MCPcopy Create free account
hub / github.com/0xShug0/audio.cpp / load_conv1d

Function load_conv1d

src/models/vibevoice/tokenizer_audio.cpp:160–172  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

158}
159
160modules::Conv1dWeights load_conv1d(
161 core::BackendWeightStore & store,
162 const assets::TensorSource & source,
163 const std::string & prefix,
164 int64_t in_channels,
165 int64_t out_channels,
166 int64_t kernel_size,
167 assets::TensorStorageType weight_storage_type) {
168 modules::Conv1dWeights weights;
169 weights.weight = store.load_tensor(source, prefix + ".weight", weight_storage_type, {out_channels, in_channels, kernel_size});
170 weights.bias = store.load_tensor(source, prefix + ".bias", assets::TensorStorageType::F32, {out_channels});
171 return weights;
172}
173
174modules::DepthwiseConv1dWeights load_depthwise_conv1d(
175 core::BackendWeightStore & store,

Callers 2

load_encoderFunction · 0.70
load_decoderFunction · 0.70

Calls 1

load_tensorMethod · 0.80

Tested by

no test coverage detected