) (speeq.models.decoders.LocationAwareAttDecoder attribute) (speeq.models.decoders.SpeechTransformerDecoder attribute) (speeq.models.decoders.TransducerRNNDecoder attribute) (speeq.models.decoders.TransformerDecoder attribute) (speeq.models.decoders.TransformerTransducerDecoder attribute) (speeq.models.encoders.ConformerEncoder attribute) (speeq.models.encoders.ContextNetEncoder attribute) (speeq.models.encoders.DeepSpeechV1Encoder attribute) (speeq.models.encoders.DeepSpeechV2Encoder attribute) (speeq.models.encoders.JasperEncoder attribute) (speeq.models.encoders.PyramidRNNEncoder attribute) (speeq.models.encoders.QuartzNetEncoder attribute) (speeq.models.encoders.RNNEncoder attribute) (speeq.models.encoders.SpeechTransformerEncoder attribute) (speeq.models.encoders.SqueezeformerEncoder attribute) (speeq.models.encoders.TransformerTransducerEncoder attribute) (speeq.models.encoders.VGGTransformerEncoder attribute) (speeq.models.encoders.Wav2LetterEncoder attribute) (speeq.models.layers.AddAndNorm attribute) (speeq.models.layers.CausalVGGBlock attribute) (speeq.models.layers.ConformerBlock attribute) (speeq.models.layers.ConformerConvModule attribute) (speeq.models.layers.ConformerFeedForward attribute) (speeq.models.layers.ConformerPreNet attribute) (speeq.models.layers.ConformerRelativeMHSA attribute) (speeq.models.layers.ContextNetBlock attribute) (speeq.models.layers.ContextNetConvLayer attribute) (speeq.models.layers.ContextNetResidual attribute) (speeq.models.layers.Conv1DLayers attribute) (speeq.models.layers.ConvPredModule attribute) (speeq.models.layers.FeedForwardModule attribute) (speeq.models.layers.GlobalMulAttention attribute) (speeq.models.layers.GroupsShuffle attribute) (speeq.models.layers.JasperBlock attribute) (speeq.models.layers.JasperBlocks attribute) (speeq.models.layers.JasperResidual attribute) (speeq.models.layers.JasperSubBlock attribute) (speeq.models.layers.LocAwareGlobalAddAttention attribute) (speeq.models.layers.MaskedMultiHeadAtt attribute) (speeq.models.layers.MultiHeadAtt attribute) (speeq.models.layers.MultiHeadAtt2d attribute) (speeq.models.layers.PackedGRU attribute) (speeq.models.layers.PackedLSTM attribute) (speeq.models.layers.PackedRNN attribute) (speeq.models.layers.PositionalEmbedding attribute) (speeq.models.layers.PredModule attribute) (speeq.models.layers.QuartzBlock attribute) (speeq.models.layers.QuartzBlocks attribute) (speeq.models.layers.QuartzSubBlock attribute) (speeq.models.layers.RowConv1D attribute) (speeq.models.layers.Scaling1d attribute) (speeq.models.layers.SpeechTransformerDecLayer attribute) (speeq.models.layers.SpeechTransformerEncLayer attribute) (speeq.models.layers.SqueezeAndExcit1D attribute) (speeq.models.layers.SqueezeformerBlock attribute) (speeq.models.layers.SqueezeformerConvModule attribute) (speeq.models.layers.SqueezeformerFeedForward attribute) (speeq.models.layers.SqueezeformerRelativeMHSA attribute) (speeq.models.layers.TransformerDecLayer attribute) (speeq.models.layers.TransformerEncLayer attribute) (speeq.models.layers.TransformerEncLayerWithAttTruncation attribute) (speeq.models.layers.TransformerTransducerLayer attribute) (speeq.models.layers.TruncatedRelativeMHSA attribute) (speeq.models.layers.TruncatedSelfAttention attribute) (speeq.models.layers.VGGTransformerPreNet attribute) (speeq.models.seq2seq.BasicAttSeq2SeqRNN attribute) (speeq.models.seq2seq.LAS attribute) (speeq.models.seq2seq.RNNWithLocationAwareAtt attribute) (speeq.models.seq2seq.SpeechTransformer attribute) (speeq.models.skeletons.CTCSkeleton attribute) (speeq.models.skeletons.Seq2SeqSkeleton attribute) (speeq.models.skeletons.TransducerSkeleton attribute) (speeq.models.transducers.ConformerTransducer attribute) (speeq.models.transducers.ContextNet attribute) (speeq.models.transducers.RNNTransducer attribute) (speeq.models.transducers.TransformerTransducer attribute) (speeq.models.transducers.VGGTransformerTransducer attribute) (speeq.trainers.criterions.RNNTLoss attribute)

training_path (speeq.config.ASRDataConfig attribute) TransducerBuilderTemp (class in speeq.models.templates) TransducerPredictor (class in speeq.predictors) TransducerRNNDecoder (class in speeq.models.decoders) TransducerSkeleton (class in speeq.models.skeletons) TransducerTrainer (class in speeq.trainers.trainers) TransformerDecLayer (class in speeq.models.layers) TransformerDecoder (class in speeq.models.decoders) TransformerEncLayer (class in speeq.models.layers) TransformerEncLayerWithAttTruncation (class in speeq.models.layers) TransformerTransducer (class in speeq.models.transducers) TransformerTransducerDecoder (class in speeq.models.decoders) TransformerTransducerEncoder (class in speeq.models.encoders) TransformerTransducerLayer (class in speeq.models.layers) TransformerTransducerTemp (class in speeq.models.templates) truncate_attention_mask() (in module speeq.utils.utils) TruncatedRelativeMHSA (class in speeq.models.layers) TruncatedSelfAttention (class in speeq.models.layers) type (speeq.config.ASRDataConfig attribute) (speeq.models.templates.BaseTemplate property) (speeq.trainers.templates.BaseSchedulerTemplate property)

U

use_blank_as_pad (speeq.config.ASRDataConfig attribute)

V

VariableAttenuator (class in speeq.data.augmenters) vgg_pooling_kernel_size (speeq.models.templates.VGGTransformerTransducerTemp attribute) VGGTransformerEncoder (class in speeq.models.encoders) VGGTransformerPreNet (class in speeq.models.layers)

VGGTransformerTransducer (class in speeq.models.transducers) VGGTransformerTransducerTemp (class in speeq.models.templates) vocab_size (speeq.data.tokenizers.BaseTokenizer property) VolumeChanger (class in speeq.data.augmenters)

W

warmup_staps (speeq.trainers.templates.NoamSchedulerTemp attribute) (speeq.trainers.templates.SqueezeformerNoamSchedulerTemp attribute) Wav2Letter (class in speeq.models.ctc) Wav2LetterEncoder (class in speeq.models.encoders)

Wav2LetterTemp (class in speeq.models.templates) wav_kernel_size (speeq.models.templates.Wav2LetterTemp attribute) wav_stride (speeq.models.templates.Wav2LetterTemp attribute) WhiteNoiseInjector (class in speeq.data.augmenters) WordTokenizer (class in speeq.data.tokenizers)

Z

zero_grad() (speeq.trainers.schedulers.Scheduler method)

zero_infinity (speeq.trainers.criterions.CTCLoss attribute)