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

Method can_load

src/models/vevo2/loader.cpp:65–76  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

63 {"top_p", "value", "AR nucleus sampling threshold."},
64 {"repetition_penalty", "value", "AR repetition penalty."},
65 {"max_tokens", "n", "Maximum generated AR content/style tokens."},
66 {"num_inference_steps", "n", "Flow-matching denoising steps."},
67 {"seed", "n", "Request seed; omitted means random."},
68 {"fm_noise_file", "path", "Optional flow-matching noise tensor file."},
69 };
70 out.session_options = {
71 {"vevo2.weight_type", "native|f32|f16|bf16|q8_0", "Default matmul weight storage type."},
72 {"vevo2.conv_weight_type", "native|f32|f16", "Default convolution weight storage type."},
73 {"vevo2.ar_weight_type", "native|f32|f16|bf16|q8_0", "AR matmul weight storage type."},
74 {"vevo2.ar_weight_context_mb", "n", "AR weight context size."},
75 {"vevo2.ar_prefill_graph_context_mb", "n", "AR prefill graph context size."},
76 {"vevo2.ar_decode_graph_context_mb", "n", "AR decode graph context size."},
77 {"vevo2.whisper_weight_type", "native|f32|f16|bf16|q8_0", "Whisper matmul weight storage type."},
78 {"vevo2.whisper_conv_weight_type", "native|f32|f16", "Whisper convolution weight storage type."},
79 {"vevo2.whisper_weight_context_mb", "n", "Whisper weight context size."},

Callers 1

find_loaderMethod · 0.45

Calls 3

is_existing_directoryFunction · 0.85
has_vevo2_assetsFunction · 0.85
is_existing_fileFunction · 0.85

Tested by

no test coverage detected