Upload 12 files

hifigan_speech/args.json

@@ -0,0 +1,280 @@
+{
+    "base_config": "egs/vocoder/gan/exp_config_base.json",
+    "dataset": [
+        "libritts",
+        "ljspeech",
+        "vctk",
+    ],
+    "exp_name": "hifigan_speech",
+    "inference": {
+        "batch_size": 1,
+    },
+    "model": {
+        "discriminators": [
+            "msd",
+            "mpd",
+            "mssbcqtd",
+            "msstftd",
+        ],
+        "generator": "hifigan",
+        "hifigan": {
+            "resblock": "1",
+            "resblock_dilation_sizes": [
+                [
+                    1,
+                    3,
+                    5,
+                ],
+                [
+                    1,
+                    3,
+                    5,
+                ],
+                [
+                    1,
+                    3,
+                    5,
+                ],
+            ],
+            "resblock_kernel_sizes": [
+                3,
+                5,
+                7,
+            ],
+            "upsample_initial_channel": 768,
+            "upsample_kernel_sizes": [
+                16,
+                8,
+                4,
+                4,
+                4,
+            ],
+            "upsample_rates": [
+                8,
+                4,
+                2,
+                2,
+                2,
+            ],
+        },
+        "mpd": {
+            "discriminator_channel_mult_factor": 1,
+            "discriminator_channel_multi": 1,
+            "mpd_reshapes": [
+                2,
+                3,
+                5,
+                7,
+                11,
+                17,
+                23,
+                37,
+            ],
+            "use_spectral_norm": false,
+        },
+        "mrd": {
+            "discriminator_channel_mult_factor": 1,
+            "mrd_override": false,
+            "resolutions": [
+                [
+                    1024,
+                    120,
+                    600,
+                ],
+                [
+                    2048,
+                    240,
+                    1200,
+                ],
+                [
+                    512,
+                    50,
+                    240,
+                ],
+            ],
+            "use_spectral_norm": false,
+        },
+        "mssbcqtd": {
+            "bins_per_octaves": [
+                24,
+                36,
+                48,
+            ],
+            "dilations": [
+                1,
+                2,
+                4,
+            ],
+            "filters": 32,
+            "filters_scale": 1,
+            "hop_lengths": [
+                512,
+                256,
+                256,
+            ],
+            "in_channels": 1,
+            "max_filters": 1024,
+            "n_octaves": [
+                9,
+                9,
+                9,
+            ],
+            "out_channels": 1,
+        },
+        "msstftd": {
+            "filters": 32,
+        },
+    },
+    "model_type": "GANVocoder",
+    "preprocess": {
+        "audio_dir": "audios",
+        "bits": 8,
+        "contentvec_dir": "contentvec",
+        "cut_mel_frame": 32,
+        "data_augment": false,
+        "dur_dir": "durs",
+        "duration_dir": "duration",
+        "emo2id": "emo2id.json",
+        "energy_dir": "energys",
+        "energy_extract_mode": "from_mel",
+        "energy_norm": false,
+        "extract_amplitude_phase": false,
+        "extract_audio": true,
+        "extract_contentvec_feature": false,
+        "extract_duration": false,
+        "extract_energy": false,
+        "extract_label": false,
+        "extract_linear_spec": false,
+        "extract_mcep": false,
+        "extract_mel": true,
+        "extract_mert_feature": false,
+        "extract_one_hot": false,
+        "extract_pitch": false,
+        "extract_uv": false,
+        "extract_wenet_feature": false,
+        "extract_whisper_feature": false,
+        "f0_max": 1100,
+        "f0_min": 50,
+        "file_lst": "file.lst",
+        "fmax": 12000,
+        "fmin": 0,
+        "hop_size": 256,
+        "imaginary_dir": "imaginarys",
+        "is_mu_law": false,
+        "lab_dir": "labs",
+        "label_dir": "labels",
+        "linear_dir": "linears",
+        "log_amplitude_dir": "log_amplitudes",
+        "mcep_dir": "mcep",
+        "mel_dir": "mels",
+        "mel_min_max_norm": false,
+        "min_level_db": -115,
+        "n_fft": 1024,
+        "n_mel": 100,
+        "num_silent_frames": 8,
+        "phase_dir": "phases",
+        "phone_seq_file": "phone_seq_file",
+        "pitch_bin": 256,
+        "pitch_dir": "pitches",
+        "pitch_extractor": "parselmouth",
+        "pitch_max": 1100.0,
+        "pitch_min": 50.0,
+        "pitch_norm": false,
+        "real_dir": "reals",
+        "ref_level_db": 20,
+        "sample_rate": 24000,
+        "spk2id": "singers.json",
+        "train_file": "train.json",
+        "trim_fft_size": 512,
+        "trim_hop_size": 128,
+        "trim_silence": false,
+        "trim_top_db": 30,
+        "trimmed_wav_dir": "trimmed_wavs",
+        "use_amplitude_phase": false,
+        "use_audio": true,
+        "use_dur": false,
+        "use_emoid": false,
+        "use_frame_duration": false,
+        "use_frame_energy": false,
+        "use_frame_pitch": false,
+        "use_lab": false,
+        "use_label": false,
+        "use_linear": false,
+        "use_log_scale_energy": false,
+        "use_log_scale_pitch": false,
+        "use_mel": true,
+        "use_min_max_norm_mel": false,
+        "use_one_hot": false,
+        "use_phn_seq": false,
+        "use_phone": false,
+        "use_phone_duration": false,
+        "use_phone_energy": false,
+        "use_phone_pitch": false,
+        "use_spkid": false,
+        "use_text": false,
+        "use_uv": false,
+        "use_wav": false,
+        "use_wenet": false,
+        "utt2emo": "utt2emo",
+        "utt2spk": "utt2spk",
+        "uv_dir": "uvs",
+        "valid_file": "test.json",
+        "wav_dir": "wavs",
+        "wenet_dir": "wenet",
+        "win_size": 1024,
+    },
+    "supported_model_type": [
+        "GANVocoder",
+        "Fastspeech2",
+        "DiffSVC",
+        "Transformer",
+        "EDM",
+        "CD",
+    ],
+    "train": {
+        "adamw": {
+            "adam_b1": 0.8,
+            "adam_b2": 0.99,
+            "lr": 0.0002,
+        },
+        "batch_size": 16,
+        "criterions": [
+            "feature",
+            "discriminator",
+            "generator",
+            "mel",
+        ],
+        "dataloader": {
+            "num_worker": 4,
+            "pin_memory": true,
+        },
+        "ddp": true,
+        "epochs": 50000,
+        "exponential_lr": {
+            "lr_decay": 0.999,
+        },
+        "gradient_accumulation_step": 1,
+        "keep_checkpoint_max": 5,
+        "max_epoch": 1000000,
+        "max_steps": 1000000,
+        "multi_speaker_training": false,
+        "random_seed": 114514,
+        "run_eval": [
+            true,
+        ],
+        "sampler": {
+            "drop_last": true,
+            "holistic_shuffle": true,
+        },
+        "save_checkpoint_stride": [
+            50,
+        ],
+        "save_checkpoints_steps": 10000,
+        "save_summary_steps": 500,
+        "total_training_steps": 50000,
+        "tracker": [
+            "tensorboard",
+        ],
+        "valid_interval": 10000,
+    },
+}

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/ckpts.json

@@ -0,0 +1,3 @@
+[
+    []
+]

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5fb0647436a074b05d2c0503c858d694bfd7eb4ac67be1597dcd0880c2bbe8b
+size 189448709

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/optimizer_1.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf5a1192dd52c5ed1849269b371631a42d89dd99348af14628eeb32936805cab
+size 767969069

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7097c8afa12e4f7f93c024ddde568e0288c1bc99de277e711d64972e7a38858c
+size 94703001

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/pytorch_model_1.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8274f13e7d8f0cbd2a45810f3d7ac184234b74eba622b3bccb4928e381507e2d
+size 118577541

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/pytorch_model_2.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34613c36c68bd754943919e0955ddbd48c2664fc2cefa14f2dc905c53079d63a
+size 263128113

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/pytorch_model_3.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc81b27297dfd94d9dd216c678faa1323d5a34c267707deb6969a1fb7d66b6fa
+size 1593185

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/pytorch_model_4.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d28f2b1a533da4ce377670cac75e77cb3855ac54b5ccb145fc551a855c896f3
+size 1169411

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/random_states_0.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33c4c0a3a856d84dd933bc33c2f13bc4f219fb6230ffafebcd9e562cf0431886
+size 15691

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/scheduler.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ead7b9a7d4eb95bd33a6dc062a4c59873f58e56d9b1c0ead06dd4594f05bfda
+size 563

hifigan_speech/checkpoint/epoch-0200_step-2841939_loss-38.503523/scheduler_1.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbca31a86f051d55ec6013b6fd1191538c7e699f1cd1aa25fd1e2d3cbe7ddf05
+size 631