diff --git a/apps/hexagon_launcher/README.md b/apps/hexagon_launcher/README.md index 210759a80c7c..cc433f245759 100644 --- a/apps/hexagon_launcher/README.md +++ b/apps/hexagon_launcher/README.md @@ -118,7 +118,7 @@ mod, params = relay.frontend.from_tflite( tflite_model, shape_dict=shape_dict, dtype_dict=dtype_dict ) -target = tvm.target.hexagon('v68', link_params=True) +target = tvm.target.hexagon('v68') with tvm.transform.PassContext(opt_level=3): lib = relay.build(mod, tvm.target.Target(target, host=target), params=params, mod_name="default") @@ -172,7 +172,7 @@ A sample output JSON from running the Inception V3 model may look like When using AoT, the `target` needs to be `llvm`: ``` -aot_target = "llvm -keys=hexagon -link-params=0 -mattr=+hvxv69,+hvx-length128b,+hvx-qfloat,-hvx-ieee-fp -mcpu=hexagonv69 -mtriple=hexagon" +aot_target = "llvm -keys=hexagon -mattr=+hvxv69,+hvx-length128b,+hvx-qfloat,-hvx-ieee-fp -mcpu=hexagonv69 -mtriple=hexagon" aot_host_target = aot_target ``` diff --git a/apps/howto_deploy/prepare_test_libs.py b/apps/howto_deploy/prepare_test_libs.py index a6c7688d2084..8e9f8b5f7335 100644 --- a/apps/howto_deploy/prepare_test_libs.py +++ b/apps/howto_deploy/prepare_test_libs.py @@ -33,7 +33,7 @@ def prepare_test_libs(base_path): fadd_dylib.export_library(dylib_path) # Compile library in system library mode - fadd_syslib = tvm.build(s, [A, B], "llvm --system-lib", name="addonesys") + fadd_syslib = tvm.build(s, [A, B], "llvm", name="addonesys") syslib_path = os.path.join(base_path, "test_addone_sys.o") fadd_syslib.save(syslib_path) diff --git a/apps/sgx/src/build_model.py b/apps/sgx/src/build_model.py index 1fc297d8a094..ea3b4ed992ad 100755 --- a/apps/sgx/src/build_model.py +++ b/apps/sgx/src/build_model.py @@ -39,7 +39,12 @@ def main(): ) with tvm.transform.PassContext(opt_level=3): - graph, lib, params = relay.build(net, "llvm --system-lib", params=params) + graph, lib, params = relay.build( + net, + "llvm", + params=params, + runtime=tvm.relay.backend.Runtime("cpp", {"system-lib": True}), + ) build_dir = osp.abspath(sys.argv[1]) if not osp.isdir(build_dir): diff --git a/apps/wasm-standalone/wasm-graph/tools/build_graph_lib.py b/apps/wasm-standalone/wasm-graph/tools/build_graph_lib.py index 9b262c398e00..c2f9089710a3 100755 --- a/apps/wasm-standalone/wasm-graph/tools/build_graph_lib.py +++ b/apps/wasm-standalone/wasm-graph/tools/build_graph_lib.py @@ -72,10 +72,15 @@ def build_graph_lib(opt_level): shape_dict = {input_name: img_data.shape} mod, params = relay.frontend.from_onnx(onnx_model, shape_dict) - target = "llvm -mtriple=wasm32-unknown-unknown -mattr=+simd128 --system-lib" + target = "llvm -mtriple=wasm32-unknown-unknown -mattr=+simd128" with tvm.transform.PassContext(opt_level=opt_level): - factory = relay.build(mod, target=target, params=params) + factory = relay.build( + mod, + target=target, + params=params, + runtime=tvm.relay.backend.Runtime("cpp", {"system-lib": True}), + ) # Save the model artifacts to obj_file obj_file = os.path.join(out_dir, "graph.o") diff --git a/gallery/how_to/tune_with_autoscheduler/ci_logs/matmul.json b/gallery/how_to/tune_with_autoscheduler/ci_logs/matmul.json index 2e3a98404dda..b0d33a911a63 100644 --- a/gallery/how_to/tune_with_autoscheduler/ci_logs/matmul.json +++ b/gallery/how_to/tune_with_autoscheduler/ci_logs/matmul.json @@ -1,2 +1,2 @@ # Keep a valid schedule for demonstraction. This is used to prevent flasky errors in CI. -{"i": [["[\"matmul_add\", 1024, 1024, 1024, \"float32\"]", "llvm -keys=cpu -link-params=0", [18, 64, 64, 0, 0, 0, 0, 0]], [[], [["SP", 2, 0, 1024, [2, 1, 4], 1], ["SP", 2, 4, 1024, [1, 1, 8], 1], ["SP", 2, 8, 1024, [4], 1], ["RE", 2, [0, 4, 1, 5, 8, 2, 6, 9, 3, 7]], ["FSP", 4, 0, 0, 2], ["FSP", 4, 3, 1, 2], ["RE", 4, [0, 3, 1, 4, 2, 5]], ["CA", 2, 4, 3], ["FU", 4, [0, 1]], ["AN", 4, 0, 3], ["PR", 2, 0, "auto_unroll_max_step$8"], ["AN", 2, 9, 2], ["AN", 4, 4, 2]]]], "r": [[0.0044742], 0, 0.335558, 1607112214], "v": "v0.3"} +{"i": [["[\"matmul_add\", 1024, 1024, 1024, \"float32\"]", "llvm -keys=cpu", [18, 64, 64, 0, 0, 0, 0, 0]], [[], [["SP", 2, 0, 1024, [2, 1, 4], 1], ["SP", 2, 4, 1024, [1, 1, 8], 1], ["SP", 2, 8, 1024, [4], 1], ["RE", 2, [0, 4, 1, 5, 8, 2, 6, 9, 3, 7]], ["FSP", 4, 0, 0, 2], ["FSP", 4, 3, 1, 2], ["RE", 4, [0, 3, 1, 4, 2, 5]], ["CA", 2, 4, 3], ["FU", 4, [0, 1]], ["AN", 4, 0, 3], ["PR", 2, 0, "auto_unroll_max_step$8"], ["AN", 2, 9, 2], ["AN", 4, 4, 2]]]], "r": [[0.0044742], 0, 0.335558, 1607112214], "v": "v0.3"} diff --git a/gallery/how_to/tune_with_autoscheduler/ci_logs/resnet-50-NHWC-B1-llvm.json b/gallery/how_to/tune_with_autoscheduler/ci_logs/resnet-50-NHWC-B1-llvm.json index 3dd4541fd33a..4fb148c887bd 100644 --- a/gallery/how_to/tune_with_autoscheduler/ci_logs/resnet-50-NHWC-B1-llvm.json +++ b/gallery/how_to/tune_with_autoscheduler/ci_logs/resnet-50-NHWC-B1-llvm.json @@ -1,28 +1,28 @@ # Provide valid schedules for resnet-50 for CPU. # This is used to run the tutorial on the documentation web server. -{"i": [["[\"d7b65649a4dd54becea0a52aabbc5af5\", 1, 1000, 1, 1000]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 1, 1000, [50], 1], ["RF", 3, 2, 1], ["RE", 3, [0, 2, 1]], ["SP", 1, 1, 1000, [20], 1], ["RF", 1, 2, 1], ["RE", 1, [0, 2, 1]], ["CR", 6], ["CA", 5, 6, 1], ["CR", 4], ["CA", 2, 3, 1], ["AN", 1, 0, 3], ["FU", 3, [0, 1]], ["AN", 3, 0, 3], ["AN", 4, 0, 3], ["FU", 6, [0, 1]], ["AN", 6, 0, 3], ["PR", 1, 0, "auto_unroll_max_step$16"], ["PR", 2, 0, "auto_unroll_max_step$16"], ["PR", 4, 0, "auto_unroll_max_step$16"], ["PR", 5, 0, "auto_unroll_max_step$64"]]]], "r": [[8.75e-06, 1.0781e-05, 9.875e-06, 9.836e-06, 1.0357e-05, 1.0238e-05, 1.0341e-05, 9.75e-06, 9.561e-06, 1.0122e-05], 0, 0.17921, 1606960872], "v": "v0.5"} -{"i": [["[\"69115f188984ae34ede37c3b8ca40b43\", 1, 7, 7, 2048, 1, 1, 1, 2048]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CA", 1, 2, 3], ["FU", 2, [0, 1, 2, 3]], ["AN", 2, 0, 3], ["PR", 1, 0, "auto_unroll_max_step$16"]]]], "r": [[6.28e-06, 8.176e-06, 8.048e-06, 7.942e-06, 7.977e-06, 8.002e-06, 8.093e-06, 7.924e-06, 7.943e-06, 7.924e-06], 0, 0.130759, 1606960900], "v": "v0.5"} -{"i": [["[\"875556d12d0be2269206a7775d5296a6\", 1, 7, 7, 512, 1, 1, 512, 2048, 1, 7, 7, 2048, 1, 1, 1, 2048, 1, 1, 1, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 9], ["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 7, 1], 1], ["SP", 3, 8, 7, [1, 1, 1], 1], ["SP", 3, 12, 2048, [8, 2, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 10, 0, 3, 2], ["FSP", 10, 3, 4, 2], ["FSP", 10, 6, 5, 2], ["FSP", 10, 9, 6, 2], ["RE", 10, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 10, 7], ["CI", 1], ["FU", 10, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 10, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 10, 4, 2]]]], "r": [[0.000175984, 0.000171372, 0.00018538, 0.000178085, 0.00017879, 0.000179878, 0.000179221, 0.000178598, 0.000176714, 0.000168318], 0, 0.277929, 1606960917], "v": "v0.5"} -{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 14, 14, 1024, 1, 1, 1024, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 1, 7], 1], ["SP", 3, 8, 7, [1, 1, 1], 1], ["SP", 3, 12, 512, [32, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["FU", 3, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 3, 0, 3], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 14, 2]]]], "r": [[0.00012651, 0.00012801, 0.000128605, 0.00013267, 0.00012596, 0.000126418, 0.000121995, 0.000127242, 0.000128152, 0.000129989], 0, 0.310011, 1606960986], "v": "v0.5"} -{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 14, 14, 256, 1, 1, 256, 1024, 1, 14, 14, 1024, 1, 1, 1, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 1, 2], 1], ["SP", 3, 8, 14, [2, 1, 1], 1], ["SP", 3, 12, 1024, [1, 1, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["FU", 3, [0, 1, 2, 3, 4, 5, 6]], ["AN", 3, 0, 3], ["FU", 8, [0, 1, 2, 3]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"]]]], "r": [[0.000183629, 0.000188334, 0.000195553, 0.000187308, 0.000196409, 0.000190496, 0.000190344, 0.000188567, 0.000186319, 0.000187136], 0, 0.384722, 1606961002], "v": "v0.5"} -{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 28, 28, 512, 1, 1, 512, 256, 1, 1, 1, 256, 1, 14, 14, 256]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [2, 1, 7], 1], ["SP", 3, 8, 14, [2, 1, 1], 1], ["SP", 3, 12, 256, [16, 4, 4], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [64], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["CR", 6], ["FU", 3, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 3, 0, 3], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 14, 2]]]], "r": [[0.000118033, 0.000116806, 0.000134047, 0.000116701, 0.000116219, 0.000116834, 0.000117132, 0.000117029, 0.000116393, 0.000116778], 0, 0.31025, 1606961069], "v": "v0.5"} -{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 28, 28, 128, 1, 1, 128, 512, 1, 28, 28, 512, 1, 1, 1, 512, 1, 28, 28, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [1, 1, 1], 1], ["SP", 3, 8, 28, [7, 1, 1], 1], ["SP", 3, 12, 512, [1, 2, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 128, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 8, 0, 2, 2], ["FSP", 8, 3, 3, 2], ["FSP", 8, 6, 4, 2], ["FSP", 8, 9, 5, 2], ["RE", 8, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 8, 7], ["CI", 1], ["FU", 8, [0, 1, 2, 3]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$0"], ["AN", 3, 21, 2]]]], "r": [[0.00019554, 0.000203491, 0.000199599, 0.000194289, 0.000197556, 0.000199504, 0.000198527, 0.000200656, 0.000200037, 0.000201954], 0, 0.240599, 1606961080], "v": "v0.5"} -{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 56, 56, 256, 1, 1, 256, 128, 1, 1, 1, 128, 1, 28, 28, 128]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [14, 1, 2], 1], ["SP", 3, 8, 28, [2, 1, 1], 1], ["SP", 3, 12, 128, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [16], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 1], ["FSP", 6, 2, 2, 1], ["FSP", 6, 4, 3, 1], ["FSP", 6, 6, 4, 1], ["RE", 6, [0, 2, 4, 6, 1, 3, 5, 7]], ["CA", 3, 6, 3], ["CI", 1], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000128461, 0.000158344, 0.000154659, 0.000148478, 0.000162668, 0.000155789, 0.000149412, 0.000141607, 0.000148815, 0.000165989], 0, 0.299928, 1606961156], "v": "v0.5"} -{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 56, 56, 64, 1, 1, 64, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [14, 2, 2], 1], ["SP", 3, 8, 56, [2, 1, 2], 1], ["SP", 3, 12, 64, [1, 16, 4], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 1], ["FSP", 6, 2, 2, 1], ["FSP", 6, 4, 3, 1], ["FSP", 6, 6, 4, 1], ["RE", 6, [0, 2, 4, 6, 1, 3, 5, 7]], ["CA", 3, 6, 3], ["CI", 1], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2]]]], "r": [[7.8291e-05, 7.4365e-05, 6.7147e-05, 6.7413e-05, 8.1894e-05, 7.1771e-05, 7.2916e-05, 6.6615e-05, 7.3038e-05, 7.4967e-05], 0, 1.09095, 1606961258], "v": "v0.5"} -{"i": [["[\"a5612fdeb9db4d579a75ec225ea4c06a\", 1, 112, 112, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 4], ["CA", 2, 5, 3], ["CR", 1], ["FU", 1, [0, 1, 2]], ["AN", 1, 0, 3], ["FU", 5, [0, 1, 2]], ["AN", 5, 0, 3], ["PR", 2, 0, "auto_unroll_max_step$64"]]]], "r": [[2.9217e-05, 3.1065e-05, 3.188e-05, 3.0897e-05, 3.1295e-05, 3.1307e-05, 3.19e-05, 3.1038e-05, 3.1919e-05, 3.2077e-05], 0, 0.217184, 1606961266], "v": "v0.5"} -{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 14, 14, 1024, 1, 1, 1024, 256, 1, 1, 1, 256, 1, 14, 14, 256]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [1, 1, 2], 1], ["SP", 3, 8, 14, [7, 2, 1], 1], ["SP", 3, 12, 256, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2]]]], "r": [[0.000153068, 0.000161094, 0.000164674, 0.000160245, 0.000159626, 0.000146788, 0.000140718, 0.000159237, 0.000162109, 0.000139686], 0, 0.273946, 1606961647], "v": "v0.5"} -{"i": [["[\"12b88bedece6984af589a28b43e0f3c4\", 1, 224, 224, 3, 7, 7, 3, 64, 1, 1, 1, 64, 1, 112, 112, 64]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 112, [1, 1, 4], 1], ["SP", 3, 8, 112, [4, 2, 1], 1], ["SP", 3, 12, 64, [1, 1, 16], 1], ["SP", 3, 16, 7, [7], 1], ["SP", 3, 18, 7, [7], 1], ["SP", 3, 20, 3, [3], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CA", 1, 6, 3], ["FU", 6, [0, 1, 2]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 1, 3, 2], ["AN", 3, 21, 2], ["AN", 6, 9, 2]]]], "r": [[0.000247808, 0.000233393, 0.000251767, 0.000252226, 0.000254169, 0.000254176, 0.00025333, 0.00025511, 0.000253678, 0.000251738], 0, 0.315503, 1606961659], "v": "v0.5"} -{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 7, 7, 512, 1, 1, 512, 2048, 1, 7, 7, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 1, 1], 1], ["SP", 3, 8, 7, [1, 1, 7], 1], ["SP", 3, 12, 2048, [256, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000169437, 0.000169021, 0.00016965, 0.00017079, 0.000170862, 0.0001692, 0.000164768, 0.000175541, 0.000171528, 0.000169094], 0, 0.25194, 1606961681], "v": "v0.5"} -{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 56, 56, 256, 1, 1, 256, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [4, 1, 1], 1], ["SP", 3, 8, 56, [7, 4, 1], 1], ["SP", 3, 12, 64, [4, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 6, 5, 2]]]], "r": [[0.00015141, 0.000158121, 0.000132758, 0.00015109, 0.000148266, 0.000152599, 0.000150809, 0.000151947, 0.000150702, 0.000156091], 0, 0.221869, 1606961698], "v": "v0.5"} -{"i": [["[\"2350d19dc42a0665244368384c66b3a5\", 1, 56, 56, 64, 3, 3, 64, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [1, 1, 1], 1], ["SP", 3, 8, 56, [7, 1, 4], 1], ["SP", 3, 12, 64, [1, 1, 16], 1], ["SP", 3, 16, 3, [1], 1], ["SP", 3, 18, 3, [3], 1], ["SP", 3, 20, 64, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["FU", 1, [0, 1, 2]], ["AN", 1, 0, 3], ["FU", 6, [0, 1, 2, 3, 4, 5, 6]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 6, 5, 2]]]], "r": [[0.000221341, 0.000225005, 0.000209954, 0.000209741, 0.000228281, 0.000208451, 0.000223046, 0.000222672, 0.000228098, 0.000220093], 0, 0.231218, 1606961709], "v": "v0.5"} -{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 7, 7, 2048, 1, 1, 2048, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [7, 1, 1], 1], ["SP", 3, 8, 7, [1, 7, 1], 1], ["SP", 3, 12, 512, [2, 2, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 2048, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$0"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000165941, 0.000152645, 0.000165687, 0.000166639, 0.000166094, 0.00016649, 0.000164394, 0.000169288, 0.000169497, 0.000168535], 0, 0.245559, 1606961724], "v": "v0.5"} -{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256, 1, 56, 56, 256]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [4, 2, 1], 1], ["SP", 3, 8, 56, [1, 1, 1], 1], ["SP", 3, 12, 256, [2, 4, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 5, 6, 2]]]], "r": [[0.000161206, 0.000161372, 0.000158862, 0.000159596, 0.00014964, 0.000162042, 0.000159626, 0.000158166, 0.000161209, 0.000159408], 0, 0.337652, 1606961748], "v": "v0.5"} -{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 28, 28, 512, 1, 1, 512, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 2, 1], 1], ["SP", 3, 8, 14, [14, 1, 1], 1], ["SP", 3, 12, 1024, [2, 2, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2]]]], "r": [[0.000238006, 0.000235502, 0.000239805, 0.000234637, 0.000235266, 0.000238355, 0.000240836, 0.000232856, 0.000231219, 0.000238776], 0, 0.219506, 1606961782], "v": "v0.5"} -{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 56, 56, 256, 1, 1, 256, 512, 1, 28, 28, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [2, 1, 2], 1], ["SP", 3, 8, 28, [1, 2, 1], 1], ["SP", 3, 12, 512, [16, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000213071, 0.000218117, 0.000216346, 0.000216237, 0.000214703, 0.00021605, 0.000210522, 0.000214234, 0.000218293, 0.00021484], 0, 0.291873, 1606961801], "v": "v0.5"} -{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 14, 14, 1024, 1, 1, 1024, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [7, 1, 1], 1], ["SP", 3, 8, 7, [1, 7, 1], 1], ["SP", 3, 12, 2048, [128, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000265306, 0.000259738, 0.000256412, 0.000284932, 0.000267557, 0.000266362, 0.00026533, 0.000263389, 0.000263022, 0.000263069], 0, 0.296232, 1606961838], "v": "v0.5"} -{"i": [["[\"2350d19dc42a0665244368384c66b3a5\", 1, 7, 7, 512, 3, 3, 512, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 7, 1], 1], ["SP", 3, 8, 7, [7, 1, 1], 1], ["SP", 3, 12, 512, [4, 1, 8], 1], ["SP", 3, 16, 3, [3], 1], ["SP", 3, 18, 3, [1], 1], ["SP", 3, 20, 512, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CR", 1], ["FU", 1, [0, 1, 2, 3]], ["AN", 1, 0, 3], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000269786, 0.0002657, 0.000261922, 0.000267462, 0.000270495, 0.000265371, 0.000273858, 0.000268022, 0.000266746, 0.000272337], 0, 0.331923, 1606961848], "v": "v0.5"} -{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 14, 14, 256, 1, 1, 256, 1024, 1, 14, 14, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 1, 2], 1], ["SP", 3, 8, 14, [7, 2, 1], 1], ["SP", 3, 12, 1024, [16, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000159777, 0.00015711, 0.000163052, 0.000152569, 0.00015342, 0.000154918, 0.000153887, 0.000154133, 0.000154319, 0.000150102], 0, 0.195628, 1606961878], "v": "v0.5"} -{"i": [["[\"7006235cfc29b73be524cf390ed5a977\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [56, 1, 1], 1], ["SP", 3, 8, 56, [14, 1, 2], 1], ["SP", 3, 12, 256, [1, 2, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [64], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000159044, 0.000157356, 0.000158889, 0.000160304, 0.000158648, 0.000159749, 0.000143679, 0.000156393, 0.000164916, 0.000155957], 0, 0.240777, 1606961918], "v": "v0.5"} -{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 28, 28, 128, 1, 1, 128, 512, 1, 28, 28, 512, 1, 28, 28, 512]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [1, 1, 2], 1], ["SP", 3, 8, 28, [1, 1, 1], 1], ["SP", 3, 12, 512, [4, 1, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 128, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000168259, 0.000157338, 0.0001551, 0.000156552, 0.000160492, 0.000164505, 0.000144937, 0.000138397, 0.000153011, 0.000153186], 0, 0.231498, 1606961965], "v": "v0.5"} -{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256, 1, 1, 1, 256, 1, 56, 56, 256]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [7, 1, 4], 1], ["SP", 3, 8, 56, [4, 2, 1], 1], ["SP", 3, 12, 256, [32, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 8, 0, 2, 2], ["FSP", 8, 3, 3, 2], ["FSP", 8, 6, 4, 2], ["FSP", 8, 9, 5, 2], ["RE", 8, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 8, 7], ["CI", 1], ["FU", 8, [0, 1, 2, 3, 4, 5, 6]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 8, 5, 2]]]], "r": [[0.000185957, 0.000180964, 0.000179419, 0.000168205, 0.000176155, 0.000178243, 0.000180175, 0.00017753, 0.000174475, 0.000158878], 0, 0.316404, 1606961979], "v": "v0.5"} -{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 28, 28, 512, 1, 1, 512, 128, 1, 1, 1, 128, 1, 28, 28, 128]", "llvm -keys=cpu -link-params=0 -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [7, 1, 4], 1], ["SP", 3, 8, 28, [14, 1, 1], 1], ["SP", 3, 12, 128, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000150378, 0.000154444, 0.000156051, 0.000130306, 0.000156154, 0.000131167, 0.000142357, 0.000152532, 0.000131899, 0.000157696], 0, 0.18509, 1606962002], "v": "v0.5"} +{"i": [["[\"d7b65649a4dd54becea0a52aabbc5af5\", 1, 1000, 1, 1000]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 1, 1000, [50], 1], ["RF", 3, 2, 1], ["RE", 3, [0, 2, 1]], ["SP", 1, 1, 1000, [20], 1], ["RF", 1, 2, 1], ["RE", 1, [0, 2, 1]], ["CR", 6], ["CA", 5, 6, 1], ["CR", 4], ["CA", 2, 3, 1], ["AN", 1, 0, 3], ["FU", 3, [0, 1]], ["AN", 3, 0, 3], ["AN", 4, 0, 3], ["FU", 6, [0, 1]], ["AN", 6, 0, 3], ["PR", 1, 0, "auto_unroll_max_step$16"], ["PR", 2, 0, "auto_unroll_max_step$16"], ["PR", 4, 0, "auto_unroll_max_step$16"], ["PR", 5, 0, "auto_unroll_max_step$64"]]]], "r": [[8.75e-06, 1.0781e-05, 9.875e-06, 9.836e-06, 1.0357e-05, 1.0238e-05, 1.0341e-05, 9.75e-06, 9.561e-06, 1.0122e-05], 0, 0.17921, 1606960872], "v": "v0.5"} +{"i": [["[\"69115f188984ae34ede37c3b8ca40b43\", 1, 7, 7, 2048, 1, 1, 1, 2048]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CA", 1, 2, 3], ["FU", 2, [0, 1, 2, 3]], ["AN", 2, 0, 3], ["PR", 1, 0, "auto_unroll_max_step$16"]]]], "r": [[6.28e-06, 8.176e-06, 8.048e-06, 7.942e-06, 7.977e-06, 8.002e-06, 8.093e-06, 7.924e-06, 7.943e-06, 7.924e-06], 0, 0.130759, 1606960900], "v": "v0.5"} +{"i": [["[\"875556d12d0be2269206a7775d5296a6\", 1, 7, 7, 512, 1, 1, 512, 2048, 1, 7, 7, 2048, 1, 1, 1, 2048, 1, 1, 1, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 9], ["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 7, 1], 1], ["SP", 3, 8, 7, [1, 1, 1], 1], ["SP", 3, 12, 2048, [8, 2, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 10, 0, 3, 2], ["FSP", 10, 3, 4, 2], ["FSP", 10, 6, 5, 2], ["FSP", 10, 9, 6, 2], ["RE", 10, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 10, 7], ["CI", 1], ["FU", 10, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 10, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 10, 4, 2]]]], "r": [[0.000175984, 0.000171372, 0.00018538, 0.000178085, 0.00017879, 0.000179878, 0.000179221, 0.000178598, 0.000176714, 0.000168318], 0, 0.277929, 1606960917], "v": "v0.5"} +{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 14, 14, 1024, 1, 1, 1024, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 1, 7], 1], ["SP", 3, 8, 7, [1, 1, 1], 1], ["SP", 3, 12, 512, [32, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["FU", 3, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 3, 0, 3], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 14, 2]]]], "r": [[0.00012651, 0.00012801, 0.000128605, 0.00013267, 0.00012596, 0.000126418, 0.000121995, 0.000127242, 0.000128152, 0.000129989], 0, 0.310011, 1606960986], "v": "v0.5"} +{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 14, 14, 256, 1, 1, 256, 1024, 1, 14, 14, 1024, 1, 1, 1, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 1, 2], 1], ["SP", 3, 8, 14, [2, 1, 1], 1], ["SP", 3, 12, 1024, [1, 1, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["FU", 3, [0, 1, 2, 3, 4, 5, 6]], ["AN", 3, 0, 3], ["FU", 8, [0, 1, 2, 3]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"]]]], "r": [[0.000183629, 0.000188334, 0.000195553, 0.000187308, 0.000196409, 0.000190496, 0.000190344, 0.000188567, 0.000186319, 0.000187136], 0, 0.384722, 1606961002], "v": "v0.5"} +{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 28, 28, 512, 1, 1, 512, 256, 1, 1, 1, 256, 1, 14, 14, 256]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [2, 1, 7], 1], ["SP", 3, 8, 14, [2, 1, 1], 1], ["SP", 3, 12, 256, [16, 4, 4], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [64], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["CI", 1], ["CR", 6], ["FU", 3, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 3, 0, 3], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 14, 2]]]], "r": [[0.000118033, 0.000116806, 0.000134047, 0.000116701, 0.000116219, 0.000116834, 0.000117132, 0.000117029, 0.000116393, 0.000116778], 0, 0.31025, 1606961069], "v": "v0.5"} +{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 28, 28, 128, 1, 1, 128, 512, 1, 28, 28, 512, 1, 1, 1, 512, 1, 28, 28, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [1, 1, 1], 1], ["SP", 3, 8, 28, [7, 1, 1], 1], ["SP", 3, 12, 512, [1, 2, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 128, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 8, 0, 2, 2], ["FSP", 8, 3, 3, 2], ["FSP", 8, 6, 4, 2], ["FSP", 8, 9, 5, 2], ["RE", 8, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 8, 7], ["CI", 1], ["FU", 8, [0, 1, 2, 3]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$0"], ["AN", 3, 21, 2]]]], "r": [[0.00019554, 0.000203491, 0.000199599, 0.000194289, 0.000197556, 0.000199504, 0.000198527, 0.000200656, 0.000200037, 0.000201954], 0, 0.240599, 1606961080], "v": "v0.5"} +{"i": [["[\"de7d1695278cf52778b038e6573d7626\", 1, 56, 56, 256, 1, 1, 256, 128, 1, 1, 1, 128, 1, 28, 28, 128]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [14, 1, 2], 1], ["SP", 3, 8, 28, [2, 1, 1], 1], ["SP", 3, 12, 128, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [16], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 1], ["FSP", 6, 2, 2, 1], ["FSP", 6, 4, 3, 1], ["FSP", 6, 6, 4, 1], ["RE", 6, [0, 2, 4, 6, 1, 3, 5, 7]], ["CA", 3, 6, 3], ["CI", 1], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000128461, 0.000158344, 0.000154659, 0.000148478, 0.000162668, 0.000155789, 0.000149412, 0.000141607, 0.000148815, 0.000165989], 0, 0.299928, 1606961156], "v": "v0.5"} +{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 56, 56, 64, 1, 1, 64, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [14, 2, 2], 1], ["SP", 3, 8, 56, [2, 1, 2], 1], ["SP", 3, 12, 64, [1, 16, 4], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 1], ["FSP", 6, 2, 2, 1], ["FSP", 6, 4, 3, 1], ["FSP", 6, 6, 4, 1], ["RE", 6, [0, 2, 4, 6, 1, 3, 5, 7]], ["CA", 3, 6, 3], ["CI", 1], ["FU", 6, [0, 1, 2, 3]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2]]]], "r": [[7.8291e-05, 7.4365e-05, 6.7147e-05, 6.7413e-05, 8.1894e-05, 7.1771e-05, 7.2916e-05, 6.6615e-05, 7.3038e-05, 7.4967e-05], 0, 1.09095, 1606961258], "v": "v0.5"} +{"i": [["[\"a5612fdeb9db4d579a75ec225ea4c06a\", 1, 112, 112, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 4], ["CA", 2, 5, 3], ["CR", 1], ["FU", 1, [0, 1, 2]], ["AN", 1, 0, 3], ["FU", 5, [0, 1, 2]], ["AN", 5, 0, 3], ["PR", 2, 0, "auto_unroll_max_step$64"]]]], "r": [[2.9217e-05, 3.1065e-05, 3.188e-05, 3.0897e-05, 3.1295e-05, 3.1307e-05, 3.19e-05, 3.1038e-05, 3.1919e-05, 3.2077e-05], 0, 0.217184, 1606961266], "v": "v0.5"} +{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 14, 14, 1024, 1, 1, 1024, 256, 1, 1, 1, 256, 1, 14, 14, 256]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [1, 1, 2], 1], ["SP", 3, 8, 14, [7, 2, 1], 1], ["SP", 3, 12, 256, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2]]]], "r": [[0.000153068, 0.000161094, 0.000164674, 0.000160245, 0.000159626, 0.000146788, 0.000140718, 0.000159237, 0.000162109, 0.000139686], 0, 0.273946, 1606961647], "v": "v0.5"} +{"i": [["[\"12b88bedece6984af589a28b43e0f3c4\", 1, 224, 224, 3, 7, 7, 3, 64, 1, 1, 1, 64, 1, 112, 112, 64]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 112, [1, 1, 4], 1], ["SP", 3, 8, 112, [4, 2, 1], 1], ["SP", 3, 12, 64, [1, 1, 16], 1], ["SP", 3, 16, 7, [7], 1], ["SP", 3, 18, 7, [7], 1], ["SP", 3, 20, 3, [3], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CA", 1, 6, 3], ["FU", 6, [0, 1, 2]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 1, 3, 2], ["AN", 3, 21, 2], ["AN", 6, 9, 2]]]], "r": [[0.000247808, 0.000233393, 0.000251767, 0.000252226, 0.000254169, 0.000254176, 0.00025333, 0.00025511, 0.000253678, 0.000251738], 0, 0.315503, 1606961659], "v": "v0.5"} +{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 7, 7, 512, 1, 1, 512, 2048, 1, 7, 7, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 1, 1], 1], ["SP", 3, 8, 7, [1, 1, 7], 1], ["SP", 3, 12, 2048, [256, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000169437, 0.000169021, 0.00016965, 0.00017079, 0.000170862, 0.0001692, 0.000164768, 0.000175541, 0.000171528, 0.000169094], 0, 0.25194, 1606961681], "v": "v0.5"} +{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 56, 56, 256, 1, 1, 256, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [4, 1, 1], 1], ["SP", 3, 8, 56, [7, 4, 1], 1], ["SP", 3, 12, 64, [4, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 6, 5, 2]]]], "r": [[0.00015141, 0.000158121, 0.000132758, 0.00015109, 0.000148266, 0.000152599, 0.000150809, 0.000151947, 0.000150702, 0.000156091], 0, 0.221869, 1606961698], "v": "v0.5"} +{"i": [["[\"2350d19dc42a0665244368384c66b3a5\", 1, 56, 56, 64, 3, 3, 64, 64, 1, 1, 1, 64, 1, 56, 56, 64]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [1, 1, 1], 1], ["SP", 3, 8, 56, [7, 1, 4], 1], ["SP", 3, 12, 64, [1, 1, 16], 1], ["SP", 3, 16, 3, [1], 1], ["SP", 3, 18, 3, [3], 1], ["SP", 3, 20, 64, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["FU", 1, [0, 1, 2]], ["AN", 1, 0, 3], ["FU", 6, [0, 1, 2, 3, 4, 5, 6]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 6, 5, 2]]]], "r": [[0.000221341, 0.000225005, 0.000209954, 0.000209741, 0.000228281, 0.000208451, 0.000223046, 0.000222672, 0.000228098, 0.000220093], 0, 0.231218, 1606961709], "v": "v0.5"} +{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 7, 7, 2048, 1, 1, 2048, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [7, 1, 1], 1], ["SP", 3, 8, 7, [1, 7, 1], 1], ["SP", 3, 12, 512, [2, 2, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 2048, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$0"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000165941, 0.000152645, 0.000165687, 0.000166639, 0.000166094, 0.00016649, 0.000164394, 0.000169288, 0.000169497, 0.000168535], 0, 0.245559, 1606961724], "v": "v0.5"} +{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256, 1, 56, 56, 256]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [4, 2, 1], 1], ["SP", 3, 8, 56, [1, 1, 1], 1], ["SP", 3, 12, 256, [2, 4, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 5, 6, 2]]]], "r": [[0.000161206, 0.000161372, 0.000158862, 0.000159596, 0.00014964, 0.000162042, 0.000159626, 0.000158166, 0.000161209, 0.000159408], 0, 0.337652, 1606961748], "v": "v0.5"} +{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 28, 28, 512, 1, 1, 512, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 2, 1], 1], ["SP", 3, 8, 14, [14, 1, 1], 1], ["SP", 3, 12, 1024, [2, 2, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2]]]], "r": [[0.000238006, 0.000235502, 0.000239805, 0.000234637, 0.000235266, 0.000238355, 0.000240836, 0.000232856, 0.000231219, 0.000238776], 0, 0.219506, 1606961782], "v": "v0.5"} +{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 56, 56, 256, 1, 1, 256, 512, 1, 28, 28, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [2, 1, 2], 1], ["SP", 3, 8, 28, [1, 2, 1], 1], ["SP", 3, 12, 512, [16, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [8], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000213071, 0.000218117, 0.000216346, 0.000216237, 0.000214703, 0.00021605, 0.000210522, 0.000214234, 0.000218293, 0.00021484], 0, 0.291873, 1606961801], "v": "v0.5"} +{"i": [["[\"f4380bb1dc62422a69ad4a1a9771f927\", 1, 14, 14, 1024, 1, 1, 1024, 2048, 1, 7, 7, 2048]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [7, 1, 1], 1], ["SP", 3, 8, 7, [1, 7, 1], 1], ["SP", 3, 12, 2048, [128, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 1024, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000265306, 0.000259738, 0.000256412, 0.000284932, 0.000267557, 0.000266362, 0.00026533, 0.000263389, 0.000263022, 0.000263069], 0, 0.296232, 1606961838], "v": "v0.5"} +{"i": [["[\"2350d19dc42a0665244368384c66b3a5\", 1, 7, 7, 512, 3, 3, 512, 512, 1, 1, 1, 512, 1, 7, 7, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 7, [1, 7, 1], 1], ["SP", 3, 8, 7, [7, 1, 1], 1], ["SP", 3, 12, 512, [4, 1, 8], 1], ["SP", 3, 16, 3, [3], 1], ["SP", 3, 18, 3, [1], 1], ["SP", 3, 20, 512, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CR", 1], ["FU", 1, [0, 1, 2, 3]], ["AN", 1, 0, 3], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000269786, 0.0002657, 0.000261922, 0.000267462, 0.000270495, 0.000265371, 0.000273858, 0.000268022, 0.000266746, 0.000272337], 0, 0.331923, 1606961848], "v": "v0.5"} +{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 14, 14, 256, 1, 1, 256, 1024, 1, 14, 14, 1024, 1, 14, 14, 1024]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 14, [7, 1, 2], 1], ["SP", 3, 8, 14, [7, 2, 1], 1], ["SP", 3, 12, 1024, [16, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 256, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$16"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000159777, 0.00015711, 0.000163052, 0.000152569, 0.00015342, 0.000154918, 0.000153887, 0.000154133, 0.000154319, 0.000150102], 0, 0.195628, 1606961878], "v": "v0.5"} +{"i": [["[\"7006235cfc29b73be524cf390ed5a977\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CHW", 3, "local"], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [56, 1, 1], 1], ["SP", 3, 8, 56, [14, 1, 2], 1], ["SP", 3, 12, 256, [1, 2, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [64], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 4, 0, 1, 2], ["FSP", 4, 3, 2, 2], ["FSP", 4, 6, 3, 2], ["FSP", 4, 9, 4, 2], ["RE", 4, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 4, 7], ["CI", 1], ["FU", 4, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 4, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 4, 4, 2]]]], "r": [[0.000159044, 0.000157356, 0.000158889, 0.000160304, 0.000158648, 0.000159749, 0.000143679, 0.000156393, 0.000164916, 0.000155957], 0, 0.240777, 1606961918], "v": "v0.5"} +{"i": [["[\"1cc666833c122282e3fcf3595901b12b\", 1, 28, 28, 128, 1, 1, 128, 512, 1, 28, 28, 512, 1, 28, 28, 512]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [1, 1, 2], 1], ["SP", 3, 8, 28, [1, 1, 1], 1], ["SP", 3, 12, 512, [4, 1, 32], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 128, [4], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 5, 0, 0, 2], ["FSP", 5, 3, 1, 2], ["FSP", 5, 6, 2, 2], ["FSP", 5, 9, 3, 2], ["RE", 5, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 5, 7], ["CI", 1], ["FU", 5, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 5, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 5, 4, 2]]]], "r": [[0.000168259, 0.000157338, 0.0001551, 0.000156552, 0.000160492, 0.000164505, 0.000144937, 0.000138397, 0.000153011, 0.000153186], 0, 0.231498, 1606961965], "v": "v0.5"} +{"i": [["[\"1b524af89dd867d26059e1f621cf987c\", 1, 56, 56, 64, 1, 1, 64, 256, 1, 56, 56, 256, 1, 1, 1, 256, 1, 56, 56, 256]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 7], ["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 56, [7, 1, 4], 1], ["SP", 3, 8, 56, [4, 2, 1], 1], ["SP", 3, 12, 256, [32, 1, 8], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 64, [2], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 8, 0, 2, 2], ["FSP", 8, 3, 3, 2], ["FSP", 8, 6, 4, 2], ["FSP", 8, 9, 5, 2], ["RE", 8, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 8, 7], ["CI", 1], ["FU", 8, [0, 1, 2, 3, 4, 5, 6]], ["AN", 8, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$64"], ["AN", 3, 21, 2], ["AN", 8, 5, 2]]]], "r": [[0.000185957, 0.000180964, 0.000179419, 0.000168205, 0.000176155, 0.000178243, 0.000180175, 0.00017753, 0.000174475, 0.000158878], 0, 0.316404, 1606961979], "v": "v0.5"} +{"i": [["[\"6b7583cf23c7c37d3212cad9d06e58c1\", 1, 28, 28, 512, 1, 1, 512, 128, 1, 1, 1, 128, 1, 28, 28, 128]", "llvm -keys=cpu -mcpu=core-avx2", [8, 64, 64, 0, 0, 0, 0, 0], "", 2], [[], [["CI", 5], ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 28, [7, 1, 4], 1], ["SP", 3, 8, 28, [14, 1, 1], 1], ["SP", 3, 12, 128, [1, 1, 16], 1], ["SP", 3, 16, 1, [1], 1], ["SP", 3, 18, 1, [1], 1], ["SP", 3, 20, 512, [1], 1], ["RE", 3, [0, 4, 8, 12, 1, 5, 9, 13, 16, 18, 20, 2, 6, 10, 14, 17, 19, 21, 3, 7, 11, 15]], ["FSP", 6, 0, 1, 2], ["FSP", 6, 3, 2, 2], ["FSP", 6, 6, 3, 2], ["FSP", 6, 9, 4, 2], ["RE", 6, [0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11]], ["CA", 3, 6, 7], ["CI", 1], ["FU", 6, [0, 1, 2, 3, 4, 5, 6, 7]], ["AN", 6, 0, 3], ["PR", 3, 0, "auto_unroll_max_step$512"], ["AN", 3, 21, 2], ["AN", 6, 4, 2]]]], "r": [[0.000150378, 0.000154444, 0.000156051, 0.000130306, 0.000156154, 0.000131167, 0.000142357, 0.000152532, 0.000131899, 0.000157696], 0, 0.18509, 1606962002], "v": "v0.5"} diff --git a/gallery/how_to/tune_with_autoscheduler/ci_logs/sparse_dense.json b/gallery/how_to/tune_with_autoscheduler/ci_logs/sparse_dense.json index 7c1c100124dc..9bf6af0b17d8 100644 --- a/gallery/how_to/tune_with_autoscheduler/ci_logs/sparse_dense.json +++ b/gallery/how_to/tune_with_autoscheduler/ci_logs/sparse_dense.json @@ -1,2 +1,2 @@ # Keep a valid schedule for demonstraction. This is used to prevent flasky errors in CI. -{"i": [["[\"sparse_dense\", 512, 512, 512, [9831, 16, 1], [9831], [33], \"float32\"]", "llvm -keys=cpu -link-params=0", [6, 64, 64, 0, 0, 0, 0, 0], "", 1, ["sparse_dense_bsr_512_512_512_16_1_0.60_W_data", "sparse_dense_bsr_512_512_512_16_1_0.60_W_indices", "sparse_dense_bsr_512_512_512_16_1_0.60_W_indptr"]], [[], [["CI", 8], ["CI", 6], ["SP", 5, 0, 512, [1, 8], 1], ["FSP", 9, 0, 2, 1], ["SP", 5, 3, 32, [32], 1], ["FSP", 9, 2, 4, 1], ["RE", 5, [0, 3, 1, 4, 6, 2, 5, 7]], ["RE", 9, [0, 2, 1, 3]], ["CA", 5, 9, 1], ["CI", 4], ["FU", 9, [0, 1]], ["AN", 9, 0, 3], ["PR", 5, 0, "auto_unroll_max_step$0"], ["AN", 9, 2, 2]]]], "r": [[0.000957008], 0, 0.605709, 1614689820], "v": "v0.6"} +{"i": [["[\"sparse_dense\", 512, 512, 512, [9831, 16, 1], [9831], [33], \"float32\"]", "llvm -keys=cpu", [6, 64, 64, 0, 0, 0, 0, 0], "", 1, ["sparse_dense_bsr_512_512_512_16_1_0.60_W_data", "sparse_dense_bsr_512_512_512_16_1_0.60_W_indices", "sparse_dense_bsr_512_512_512_16_1_0.60_W_indptr"]], [[], [["CI", 8], ["CI", 6], ["SP", 5, 0, 512, [1, 8], 1], ["FSP", 9, 0, 2, 1], ["SP", 5, 3, 32, [32], 1], ["FSP", 9, 2, 4, 1], ["RE", 5, [0, 3, 1, 4, 6, 2, 5, 7]], ["RE", 9, [0, 2, 1, 3]], ["CA", 5, 9, 1], ["CI", 4], ["FU", 9, [0, 1]], ["AN", 9, 0, 3], ["PR", 5, 0, "auto_unroll_max_step$0"], ["AN", 9, 2, 2]]]], "r": [[0.000957008], 0, 0.605709, 1614689820], "v": "v0.6"} diff --git a/gallery/how_to/tune_with_autotvm/tune_relay_x86.py b/gallery/how_to/tune_with_autotvm/tune_relay_x86.py index 6e46fbd8ffc8..2ba597d1da19 100644 --- a/gallery/how_to/tune_with_autotvm/tune_relay_x86.py +++ b/gallery/how_to/tune_with_autotvm/tune_relay_x86.py @@ -298,7 +298,7 @@ def tune_and_evaluate(tuning_opt): # # Evaluation of the network been tuned on graph level: # Compile... -# Config for target=llvm -keys=cpu -link-params=0, workload=('dense_nopack.x86', ('TENSOR', (1, 512), 'float32'), ('TENSOR', (1000, 512), 'float32'), None, 'float32') is missing in ApplyGraphBest context. A fallback configuration is used, which may bring great performance regression. -# Config for target=llvm -keys=cpu -link-params=0, workload=('dense_pack.x86', ('TENSOR', (1, 512), 'float32'), ('TENSOR', (1000, 512), 'float32'), None, 'float32') is missing in ApplyGraphBest context. A fallback configuration is used, which may bring great performance regression. +# Config for target=llvm -keys=cpu, workload=('dense_nopack.x86', ('TENSOR', (1, 512), 'float32'), ('TENSOR', (1000, 512), 'float32'), None, 'float32') is missing in ApplyGraphBest context. A fallback configuration is used, which may bring great performance regression. +# Config for target=llvm -keys=cpu, workload=('dense_pack.x86', ('TENSOR', (1, 512), 'float32'), ('TENSOR', (1000, 512), 'float32'), None, 'float32') is missing in ApplyGraphBest context. A fallback configuration is used, which may bring great performance regression. # Evaluate inference time cost... # Mean inference time (std dev): 3.16 ms (0.03 ms) diff --git a/gallery/how_to/work_with_microtvm/micro_tvmc.sh b/gallery/how_to/work_with_microtvm/micro_tvmc.sh index 5ec718884559..0eaef9c6a836 100755 --- a/gallery/how_to/work_with_microtvm/micro_tvmc.sh +++ b/gallery/how_to/work_with_microtvm/micro_tvmc.sh @@ -99,7 +99,7 @@ wget https://github.com/tensorflow/tflite-micro/raw/main/tensorflow/lite/micro/e # # bash tvmc compile magic_wand.tflite \ - --target='c -keys=cpu -link-params=0 -model=host' \ + --target='c -keys=cpu -model=host' \ --runtime=crt \ --runtime-crt-system-lib 1 \ --executor='graph' \ @@ -111,7 +111,7 @@ tvmc compile magic_wand.tflite \ # bash # This will generate a ``model.tar`` file which contains TVM compiler output files. To run this command for # a different Zephyr device, you need to update ``target``. For instance, for ``nrf5340dk_nrf5340_cpuapp`` board -# the target is ``--target='c -keys=cpu -link-params=0 -model=nrf5340dk'``. +# the target is ``--target='c -keys=cpu -model=nrf5340dk'``. # diff --git a/gallery/tutorial/auto_scheduler_matmul_x86.py b/gallery/tutorial/auto_scheduler_matmul_x86.py index 279987f00d81..98fd95c33878 100644 --- a/gallery/tutorial/auto_scheduler_matmul_x86.py +++ b/gallery/tutorial/auto_scheduler_matmul_x86.py @@ -44,8 +44,6 @@ testing.utils.install_request_hook(depth=3) # sphinx_gallery_end_ignore -import os - import numpy as np import tvm from tvm import te, auto_scheduler diff --git a/python/tvm/contrib/hexagon/pytest_plugin.py b/python/tvm/contrib/hexagon/pytest_plugin.py index 6b61b6f4ba55..f735c81ee0aa 100644 --- a/python/tvm/contrib/hexagon/pytest_plugin.py +++ b/python/tvm/contrib/hexagon/pytest_plugin.py @@ -245,7 +245,7 @@ def terminate_rpc_servers(): aot_host_target = tvm.testing.parameter( "c", - "llvm -keys=hexagon -link-params=0 " + "llvm -keys=hexagon " "-mattr=+hvxv68,+hvx-length128b,+hvx-qfloat,-hvx-ieee-fp " "-mcpu=hexagonv68 -mtriple=hexagon", ) diff --git a/python/tvm/relay/build_module.py b/python/tvm/relay/build_module.py index f3de1a085692..6cdc79ceb587 100644 --- a/python/tvm/relay/build_module.py +++ b/python/tvm/relay/build_module.py @@ -274,69 +274,6 @@ def _build_module_no_factory(mod, target=None, target_host=None, params=None, mo return _build_module_no_factory_impl(mod, target, target_host, params, mod_name) -def _reconstruct_from_deprecated_options(deprecated_params_target): - executor = None - runtime = None - - deprecated_executor = None - deprecated_executor_args = {} - if "executor" in deprecated_params_target.attrs: - _deprecated_target_param_warning("Executor", "executor") - deprecated_executor = deprecated_params_target.attrs.get("executor", "graph") - if "interface-api" in deprecated_params_target.attrs: - _deprecated_target_sub_param_warning("Executor", "interface-api") - deprecated_executor_args.update( - {"interface-api": deprecated_params_target.attrs["interface-api"]} - ) - if "unpacked-api" in deprecated_params_target.attrs: - _deprecated_target_sub_param_warning("Executor", "unpacked-api") - deprecated_executor_args.update( - {"unpacked-api": deprecated_params_target.attrs["unpacked-api"]} - ) - if ( - "link-params" in deprecated_params_target.attrs - and deprecated_params_target.attrs["link-params"] - ): - _deprecated_target_sub_param_warning("Executor", "link-params") - if deprecated_executor != "aot": - deprecated_executor_args.update( - {"link-params": deprecated_params_target.attrs["link-params"]} - ) - if deprecated_executor or deprecated_executor_args: - executor = Executor(deprecated_executor or "graph", deprecated_executor_args) - - deprecated_runtime = None - deprecated_runtime_args = {} - if "runtime" in deprecated_params_target.attrs: - _deprecated_target_param_warning("Runtime", "runtime") - deprecated_runtime = deprecated_params_target.attrs.get("runtime", "cpp") - if deprecated_runtime == "c": - deprecated_runtime = "crt" - if "system-lib" in deprecated_params_target.attrs: - _deprecated_target_sub_param_warning("Runtime", "system-lib") - deprecated_runtime_args.update({"system-lib": deprecated_params_target.attrs["system-lib"]}) - if deprecated_runtime or deprecated_runtime_args: - runtime = Runtime(deprecated_runtime or "cpp", deprecated_runtime_args) - - return executor, runtime - - -def _deprecated_target_param_warning(registry, param): - warnings.warn( - f"Please use {registry} (tvm.relay.backend.{registry}) " - f"instead of deprecated Target parameter -{param}", - DeprecationWarning, - ) - - -def _deprecated_target_sub_param_warning(registry, param): - warnings.warn( - f"Please use {registry} (tvm.relay.backend.{registry}) parameter {param} " - f"instead of deprecated Target parameter -{param}", - DeprecationWarning, - ) - - def build( ir_mod, target=None, @@ -415,17 +352,6 @@ def build( assert len(raw_targets) > 0 target_host = raw_targets[0].host - # All of this logic is to raise deprecation warnings for various parameters - # TODO(Mousius) Remove these after some time - deprecated_params_target = target_host or list(raw_targets)[0] - deprecated_executor, deprecated_runtime = _reconstruct_from_deprecated_options( - deprecated_params_target - ) - if deprecated_executor: - executor = deprecated_executor - if deprecated_runtime: - runtime = deprecated_runtime - # If current dispatch context is fallback context (the default root context), # then load pre-tuned parameters from TopHub if isinstance(autotvm.DispatchContext.current, autotvm.FallbackContext): @@ -756,9 +682,5 @@ def create_executor(kind="debug", mod=None, device=None, target="llvm", params=N if kind == "vm": return VMExecutor(mod, device, raw_targets) if kind == "aot": - # The AOT requires the executor as a target attribute. - # (The compilation paths for the other executors currently do not always provide this - # attribute, hence the above generic assert is more forgiving). - assert "executor" in raw_targets[0].attrs return AotExecutor(mod, device, raw_targets) raise RuntimeError("unknown execution strategy: {0}".format(kind)) diff --git a/python/tvm/target/target.py b/python/tvm/target/target.py index e0e5f0177b5e..a558fcbeaf5b 100644 --- a/python/tvm/target/target.py +++ b/python/tvm/target/target.py @@ -636,8 +636,6 @@ def hexagon(cpu_ver="v66", **kwargs): Whether to use QFloat HVX instructions. use_ieee_fp : bool (default: False) Whether to use IEEE HVX instructions - link_params : bool (default: False) - Whether to link graph parameters into the LLVM module. Note: Floating point support in HVX requires LLVM 14+. """ @@ -671,7 +669,6 @@ def get_arch_version(cpu_ver): "llvm_options": None, "use_qfloat": arch_version >= 68, "use_ieee_fp": False, - "link_params": False, } config.update(kwargs) @@ -738,24 +735,10 @@ def create_llvm_options(cpu_ver, config): # pylint: disable=unused-argument args = [s.replace("=", "@") for s in llvm_options.split()] return "--llvm-options=" + ",".join(args) - # TVM target attributes string - def create_tvm_options(cpu_ver, config): # pylint: disable=unused-argument - """Create TVM target features string.""" - - features = { - "link_params": "link-params", - } - opts = "" - for k in config: - if k in features: - opts += " --" + features[k] + "=" + str(config[k]) - return opts - target_str = create_llvm_target(cpu_ver, config) llvm_str = create_llvm_options(cpu_ver, config) - tvm_str = create_tvm_options(cpu_ver, config) - args_list = target_str.split() + llvm_str.split() + tvm_str.split() + args_list = target_str.split() + llvm_str.split() return Target(" ".join(["hexagon"] + args_list)) diff --git a/src/target/target_kind.cc b/src/target/target_kind.cc index 38ee536e7818..e3b2d7b096fd 100644 --- a/src/target/target_kind.cc +++ b/src/target/target_kind.cc @@ -264,12 +264,7 @@ TVM_REGISTER_TARGET_KIND("llvm", kDLCPU) .add_attr_option("mtriple") .add_attr_option("mfloat-abi") .add_attr_option("mabi") - .add_attr_option("system-lib") - .add_attr_option("runtime") .add_attr_option("num-cores") - .add_attr_option("link-params", Bool(false)) - .add_attr_option("unpacked-api") - .add_attr_option("interface-api") // Fast math flags, see https://llvm.org/docs/LangRef.html#fast-math-flags .add_attr_option("fast-math") // implies all the below .add_attr_option("fast-math-nnan") @@ -310,23 +305,16 @@ TVM_REGISTER_TARGET_KIND("llvm", kDLCPU) // Hence the type is "uint". TVM_REGISTER_TARGET_KIND("c", kDLCPU) - .add_attr_option("system-lib") - .add_attr_option("link-params", Bool(false)) - .add_attr_option("runtime") .add_attr_option("mcpu") .add_attr_option("march") - .add_attr_option("executor") .add_attr_option("workspace-byte-alignment") .add_attr_option("constants-byte-alignment") - .add_attr_option("unpacked-api") - .add_attr_option("interface-api") .set_default_keys({"cpu"}) .set_target_parser(tvm::target::parsers::cpu::ParseTarget); TVM_REGISTER_TARGET_KIND("cuda", kDLCUDA) .add_attr_option("mcpu") .add_attr_option("arch") - .add_attr_option("system-lib") .add_attr_option("max_shared_memory_per_block") .add_attr_option("max_threads_per_block") .add_attr_option("thread_warp_size", Integer(32)) @@ -338,7 +326,6 @@ TVM_REGISTER_TARGET_KIND("cuda", kDLCUDA) TVM_REGISTER_TARGET_KIND("nvptx", kDLCUDA) .add_attr_option("mcpu") .add_attr_option("mtriple") - .add_attr_option("system-lib") .add_attr_option("max_num_threads", Integer(1024)) .add_attr_option("thread_warp_size", Integer(32)) .set_default_keys({"cuda", "gpu"}) @@ -348,7 +335,6 @@ TVM_REGISTER_TARGET_KIND("rocm", kDLROCM) .add_attr_option("mcpu") .add_attr_option("mtriple") .add_attr_option>("mattr") - .add_attr_option("system-lib") // TODO(masahi): Support querying from a target device // On RDNA cards, thread_warp_size should be 32 .add_attr_option("max_num_threads", Integer(256)) @@ -359,7 +345,6 @@ TVM_REGISTER_TARGET_KIND("rocm", kDLROCM) .set_target_parser(UpdateROCmAttrs); TVM_REGISTER_TARGET_KIND("opencl", kDLOpenCL) - .add_attr_option("system-lib") .add_attr_option("max_num_threads", Integer(256)) .add_attr_option("thread_warp_size", Integer(1)) .add_attr_option("texture_spatial_limit", Integer(16384)) @@ -370,7 +355,6 @@ TVM_REGISTER_TARGET_KIND("opencl", kDLOpenCL) // information about this limitation can be found here: // https://developer.apple.com/documentation/metal/buffers/about_argument_buffers?language=objc TVM_REGISTER_TARGET_KIND("metal", kDLMetal) - .add_attr_option("system-lib") .add_attr_option("max_num_threads", Integer(256)) .add_attr_option("thread_warp_size", Integer(16)) .add_attr_option("max_function_args", Integer(31)) @@ -378,7 +362,6 @@ TVM_REGISTER_TARGET_KIND("metal", kDLMetal) TVM_REGISTER_TARGET_KIND("vulkan", kDLVulkan) .add_attr_option>("mattr") - .add_attr_option("system-lib") // Feature support .add_attr_option("supports_float16") .add_attr_option("supports_float32", Bool(true)) @@ -417,39 +400,30 @@ TVM_REGISTER_TARGET_KIND("vulkan", kDLVulkan) .set_default_keys({"vulkan", "gpu"}); TVM_REGISTER_TARGET_KIND("webgpu", kDLWebGPU) - .add_attr_option("system-lib") .add_attr_option("max_num_threads", Integer(256)) .set_default_keys({"webgpu", "gpu"}); -TVM_REGISTER_TARGET_KIND("sdaccel", kDLOpenCL) - .add_attr_option("system-lib") +TVM_REGISTER_TARGET_KIND("sdaccel", kDLOpenCL) // line break .set_default_keys({"sdaccel", "hls"}); -TVM_REGISTER_TARGET_KIND("aocl", kDLAOCL) - .add_attr_option("system-lib") +TVM_REGISTER_TARGET_KIND("aocl", kDLAOCL) // line break .set_default_keys({"aocl", "hls"}); -TVM_REGISTER_TARGET_KIND("aocl_sw_emu", kDLAOCL) - .add_attr_option("system-lib") +TVM_REGISTER_TARGET_KIND("aocl_sw_emu", kDLAOCL) // line break .set_default_keys({"aocl", "hls"}); TVM_REGISTER_TARGET_KIND("hexagon", kDLHexagon) .add_attr_option>("mattr") .add_attr_option("mcpu") .add_attr_option("mtriple") - .add_attr_option("system-lib") - .add_attr_option("link-params", Bool(false)) .add_attr_option>("llvm-options") .set_default_keys({"hexagon"}); -TVM_REGISTER_TARGET_KIND("stackvm", kDLCPU) // line break - .add_attr_option("system-lib"); +TVM_REGISTER_TARGET_KIND("stackvm", kDLCPU); -TVM_REGISTER_TARGET_KIND("ext_dev", kDLExtDev) // line break - .add_attr_option("system-lib"); +TVM_REGISTER_TARGET_KIND("ext_dev", kDLExtDev); -TVM_REGISTER_TARGET_KIND("hybrid", kDLCPU) // line break - .add_attr_option("system-lib"); +TVM_REGISTER_TARGET_KIND("hybrid", kDLCPU); TVM_REGISTER_TARGET_KIND("composite", kDLCPU) // line break .add_attr_option>("devices"); diff --git a/tests/cpp/c_codegen_test.cc b/tests/cpp/c_codegen_test.cc index 442f76a8cff3..e764d21505d4 100644 --- a/tests/cpp/c_codegen_test.cc +++ b/tests/cpp/c_codegen_test.cc @@ -33,7 +33,7 @@ TEST(CCodegen, MainFunctionOrder) { std::string tvm_module_main = std::string(runtime::symbol::tvm_module_main); - tvm::Target target_c = tvm::Target("c -keys=cpu -link-params=0"); + tvm::Target target_c = tvm::Target("c -keys=cpu"); const int n = 4; Array shape{n}; @@ -104,16 +104,16 @@ TEST(CCodegen, FunctionOrder) { using namespace tvm; using namespace tvm::te; - Target target = Target("c -keys=cpu -link-params=0"); + Target target = Target("c -keys=cpu"); // add schedules in reverse order Map inputs; - inputs.Set(Target("c -keys=cpu -link-params=0"), BuildLowered("op_2", target)); - inputs.Set(Target("c -keys=cpu -link-params=0"), BuildLowered("op_1", target)); + inputs.Set(Target("c -keys=cpu"), BuildLowered("op_2", target)); + inputs.Set(Target("c -keys=cpu"), BuildLowered("op_1", target)); for (uint32_t counter = 99; IsSorted(inputs) && counter > 0; counter--) { std::string op_name = "op_" + std::to_string(counter); - inputs.Set(Target("c -keys=cpu -link-params=0"), BuildLowered(op_name, target)); + inputs.Set(Target("c -keys=cpu"), BuildLowered(op_name, target)); } EXPECT_FALSE(IsSorted(inputs)); diff --git a/tests/cpp/target_test.cc b/tests/cpp/target_test.cc index f238393ce923..37a8eeb44840 100644 --- a/tests/cpp/target_test.cc +++ b/tests/cpp/target_test.cc @@ -493,9 +493,8 @@ TEST(TargetCreation, DeduplicateKeys) { ICHECK_EQ(target->keys.size(), 2U); ICHECK_EQ(target->keys[0], "cpu"); ICHECK_EQ(target->keys[1], "arm_cpu"); - ICHECK_EQ(target->attrs.size(), 2U); + ICHECK_EQ(target->attrs.size(), 1U); ICHECK_EQ(target->GetAttr("device"), "arm_cpu"); - ICHECK_EQ(target->GetAttr("link-params"), false); } TEST(TargetKindRegistry, ListTargetKinds) { @@ -511,5 +510,4 @@ TEST(TargetKindRegistry, ListTargetOptions) { ICHECK_EQ(attrs["mattr"], "Array"); ICHECK_EQ(attrs["mcpu"], "runtime.String"); - ICHECK_EQ(attrs["system-lib"], "IntImm"); } diff --git a/tests/python/contrib/test_hexagon/topi/test_softmax_slice.py b/tests/python/contrib/test_hexagon/topi/test_softmax_slice.py index 9bbecdd7f81b..91b51cb5cc75 100644 --- a/tests/python/contrib/test_hexagon/topi/test_softmax_slice.py +++ b/tests/python/contrib/test_hexagon/topi/test_softmax_slice.py @@ -14,7 +14,6 @@ # KIND, either express or implied. See the License for the # specific language governing permissions and limitations # under the License. -import pytest import numpy as np import tvm diff --git a/tests/python/driver/tvmc/test_target.py b/tests/python/driver/tvmc/test_target.py index 4438ec437cb4..39e90e6d6ac4 100644 --- a/tests/python/driver/tvmc/test_target.py +++ b/tests/python/driver/tvmc/test_target.py @@ -114,9 +114,7 @@ def test_parse_multiple_target(): def test_parse_hybrid_target(): """Hybrid Target and external codegen""" - targets = parse_target( - "cmsis-nn -accelerator_config=ethos-u55-256, llvm -device=arm_cpu --system-lib" - ) + targets = parse_target("cmsis-nn -accelerator_config=ethos-u55-256, llvm -device=arm_cpu") assert len(targets) == 2 assert "cmsis-nn" == targets[0]["name"] @@ -154,7 +152,7 @@ def test_parse_quotes_and_separators_on_options(): def test_parse_multiple_target_with_opts_ethos_n78(): - targets = parse_target("ethos-n -myopt=value, llvm -device=arm_cpu --system-lib") + targets = parse_target("ethos-n -myopt=value, llvm -device=arm_cpu") assert len(targets) == 2 assert "ethos-n" == targets[0]["name"] diff --git a/tests/python/driver/tvmc/test_target_options.py b/tests/python/driver/tvmc/test_target_options.py index c73dc288cdd8..891df86f0c1f 100644 --- a/tests/python/driver/tvmc/test_target_options.py +++ b/tests/python/driver/tvmc/test_target_options.py @@ -86,7 +86,7 @@ def test_skip_target_from_codegen(): def test_target_recombobulation_single(): tvm_target, _ = target_from_cli("llvm", {"llvm": {"mcpu": "cortex-m3"}}) - assert str(tvm_target) == "llvm -keys=arm_cpu,cpu -link-params=0 -mcpu=cortex-m3" + assert str(tvm_target) == "llvm -keys=arm_cpu,cpu -mcpu=cortex-m3" def test_target_recombobulation_many(): diff --git a/tests/python/relay/aot/test_cpp_aot.py b/tests/python/relay/aot/test_cpp_aot.py index 3f641c995652..4ffe302763f8 100644 --- a/tests/python/relay/aot/test_cpp_aot.py +++ b/tests/python/relay/aot/test_cpp_aot.py @@ -169,7 +169,7 @@ def test_create_executor(): x = tvm.relay.var("x", tvm.relay.TensorType([1], dtype="float32")) expr = tvm.relay.add(x, tvm.relay.Constant(tvm.nd.array(np.array([1], dtype="float32")))) actual = relay.create_executor( - "aot", mod=tvm.IRModule.from_expr(tvm.relay.Function([x], expr)), target="c -executor=aot" + "aot", mod=tvm.IRModule.from_expr(tvm.relay.Function([x], expr)), target="c" ).evaluate()(np.array([2], dtype="float32")) np.isfinite(np.array([3], dtype="float32")) diff --git a/tests/python/relay/aot/test_crt_aot.py b/tests/python/relay/aot/test_crt_aot.py index edf23ff22781..c3426f147e0d 100644 --- a/tests/python/relay/aot/test_crt_aot.py +++ b/tests/python/relay/aot/test_crt_aot.py @@ -710,42 +710,6 @@ def test_name_sanitiser_name_clash(): ) -# This tests for deprecated AOT executor arguments -# TODO(Mousius) Remove deprecated arguments later -def test_deprecated_target_arguments(): - """Tests we can still use relay.build with -executor, -runtime and -link-params""" - - interface_api = "c" - use_unpacked_api = True - test_runner = AOT_DEFAULT_RUNNER - - input_x = relay.var("x", shape=(1, 10)) - input_y = relay.var("y", shape=(1, 10)) - func_add = relay.add(input_x, input_y) - func = relay.Function([input_x, input_y], func_add) - - x_in = np.ones((1, 10)).astype("float32") - y_in = np.random.uniform(size=(1, 10)).astype("float32") - - params = {"x": x_in} - inputs = {"y": y_in} - output_list = generate_ref_data(func, inputs, params) - - compile_and_run( - AOTTestModel( - module=IRModule.from_expr(func), - inputs=inputs, - outputs=output_list, - params=params, - ), - test_runner, - interface_api, - use_unpacked_api, - use_runtime_executor=False, - target="c -executor=aot --link-params -runtime=c -interface-api=c --unpacked-api", - ) - - def test_aot_codegen_backend_alloc_workspace_calls(): """This test checks whether AoT lowering creates TVMBackendAllocWorkspace calls""" diff --git a/tests/python/relay/test_build_module.py b/tests/python/relay/test_build_module.py index d51cfd29dc97..5cfc27330aff 100644 --- a/tests/python/relay/test_build_module.py +++ b/tests/python/relay/test_build_module.py @@ -22,48 +22,23 @@ from tvm import relay from tvm.target.target import Target from tvm.relay.backend import Runtime, Executor, graph_executor_codegen -from tvm.relay.build_module import _reconstruct_from_deprecated_options @pytest.mark.parametrize( - "target,executor,runtime", + "test_target,unsupported_config", [ - [Target("c"), None, None], - [Target("c -runtime=c"), None, Runtime("crt")], - [Target("c -system-lib"), None, Runtime("cpp", {"system-lib": True})], - [Target("c -runtime=c -system-lib"), None, Runtime("crt", {"system-lib": True})], - [Target("c -executor=aot"), Executor("aot"), None], - [ - Target("c -executor=aot -interface-api=c"), - Executor("aot", {"interface-api": "c"}), - None, - ], - [ - Target("c -executor=aot -unpacked-api=1"), - Executor("aot", {"unpacked-api": 1}), - None, - ], - [Target("c -executor=aot -link-params=1"), Executor("aot"), None], - [Target("c -link-params=1"), Executor("graph", {"link-params": 1}), None], - [ - Target( - "c -executor=aot -link-params=1 -interface-api=c" - " -unpacked-api=1 -runtime=c -system-lib" - ), - Executor("aot", {"unpacked-api": 1, "interface-api": "c"}), - Runtime("crt", {"system-lib": True}), - ], + ["c", "-runtime=c"], + ["c", "-system-lib=1"], + ["c", "-executor=aot"], + ["c", "-interface-api=c"], + ["c", "-unpacked-api=1"], + ["c", "-link-params=1"], ], ) -def test_deprecated_target_parameters(target, executor, runtime): - actual_executor, actual_runtime = _reconstruct_from_deprecated_options(target) - - assert (executor is None and actual_executor is None) or (executor.name == actual_executor.name) - # sort as TVM Map cannot guarantee round-trip order. - assert (executor is None and actual_executor is None) or ( - sorted(executor.attrs.items()) == sorted(actual_executor.attrs.items()) - ) - assert runtime == actual_runtime +def test_deprecated_target_parameters(test_target, unsupported_config): + with pytest.raises(ValueError) as e_info: + Target(f"{test_target} {unsupported_config}") + assert f"Cannot recognize '{unsupported_config}" in str(e_info.execption) def test_build_relay_graph_(): diff --git a/tests/python/unittest/test_tir_transform_common_subexpr_elim.py b/tests/python/unittest/test_tir_transform_common_subexpr_elim.py index a546c16a648e..be229a580f01 100644 --- a/tests/python/unittest/test_tir_transform_common_subexpr_elim.py +++ b/tests/python/unittest/test_tir_transform_common_subexpr_elim.py @@ -449,7 +449,7 @@ def test_deterministic_cse(): # Needed for the second test on determinism LOG_LINE = '{"i": [["[\\"conv2d_layer\\", 1, 7, 7, 512, 512, 3, 3, [1, 1], [1, 1]]", \ - "llvm -keys=cpu -link-params=0 -mcpu=broadwell -num-cores=2", \ + "llvm -keys=cpu -mcpu=broadwell -num-cores=2", \ [8, 64, 64, 0, 0, 0, 0, 0], "", 1, []], [[], [["CI", 5], \ ["SP", 3, 0, 1, [1, 1, 1], 1], ["SP", 3, 4, 512, [1, 32, 16], 1], \ ["SP", 3, 8, 7, [7, 1, 1], 1], ["SP", 3, 12, 7, [1, 1, 1], 1], \ diff --git a/tests/python/unittest/test_tvmscript_roundtrip.py b/tests/python/unittest/test_tvmscript_roundtrip.py index 0a2cec6011ef..e5f5ae752aac 100644 --- a/tests/python/unittest/test_tvmscript_roundtrip.py +++ b/tests/python/unittest/test_tvmscript_roundtrip.py @@ -3087,9 +3087,7 @@ def func_with_target_spec_by_config() -> None: "kind": "cuda", "tag": "", "keys": ["cuda", "gpu"], - "host": T.target( - {"kind": "llvm", "tag": "", "keys": ["cpu"], "link-params": False} - ), + "host": T.target({"kind": "llvm", "tag": "", "keys": ["cpu"]}), } ) } diff --git a/tests/scripts/task_python_docs.sh b/tests/scripts/task_python_docs.sh index 8b390c962e98..d8578fde2817 100755 --- a/tests/scripts/task_python_docs.sh +++ b/tests/scripts/task_python_docs.sh @@ -83,7 +83,7 @@ IGNORED_WARNINGS=( 'strategy:depthwise_conv2d NHWC layout is not optimized for x86 with autotvm.' 'strategy:depthwise_conv2d with layout NHWC is not optimized for arm cpu.' 'strategy:dense is not optimized for arm cpu.' - 'autotvm:Cannot find config for target=llvm -keys=cpu -link-params=0' + 'autotvm:Cannot find config for target=llvm -keys=cpu' 'autotvm:One or more operators have not been tuned. Please tune your model for better performance. Use DEBUG logging level to see more details.' 'autotvm:Cannot find config for target=cuda -keys=cuda,gpu' # Warning is thrown during TFLite quantization for micro_train tutorial