aotrih's picture
initial commit
e8f351b
[
{
"metadataOutputVersion" : "3.0",
"storagePrecision" : "Float16",
"outputSchema" : [
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 1 × 51866)",
"shortDescription" : "",
"shape" : "[1, 1, 51866]",
"name" : "logits",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 1)",
"shortDescription" : "",
"shape" : "[1, 5120, 1, 1]",
"name" : "key_cache_updates",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 5120 × 1 × 1)",
"shortDescription" : "",
"shape" : "[1, 5120, 1, 1]",
"name" : "value_cache_updates",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 1536)",
"shortDescription" : "",
"shape" : "[1, 1536]",
"name" : "alignment_heads_weights",
"type" : "MultiArray"
}
],
"modelParameters" : [
],
"specificationVersion" : 9,
"mlProgramOperationTypeHistogram" : {
"Ios18.expandDims" : 8,
"Ios18.softmax" : 8,
"Ios18.mul" : 16,
"Ios18.matmul" : 16,
"Ios18.batchNorm" : 13,
"Ios16.reduceMean" : 1,
"Split" : 2,
"Ios18.readState" : 5,
"Ios18.gather" : 2,
"Ios18.add" : 29,
"Ios18.layerNorm" : 13,
"Ios18.reshape" : 32,
"Ios18.linear" : 1,
"Ios18.conv" : 32,
"Ios18.gelu" : 4,
"Ios18.concat" : 3,
"Ios18.cast" : 1,
"Ios18.transpose" : 1,
"Ios18.sliceByIndex" : 20,
"Ios18.squeeze" : 1
},
"computePrecision" : "Mixed (Float16, Int32, UInt16)",
"isUpdatable" : "0",
"stateSchema" : [
{
"dataType" : "Float16",
"isOptional" : "0",
"formattedType" : "State (Float16 1 × 1536)",
"shortDescription" : "",
"shape" : "[1, 1536]",
"name" : "encoder_attn_key_padding_mask",
"type" : "State"
},
{
"dataType" : "Float16",
"isOptional" : "0",
"formattedType" : "State (Float16 4 × 1280 × 1 × 1536)",
"shortDescription" : "",
"shape" : "[4, 1280, 1, 1536]",
"name" : "encoder_attn_key_cache",
"type" : "State"
},
{
"dataType" : "Float16",
"isOptional" : "0",
"formattedType" : "State (Float16 4 × 1280 × 1 × 1536)",
"shortDescription" : "",
"shape" : "[4, 1280, 1, 1536]",
"name" : "encoder_attn_value_cache",
"type" : "State"
},
{
"dataType" : "Float16",
"isOptional" : "0",
"formattedType" : "State (Float16 4 × 1280 × 1 × 448)",
"shortDescription" : "",
"shape" : "[4, 1280, 1, 448]",
"name" : "self_attn_key_cache",
"type" : "State"
},
{
"dataType" : "Float16",
"isOptional" : "0",
"formattedType" : "State (Float16 4 × 1280 × 1 × 448)",
"shortDescription" : "",
"shape" : "[4, 1280, 1, 448]",
"name" : "self_attn_value_cache",
"type" : "State"
}
],
"availability" : {
"macOS" : "15.0",
"tvOS" : "18.0",
"visionOS" : "2.0",
"watchOS" : "11.0",
"iOS" : "18.0",
"macCatalyst" : "18.0"
},
"modelType" : {
"name" : "MLModelType_mlProgram"
},
"userDefinedMetadata" : {
"com.github.apple.coremltools.source_dialect" : "TorchScript",
"com.github.apple.coremltools.source" : "torch==2.5.1",
"com.github.apple.coremltools.version" : "8.0"
},
"inputSchema" : [
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Int32",
"formattedType" : "MultiArray (Int32 1)",
"shortDescription" : "",
"shape" : "[1]",
"name" : "input_ids",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Int32",
"formattedType" : "MultiArray (Int32 1)",
"shortDescription" : "",
"shape" : "[1]",
"name" : "cache_length",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 448)",
"shortDescription" : "",
"shape" : "[1, 448]",
"name" : "kv_cache_update_mask",
"type" : "MultiArray"
},
{
"hasShapeFlexibility" : "0",
"isOptional" : "0",
"dataType" : "Float16",
"formattedType" : "MultiArray (Float16 1 × 448)",
"shortDescription" : "",
"shape" : "[1, 448]",
"name" : "decoder_key_padding_mask",
"type" : "MultiArray"
}
],
"generatedClassName" : "TextDecoderStateful",
"method" : "predict"
}
]