Skip to content

Support espdl #1519

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
sun-xiangyu wants to merge 4 commits into
base: main
Choose a base branch
from
Open

Support espdl #1519

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
83bbace
feat: support espdl format
sun-xiangyu Dec 3, 2025
929e945
Add script to generate espdl schema and metadata files
sun-xiangyu Dec 5, 2025
2f71d59
Add espdl test cases
sun-xiangyu Dec 5, 2025
efb6cf0
fix: fix ci errors
sun-xiangyu Dec 8, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions publish/electron-builder.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
{ "ext": "cmf", "name": "CNTK Model" },
{ "ext": "dlc", "name": "DLC Model" },
{ "ext": "dnn", "name": "CNTK Model" },
{ "ext": "espdl", "name": "ESPDL Model" },
{ "ext": "gguf", "name": "GGUF Model" },
{ "ext": "h5", "name": "Keras Model" },
{ "ext": "hd5", "name": "Keras Model" },
Expand Down
282 changes: 282 additions & 0 deletions source/espdl-metadata.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,282 @@
[
{
"name": "Add",
"module": "espdl",
"version": 1,
"description": "Element-wise addition of two tensors.",
"inputs": [
{ "name": "A", "description": "First input tensor" },
{ "name": "B", "description": "Second input tensor" }
],
"outputs": [
{ "name": "C", "description": "Output tensor" }
],
"category": "Tensor"
},
{
"name": "AveragePool",
"module": "espdl",
"version": 1,
"description": "Average pooling operation.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Pool"
},
{
"name": "BatchNormalization",
"module": "espdl",
"version": 1,
"description": "Batch normalization.",
"inputs": [
{ "name": "X", "description": "Input data tensor" },
{ "name": "scale", "description": "Scale tensor" },
{ "name": "B", "description": "Bias tensor" },
{ "name": "mean", "description": "Mean tensor" },
{ "name": "var", "description": "Variance tensor" }
],
"outputs": [
{ "name": "Y", "description": "Output data tensor" },
{ "name": "mean", "description": "Updated mean tensor (optional)" },
{ "name": "var", "description": "Updated variance tensor (optional)" },
{ "name": "saved_mean", "description": "Saved mean tensor (optional)" },
{ "name": "saved_var", "description": "Saved variance tensor (optional)" }
],
"category": "Normalization"
},
{
"name": "Clip",
"module": "espdl",
"version": 1,
"category": "Activation",
"description": "Clip operator limits values to a specified range.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Clipped output tensor" }
]
},
{
"name": "Concat",
"module": "espdl",
"version": 1,
"description": "Concatenates tensors along a given axis.",
"inputs": [
{ "name": "inputs", "list": true, "description": "Input tensors to concatenate" }
],
"outputs": [
{ "name": "output", "description": "Concatenated output tensor" }
],
"category": "Tensor"
},
{
"name": "Conv",
"module": "espdl",
"version": 1,
"description": "Convolution operator. Applies a convolution filter to the input.",
"inputs": [
{ "name": "input", "description": "Input feature map" },
{ "name": "weight", "description": "Convolution kernel weights" },
{ "name": "bias", "option": "optional", "description": "Bias values (optional)" }
],
"outputs": [
{ "name": "output", "description": "Output feature map" }
],
"category": "Layer"
},
{
"name": "Gemm",
"module": "espdl",
"version": 1,
"description": "General matrix multiplication: alpha * A * B + beta * C",
"inputs": [
{ "name": "A", "description": "Input tensor A" },
{ "name": "B", "description": "Input tensor B" },
{ "name": "C", "option": "optional", "description": "Input tensor C (optional)" }
],
"outputs": [
{ "name": "Y", "description": "Output tensor" }
],
"category": "Layer"
},
{
"name": "LeakyRelu",
"module": "espdl",
"version": 1,
"category": "Activation",
"description": "Leaky Rectified Linear Unit activation function.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
]
},
{
"name": "MaxPool",
"module": "espdl",
"version": 1,
"description": "Max pooling operation.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Pool"
},
{
"name": "Mul",
"module": "espdl",
"version": 1,
"description": "Element-wise multiplication of two tensors.",
"inputs": [
{ "name": "A", "description": "First input tensor" },
{ "name": "B", "description": "Second input tensor" }
],
"outputs": [
{ "name": "C", "description": "Output tensor" }
],
"category": "Tensor"
},
{
"name": "Pad",
"module": "espdl",
"version": 1,
"category": "Tensor",
"description": "Pad operator adds padding to tensor dimensions.",
"inputs": [
{ "name": "input", "description": "Input tensor" },
{ "name": "pads", "description": "Padding values" }
],
"outputs": [
{ "name": "output", "description": "Padded output tensor" }
]
},
{
"name": "Relu",
"module": "espdl",
"version": 1,
"description": "Rectified Linear Unit activation function.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Activation"
},
{
"name": "RequantizeLinear",
"module": "espdl",
"version": 1,
"category": "Quantization",
"description": "Requantize linear quantization operator.",
"inputs": [
{ "name": "input", "description": "Input tensor to requantize" },
{ "name": "scale", "description": "Scale for requantization" },
{ "name": "zero_point", "description": "Zero point for requantization" }
],
"outputs": [
{ "name": "output", "description": "Requantized output tensor" }
]
},
{
"name": "Reshape",
"module": "espdl",
"version": 1,
"description": "Reshapes a tensor to a new shape.",
"inputs": [
{ "name": "data", "description": "Input tensor" },
{ "name": "shape", "description": "New shape" }
],
"outputs": [
{ "name": "reshaped", "description": "Reshaped output tensor" }
],
"category": "Shape"
},
{
"name": "Resize",
"module": "espdl",
"version": 1,
"category": "Data",
"description": "Resize operator for spatial dimensions.",
"inputs": [
{ "name": "input", "description": "Input tensor" },
{ "name": "scales", "description": "Scale factors for each dimension" }
],
"outputs": [
{ "name": "output", "description": "Resized output tensor" }
]
},
{
"name": "Sigmoid",
"module": "espdl",
"version": 1,
"description": "Sigmoid activation function.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Activation"
},
{
"name": "Softmax",
"module": "espdl",
"version": 1,
"description": "Softmax activation function.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Activation"
},
{
"name": "Split",
"module": "espdl",
"version": 1,
"description": "Splits a tensor into multiple tensors along a given axis.",
"inputs": [
{ "name": "input", "description": "Input tensor to split" },
{ "name": "split", "description": "Optional list of split sizes or number of splits" }
],
"outputs": [
{ "name": "outputs", "list": true, "description": "Output tensors" }
],
"category": "Tensor"
},
{
"name": "Tanh",
"module": "espdl",
"version": 1,
"description": "Hyperbolic tangent activation function.",
"inputs": [
{ "name": "input", "description": "Input tensor" }
],
"outputs": [
{ "name": "output", "description": "Output tensor" }
],
"category": "Activation"
},
{
"name": "Transpose",
"module": "espdl",
"version": 1,
"description": "Transposes the dimensions of a tensor.",
"inputs": [
{ "name": "data", "description": "Input tensor" }
],
"outputs": [
{ "name": "transposed", "description": "Transposed output tensor" }
],
"category": "Tensor"
}
]
Loading
Loading