768VecOnnxExport (#328)

* Delete export_onnx.py

* Delete export_onnx_old.py

* Delete models_onnx_moess.py

* Support 768 Vec

* Add files via upload

* Support 768 Vec

Support 768 Vec

* Support 768 Vec Onnx Export

Support 768 Vec Onnx Export

infer_pack/models_onnx.py

@@ -61,7 +61,7 @@ class TextEncoder256(nn.Module):
         return m, logs, x_mask
 
 
-class TextEncoder256Sim(nn.Module):
+class TextEncoder768(nn.Module):
     def __init__(
         self,
         out_channels,
@@ -81,14 +81,14 @@ class TextEncoder256Sim(nn.Module):
         self.n_layers = n_layers
         self.kernel_size = kernel_size
         self.p_dropout = p_dropout
-        self.emb_phone = nn.Linear(256, hidden_channels)
+        self.emb_phone = nn.Linear(768, hidden_channels)
         self.lrelu = nn.LeakyReLU(0.1, inplace=True)
         if f0 == True:
             self.emb_pitch = nn.Embedding(256, hidden_channels)  # pitch 256
         self.encoder = attentions.Encoder(
             hidden_channels, filter_channels, n_heads, n_layers, kernel_size, p_dropout
         )
-        self.proj = nn.Conv1d(hidden_channels, out_channels, 1)
+        self.proj = nn.Conv1d(hidden_channels, out_channels * 2, 1)
 
     def forward(self, phone, pitch, lengths):
         if pitch == None:
@@ -102,8 +102,10 @@ class TextEncoder256Sim(nn.Module):
             x.dtype
         )
         x = self.encoder(x * x_mask, x_mask)
-        x = self.proj(x) * x_mask
-        return x, x_mask
+        stats = self.proj(x) * x_mask
+
+        m, logs = torch.split(stats, self.out_channels, dim=1)
+        return m, logs, x_mask
 
 
 class ResidualCouplingBlock(nn.Module):
@@ -527,7 +529,7 @@ sr2sr = {
 }
 
 
-class SynthesizerTrnMs256NSFsidO(nn.Module):
+class SynthesizerTrnMsNSFsidM(nn.Module):
     def __init__(
         self,
         spec_channels,
@@ -571,15 +573,26 @@ class SynthesizerTrnMs256NSFsidO(nn.Module):
         self.gin_channels = gin_channels
         # self.hop_length = hop_length#
         self.spk_embed_dim = spk_embed_dim
-        self.enc_p = TextEncoder256(
-            inter_channels,
-            hidden_channels,
-            filter_channels,
-            n_heads,
-            n_layers,
-            kernel_size,
-            p_dropout,
-        )
+        if self.gin_channels == 256:
+            self.enc_p = TextEncoder256(
+                inter_channels,
+                hidden_channels,
+                filter_channels,
+                n_heads,
+                n_layers,
+                kernel_size,
+                p_dropout,
+            )
+        else:
+            self.enc_p = TextEncoder768(
+                inter_channels,
+                hidden_channels,
+                filter_channels,
+                n_heads,
+                n_layers,
+                kernel_size,
+                p_dropout,
+            )
         self.dec = GeneratorNSF(
             inter_channels,
             resblock,
@@ -605,6 +618,7 @@ class SynthesizerTrnMs256NSFsidO(nn.Module):
             inter_channels, hidden_channels, 5, 1, 3, gin_channels=gin_channels
         )
         self.emb_g = nn.Embedding(self.spk_embed_dim, gin_channels)
+        self.speaker_map = None
         print("gin_channels:", gin_channels, "self.spk_embed_dim:", self.spk_embed_dim)
 
     def remove_weight_norm(self):
@@ -612,10 +626,24 @@ class SynthesizerTrnMs256NSFsidO(nn.Module):
         self.flow.remove_weight_norm()
         self.enc_q.remove_weight_norm()
 
-    def forward(self, phone, phone_lengths, pitch, nsff0, sid, max_len=None):
-        g = self.emb_g(sid).unsqueeze(-1)
+    def construct_spkmixmap(self, n_speaker):
+        self.speaker_map = torch.zeros((n_speaker, 1, 1, self.gin_channels))
+        for i in range(n_speaker):
+            self.speaker_map[i] = self.emb_g(torch.LongTensor([[i]]))
+        self.speaker_map = self.speaker_map.unsqueeze(0)
+
+    def forward(self, phone, phone_lengths, pitch, nsff0, g, rnd, max_len=None):
+        if self.speaker_map is not None:   # [N, S]  *  [S, B, 1, H]
+            g = g.reshape((g.shape[0], g.shape[1], 1, 1, 1))  # [N, S, B, 1, 1]
+            g = g * self.speaker_map  # [N, S, B, 1, H]
+            g = torch.sum(g, dim=1) # [N, 1, B, 1, H]
+            g = g.transpose(0, -1).transpose(0, -2).squeeze(0) # [B, H, N]
+        else:
+            g = g.unsqueeze(0)
+            g = self.emb_g(g).transpose(1,2)
+
         m_p, logs_p, x_mask = self.enc_p(phone, pitch, phone_lengths)
-        z_p = (m_p + torch.exp(logs_p) * torch.randn_like(m_p) * 0.66666) * x_mask
+        z_p = (m_p + torch.exp(logs_p) * rnd) * x_mask
         z = self.flow(z_p, x_mask, g=g, reverse=True)
         o = self.dec((z * x_mask)[:, :, :max_len], nsff0, g=g)
         return o
@@ -651,6 +679,36 @@ class MultiPeriodDiscriminator(torch.nn.Module):
         return y_d_rs, y_d_gs, fmap_rs, fmap_gs
 
 
+class MultiPeriodDiscriminatorV2(torch.nn.Module):
+    def __init__(self, use_spectral_norm=False):
+        super(MultiPeriodDiscriminatorV2, self).__init__()
+        # periods = [2, 3, 5, 7, 11, 17]
+        periods = [2, 3, 5, 7, 11, 17, 23, 37]
+
+        discs = [DiscriminatorS(use_spectral_norm=use_spectral_norm)]
+        discs = discs + [
+            DiscriminatorP(i, use_spectral_norm=use_spectral_norm) for i in periods
+        ]
+        self.discriminators = nn.ModuleList(discs)
+
+    def forward(self, y, y_hat):
+        y_d_rs = []  #
+        y_d_gs = []
+        fmap_rs = []
+        fmap_gs = []
+        for i, d in enumerate(self.discriminators):
+            y_d_r, fmap_r = d(y)
+            y_d_g, fmap_g = d(y_hat)
+            # for j in range(len(fmap_r)):
+            #     print(i,j,y.shape,y_hat.shape,fmap_r[j].shape,fmap_g[j].shape)
+            y_d_rs.append(y_d_r)
+            y_d_gs.append(y_d_g)
+            fmap_rs.append(fmap_r)
+            fmap_gs.append(fmap_g)
+
+        return y_d_rs, y_d_gs, fmap_rs, fmap_gs
+
+
 class DiscriminatorS(torch.nn.Module):
     def __init__(self, use_spectral_norm=False):
         super(DiscriminatorS, self).__init__()