| #!/usr/bin/env python3 |
| |
| # |
| # Copyright (c) 2021-2022 Arm Limited. |
| # |
| # SPDX-License-Identifier: Apache-2.0 |
| # |
| # Licensed under the Apache License, Version 2.0 (the License); you may |
| # not use this file except in compliance with the License. |
| # You may obtain a copy of the License at |
| # |
| # www.apache.org/licenses/LICENSE-2.0 |
| # |
| # Unless required by applicable law or agreed to in writing, software |
| # distributed under the License is distributed on an AS IS BASIS, WITHOUT |
| # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| # See the License for the specific language governing permissions and |
| # limitations under the License. |
| # |
| |
| import argparse |
| import multiprocessing |
| import numpy |
| import os |
| import pathlib |
| import re |
| import shutil |
| import subprocess |
| import sys |
| |
| os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' |
| from tensorflow.lite.python.interpreter import Interpreter, OpResolverType |
| |
| CORE_PLATFORM_PATH = pathlib.Path(__file__).resolve().parents[1] |
| |
| def run_cmd(cmd, **kwargs): |
| # str() is called to handle pathlib.Path objects |
| cmd_str = " ".join([str(arg) for arg in cmd]) |
| print(f"Running command: {cmd_str}") |
| return subprocess.run(cmd, check=True, **kwargs) |
| |
| def ta_parse_raw(ta_raw): |
| ta_parsed = [-1, -1] |
| if ta_raw: |
| for v in ta_raw: |
| index = v[0] |
| value = v[1] |
| if (index > 1): |
| raise Exception("Illegal index value - Should be '0' or '1'") |
| (ta_parsed)[index] = value |
| |
| return ta_parsed |
| |
| def build_core_platform(output_folder, target, toolchain, memory_model, memory_arena, pmu, |
| ta_maxr, ta_maxw, ta_maxrw, ta_rlatency, ta_wlatency, |
| ta_pulse_on, ta_pulse_off, ta_bwcap, ta_perfctrl, ta_perfcnt, |
| ta_mode, ta_histbin, ta_histcnt): |
| build_folder = output_folder/"model"/"build" |
| maxr = ta_parse_raw(ta_maxr) |
| maxw = ta_parse_raw(ta_maxw) |
| maxrw = ta_parse_raw(ta_maxrw) |
| rlatency = ta_parse_raw(ta_rlatency) |
| wlatency = ta_parse_raw(ta_wlatency) |
| pulse_on = ta_parse_raw(ta_pulse_on) |
| pulse_off = ta_parse_raw(ta_pulse_off) |
| bwcap = ta_parse_raw(ta_bwcap) |
| perfctrl = ta_parse_raw(ta_perfctrl) |
| perfcnt = ta_parse_raw(ta_perfcnt) |
| mode = ta_parse_raw(ta_mode) |
| histbin = ta_parse_raw(ta_histbin) |
| histcnt = ta_parse_raw(ta_histcnt) |
| cmake_cmd = ["cmake", |
| CORE_PLATFORM_PATH/"targets"/target, |
| f"-B{build_folder}", |
| f"-DCMAKE_TOOLCHAIN_FILE={CORE_PLATFORM_PATH/'cmake'/'toolchain'/(toolchain + '.cmake')}", |
| f"-DBAREMETAL_PATH={output_folder}", |
| f"-DMEMORY_MODEL={memory_model}", |
| f"-DMEMORY_ARENA={memory_arena}", |
| f"-DETHOSU_TA_MAXR_0={maxr[0]}", |
| f"-DETHOSU_TA_MAXR_1={maxr[1]}", |
| f"-DETHOSU_TA_MAXW_0={maxw[0]}", |
| f"-DETHOSU_TA_MAXW_1={maxw[1]}", |
| f"-DETHOSU_TA_MAXRW_0={maxrw[0]}", |
| f"-DETHOSU_TA_MAXRW_1={maxrw[1]}", |
| f"-DETHOSU_TA_RLATENCY_0={rlatency[0]}", |
| f"-DETHOSU_TA_RLATENCY_1={rlatency[1]}", |
| f"-DETHOSU_TA_WLATENCY_0={wlatency[0]}", |
| f"-DETHOSU_TA_WLATENCY_1={wlatency[1]}", |
| f"-DETHOSU_TA_PULSE_ON_0={pulse_on[0]}", |
| f"-DETHOSU_TA_PULSE_ON_1={pulse_on[1]}", |
| f"-DETHOSU_TA_PULSE_OFF_0={pulse_off[0]}", |
| f"-DETHOSU_TA_PULSE_OFF_1={pulse_off[1]}", |
| f"-DETHOSU_TA_BWCAP_0={bwcap[0]}", |
| f"-DETHOSU_TA_BWCAP_1={bwcap[1]}", |
| f"-DETHOSU_TA_PERFCTRL_0={perfctrl[0]}", |
| f"-DETHOSU_TA_PERFCTRL_1={perfctrl[1]}", |
| f"-DETHOSU_TA_PERFCNT_0={perfcnt[0]}", |
| f"-DETHOSU_TA_PERFCNT_1={perfcnt[1]}", |
| f"-DETHOSU_TA_MODE_0={mode[0]}", |
| f"-DETHOSU_TA_MODE_1={mode[1]}", |
| f"-DETHOSU_TA_HISTBIN_0={histbin[0]}", |
| f"-DETHOSU_TA_HISTBIN_1={histbin[1]}", |
| f"-DETHOSU_TA_HISTCNT_0={histcnt[0]}", |
| f"-DETHOSU_TA_HISTCNT_1={histcnt[1]}"] |
| |
| if pmu: |
| for i in range(len(pmu)): |
| cmake_cmd += [f"-DETHOSU_PMU_EVENT_{i}={pmu[i]}"] |
| run_cmd(cmake_cmd) |
| |
| make_cmd = ["make", "-C", build_folder, f"-j{multiprocessing.cpu_count()}", "baremetal_custom"] |
| run_cmd(make_cmd) |
| |
| def generate_reference_data(output_folder, non_optimized_model_path, input_path, expected_output_path): |
| interpreter = Interpreter(model_path=str(non_optimized_model_path.resolve()), experimental_op_resolver_type=OpResolverType.BUILTIN_REF) |
| |
| interpreter.allocate_tensors() |
| input_detail = interpreter.get_input_details()[0] |
| output_detail = interpreter.get_output_details()[0] |
| |
| input_data = None |
| if input_path is None: |
| # Randomly generate input data |
| dtype = input_detail["dtype"] |
| if dtype is numpy.float32: |
| rand = numpy.random.default_rng() |
| input_data = rand.random(size=input_detail["shape"], dtype=numpy.float32) |
| else: |
| input_data = numpy.random.randint(low=numpy.iinfo(dtype).min, high=numpy.iinfo(dtype).max, size=input_detail["shape"], dtype=dtype) |
| else: |
| # Load user provided input data |
| input_data = numpy.load(input_path) |
| |
| output_data = None |
| if expected_output_path is None: |
| # Run the network with input_data to get reference output |
| interpreter.set_tensor(input_detail["index"], input_data) |
| interpreter.invoke() |
| output_data = interpreter.get_tensor(output_detail["index"]) |
| else: |
| # Load user provided output data |
| output_data = numpy.load(expected_output_path) |
| |
| network_input_path = output_folder/"ref_input.bin" |
| network_output_path = output_folder/"ref_output.bin" |
| |
| with network_input_path.open("wb") as fp: |
| fp.write(input_data.tobytes()) |
| with network_output_path.open("wb") as fp: |
| fp.write(output_data.tobytes()) |
| |
| output_folder = pathlib.Path(output_folder) |
| dump_c_header(network_input_path, output_folder/"input.h", "inputData", "input_data_sec", 4) |
| dump_c_header(network_output_path, output_folder/"output.h", "expectedOutputData", "expected_output_data_sec", 4) |
| |
| def dump_c_header(input_path, output_path, array_name, section, alignment, extra_data=""): |
| byte_array = [] |
| with open(input_path, "rb") as fp: |
| byte_string = fp.read() |
| byte_array = [f"0x{format(byte, '02x')}" for byte in byte_string] |
| |
| last = byte_array[-1] |
| byte_array = [byte + "," for byte in byte_array[:-1]] + [last] |
| |
| byte_array = [" " + byte if idx % 12 == 0 else byte |
| for idx, byte in enumerate(byte_array)] |
| |
| byte_array = [byte + "\n" if (idx + 1) % 12 == 0 else byte + " " |
| for idx, byte in enumerate(byte_array)] |
| |
| with open(output_path, "w") as carray: |
| header = f"uint8_t {array_name}[] __attribute__((section(\"{section}\"), aligned({alignment}))) = {{\n" |
| carray.write(extra_data) |
| carray.write(header) |
| carray.write("".join(byte_array)) |
| carray.write("\n};\n") |
| |
| def optimize_network(output_folder, network_path, accelerator_conf): |
| vela_cmd = ["vela", |
| network_path, |
| "--output-dir", output_folder, |
| "--accelerator-config", accelerator_conf] |
| res = run_cmd(vela_cmd) |
| optimized_model_path = output_folder/(network_path.stem + "_vela.tflite") |
| model_name = network_path.stem |
| dump_c_header(optimized_model_path, output_folder/"model.h", "networkModelData", "network_model_sec", 16, extra_data=f""" |
| #include <stddef.h> |
| |
| const size_t tensorArenaSize = 2000000; |
| const char* modelName = \"{model_name}\"; |
| """) |
| |
| def run_model(output_folder): |
| build_folder = output_folder/"model"/"build" |
| model_cmd = ["ctest", "-V", "-R", "^baremetal_custom$" ] |
| res = run_cmd(model_cmd, cwd=build_folder) |
| |
| def main(): |
| target_mapping = { |
| "corstone-300": "ethos-u55-128" |
| } |
| parser = argparse.ArgumentParser() |
| parser.add_argument("-o", "--output-folder", type=pathlib.Path, default="output", help="Output folder for build and generated files") |
| parser.add_argument("--network-path", type=pathlib.Path, required=True, help="Path to .tflite file") |
| parser.add_argument("--target", choices=target_mapping, default="corstone-300", help=f"Configure target") |
| parser.add_argument("--toolchain", choices=["armclang", "arm-none-eabi-gcc"], default="armclang", help=f"Configure toolchain") |
| parser.add_argument("--memory_model", choices=["sram", "dram"], default="dram", help=f"Configure memory_model") |
| parser.add_argument("--memory_arena", choices=["sram", "dram"], default="sram", help=f"Configure memory_arena") |
| parser.add_argument("--pmu", type=int, action='append', help="PMU Event Counters") |
| parser.add_argument("--custom-input", type=pathlib.Path, help="Custom input to network") |
| parser.add_argument("--custom-output", type=pathlib.Path, help="Custom expected output data for network") |
| parser.add_argument("--ta-maxr", type=int, nargs=2, action='append', help="Max no. of pending reads") |
| parser.add_argument("--ta-maxw", type=int, nargs=2, action='append', help="Max no. of pending writes") |
| parser.add_argument("--ta-maxrw", type=int, nargs=2, action='append', help="Max no. of pending reads+writes") |
| parser.add_argument("--ta-rlatency", type=int, nargs=2, action='append', help="Minimum latency (clock cycles) from AVALID to RVALID") |
| parser.add_argument("--ta-wlatency", type=int, nargs=2, action='append', help="Minimum latency (clock cycles) from WVALID&WLAST to BVALID") |
| parser.add_argument("--ta-pulse_on", type=int, nargs=2, action='append', help="No. of cycles addresses let through (0-65535)") |
| parser.add_argument("--ta-pulse_off", type=int, nargs=2, action='append', help="No. of cycles addresses blocked (0-65535)") |
| parser.add_argument("--ta-bwcap", type=int, nargs=2, action='append', help="Max no. of 64-bit words transfered per pulse cycle 0=infinite") |
| parser.add_argument("--ta-perfctrl", type=int, nargs=2, action='append', help="selecting an event for event counter 0=default") |
| parser.add_argument("--ta-perfcnt", type=int, nargs=2, action='append', help="event counter") |
| parser.add_argument("--ta-mode", type=int, nargs=2, action='append', help="Max no. of pending reads") |
| parser.add_argument("--ta-histbin", type=int, nargs=2, action='append', help="Controlls which histogram bin (0-15) that should be accessed by HISTCNT") |
| parser.add_argument("--ta-histcnt", type=int, nargs=2, action='append', help="Read/write the selected histogram bin") |
| |
| args = parser.parse_args() |
| args.output_folder.mkdir(exist_ok=True) |
| |
| try: |
| optimize_network(args.output_folder, args.network_path, target_mapping[args.target]) |
| generate_reference_data(args.output_folder, args.network_path, args.custom_input, args.custom_output) |
| build_core_platform(args.output_folder, args.target, args.toolchain, args.memory_model, args.memory_arena, args.pmu, |
| args.ta_maxr, args.ta_maxw, args.ta_maxrw, args.ta_rlatency, args.ta_wlatency, |
| args.ta_pulse_on, args.ta_pulse_off, args.ta_bwcap, args.ta_perfctrl, args.ta_perfcnt, |
| args.ta_mode, args.ta_histbin, args.ta_histcnt) |
| run_model(args.output_folder) |
| except subprocess.CalledProcessError as err: |
| print(f"Command: '{err.cmd}' failed", file=sys.stderr) |
| return 1 |
| return 0 |
| |
| if __name__ == "__main__": |
| sys.exit(main()) |