From 9c4a006cff073065da7591ee4e6f83242419445e Mon Sep 17 00:00:00 2001 From: Ftps Date: Wed, 20 Dec 2023 22:20:52 +0700 Subject: [PATCH] add architecture of small-model --- rvc/lib/train/architecture/v1.yml | 57 +++++++++++++++++++++++++++++++ rvc/lib/train/architecture/v2.yml | 38 +++++++++++++++++++++ 2 files changed, 95 insertions(+) create mode 100644 rvc/lib/train/architecture/v1.yml create mode 100644 rvc/lib/train/architecture/v2.yml diff --git a/rvc/lib/train/architecture/v1.yml b/rvc/lib/train/architecture/v1.yml new file mode 100644 index 0000000..f0891bf --- /dev/null +++ b/rvc/lib/train/architecture/v1.yml @@ -0,0 +1,57 @@ +32k: + filter_length: 513, + a?: 32, + inter_channels: 192, + hidden_channels: 192, + filter_channels: 768, + n_heads: 2, + kernen_layersl_size: 6, + kernel_size: 3, + p_dropout: 0, + resblock: "1", + resblock_kernel_sizes: [3, 7, 11], + resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]], + upsample_rates: [10, 4, 2, 2, 2], + upsample_initial_channel: 512, + upsample_kernel_sizes: [16, 16, 4, 4, 4], + spk_embed_dim: 109, + gin_channels: 256, + sampling_rate: 32000, +40k: + filter_length: 1025, + a?: 32, # What? + inter_channels: 192, + hidden_channels: 192, + filter_channels: 768, + n_heads: 2, + kernen_layersl_size: 6, + kernel_size: 3, + p_dropout: 0, + resblock: "1", + resblock_kernel_sizes: [3, 7, 11], + resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]], + upsample_rates: [10, 10, 2, 2], + upsample_initial_channel: 512, + upsample_kernel_sizes: [16, 16, 4, 4], + spk_embed_dim: 109, + gin_channels: 256, + sampling_rate: 40000, +48k: + filter_length: 1025, + a?: 32, + inter_channels: 192, + hidden_channels: 192, + filter_channels: 768, + n_heads: 2, + kernen_layersl_size: 6, + kernel_size: 3, + p_dropout: 0, + resblock: "1", + resblock_kernel_sizes: [3, 7, 11], + resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]], + upsample_rates: [10, 6, 2, 2, 2], + upsample_initial_channel: 512, + upsample_kernel_sizes: [16, 16, 4, 4, 4], + spk_embed_dim: 109, + gin_channels: 256, + sampling_rate: 48000, diff --git a/rvc/lib/train/architecture/v2.yml b/rvc/lib/train/architecture/v2.yml new file mode 100644 index 0000000..88534d9 --- /dev/null +++ b/rvc/lib/train/architecture/v2.yml @@ -0,0 +1,38 @@ +32k: + filter_length: 513, + a?: 32, + inter_channels: 192, + hidden_channels: 192, + filter_channels: 768, + n_heads: 2, + kernen_layersl_size: 6, + kernel_size: 3, + p_dropout: 0, + resblock: "1", + resblock_kernel_sizes: [3, 7, 11], + resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]], + upsample_rates: [10, 4, 2, 2, 2], + upsample_initial_channel: 512, + upsample_kernel_sizes: [16, 16, 4, 4, 4], + spk_embed_dim: 109, + gin_channels: 256, + sampling_rate: 32000, +48k: + filter_length: 1025, + a?: 32, + inter_channels: 192, + hidden_channels: 192, + filter_channels: 768, + n_heads: 2, + kernen_layersl_size: 6, + kernel_size: 3, + p_dropout: 0, + resblock: "1", + resblock_kernel_sizes: [3, 7, 11], + resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]], + upsample_rates: [12, 10, 2, 2], + upsample_initial_channel: 512, + upsample_kernel_sizes: [24, 20, 4, 4], + spk_embed_dim: 109, + gin_channels: 256, + sampling_rate: 48000,