diff --git a/ggml/src/ggml-cpu/ggml-cpu.cpp b/ggml/src/ggml-cpu/ggml-cpu.cpp index c390957afa8e3..0b6419f83f0c2 100644 --- a/ggml/src/ggml-cpu/ggml-cpu.cpp +++ b/ggml/src/ggml-cpu/ggml-cpu.cpp @@ -394,8 +394,11 @@ static bool ggml_backend_cpu_device_supports_op(ggml_backend_dev_t dev, const st switch (op->op) { case GGML_OP_CPY: return + op->type != GGML_TYPE_IQ3_XXS && + op->type != GGML_TYPE_IQ3_S && op->type != GGML_TYPE_IQ2_XXS && op->type != GGML_TYPE_IQ2_XS && + op->type != GGML_TYPE_IQ2_S && op->type != GGML_TYPE_IQ1_S && op->type != GGML_TYPE_IQ1_M; // missing type_traits.from_float case GGML_OP_MUL_MAT: diff --git a/tests/test-backend-ops.cpp b/tests/test-backend-ops.cpp index b9454ba596c21..ccdd3fb57a504 100644 --- a/tests/test-backend-ops.cpp +++ b/tests/test-backend-ops.cpp @@ -3549,8 +3549,8 @@ static std::vector> make_test_cases_eval() { for (ggml_type type_src : {GGML_TYPE_F16, GGML_TYPE_F32}) { for (ggml_type type_dst : all_types) { - test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 4, 4, 4})); - test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 2, 3, 4}, {0, 2, 1, 3})); // cpy by rows + test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 4, 4, 4})); + test_cases.emplace_back(new test_cpy(type_src, type_dst, {256, 2, 3, 4}, {0, 2, 1, 3})); // cpy by rows } } for (ggml_type type_src : {GGML_TYPE_F16, GGML_TYPE_F32}) {