Skip to content

LLM inference in C/C++

C++ ggml

ci: fix Models Backend webgpu by supporting GGML_OP_DUP (#30216)

Pascale94acad853c0
diff --git a/ggml/src/ggml-webgpu/ggml-webgpu.cpp b/ggml/src/ggml-webgpu/ggml-webgpu.cpp
index 11a4fc46fe..b14a1fe2b0 100644
--- a/ggml/src/ggml-webgpu/ggml-webgpu.cpp
+++ b/ggml/src/ggml-webgpu/ggml-webgpu.cpp
@@ -3356,6 +3356,7 @@ static std::optional<webgpu_encoded_op> ggml_webgpu_encode(webgpu_context ctx,
case GGML_OP_TRANSPOSE:
case GGML_OP_RESHAPE:
return std::nullopt;
+ case GGML_OP_DUP:
case GGML_OP_CPY:
case GGML_OP_CONT:
return ggml_webgpu_cpy(ctx, src0, node);
@@ -4419,6 +4420,7 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const
supports_op = (src0->type == GGML_TYPE_F32 || src0->type == GGML_TYPE_F16 || src0->type == GGML_TYPE_I32 ||
src0->type == GGML_TYPE_I16);
break;
+ case GGML_OP_DUP:
case GGML_OP_CPY:
case GGML_OP_CONT:
supports_op = (op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_I32) &&