Replace ThneedModel with TinygradModel (#33532)

* squash

* bump tg

* bump tg

* debump tinygrad

* bump tinygrad

* bump tg

* Skip init iteration

* fixes

* cleanups

* skip first test sample

* typos

* linter unhappy

* update cpu usage

* OPENCL just zeros for now

* imports

* Try printing

* Runs again, but slower

* unused import

* Allow more buffer with tg and all on gpu

* bump tinygrad

---------

Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com>
Co-authored-by: Bruce Wayne <harald.the.engineer@gmail.com>
This commit is contained in:
Mitchell Goff
2024-11-11 13:32:21 -08:00
committed by GitHub
parent 127c922aed
commit da952e9b64
25 changed files with 148 additions and 1018 deletions
+13 -24
View File
@@ -13,15 +13,6 @@ common_src = [
"transforms/transform.cc",
]
thneed_src_common = [
"thneed/thneed_common.cc",
"thneed/serialize.cc",
]
thneed_src_qcom = thneed_src_common + ["thneed/thneed_qcom2.cc"]
thneed_src_pc = thneed_src_common + ["thneed/thneed_pc.cc"]
thneed_src = thneed_src_qcom if arch == "larch64" else thneed_src_pc
# SNPE except on Mac and ARM Linux
snpe_lib = []
if arch != "Darwin" and arch != "aarch64":
@@ -59,20 +50,18 @@ fn = File("models/supercombo").abspath
cmd = f'python3 {Dir("#selfdrive/modeld").abspath}/get_model_metadata.py {fn}.onnx'
lenv.Command(fn + "_metadata.pkl", [fn + ".onnx"] + tinygrad_files, cmd)
# Build thneed model
if arch == "larch64" or GetOption('pc_thneed'):
tinygrad_opts = []
if not GetOption('pc_thneed'):
# use FLOAT16 on device for speed + don't cache the CL kernels for space
tinygrad_opts += ["FLOAT16=1", "PYOPENCL_NO_CACHE=1"]
cmd = f"cd {Dir('#').abspath}/tinygrad_repo && " + ' '.join(tinygrad_opts) + f" python3 openpilot/compile2.py {fn}.onnx {fn}.thneed"
# Compile tinygrad model
# TODO this is all super hacky
pythonpath_string = 'PYTHONPATH="${PYTHONPATH}:' + env.Dir("#tinygrad_repo").abspath + '"'
if arch == 'larch64':
device_string = 'QCOM=1'
elif arch == 'Darwin' or arch == 'aarch64':
device_string = 'CLANG=1 IMAGE=0'
else:
device_string = 'GPU=1'
lenv.Command(fn + ".thneed", [fn + ".onnx"] + tinygrad_files, cmd)
for model_name in ['supercombo', 'dmonitoring_model']:
fn = File(f"models/{model_name}").abspath
cmd = f'{pythonpath_string} {device_string} python3 {Dir("#tinygrad_repo").abspath}/examples/openpilot/compile3.py {fn}.onnx {fn}_tinygrad.pkl'
lenv.Command(fn + "_tinygrad.pkl", [fn + ".onnx"] + tinygrad_files, cmd)
fn_dm = File("models/dmonitoring_model").abspath
cmd = f"cd {Dir('#').abspath}/tinygrad_repo && " + ' '.join(tinygrad_opts) + f" python3 openpilot/compile2.py {fn_dm}.onnx {fn_dm}.thneed"
lenv.Command(fn_dm + ".thneed", [fn_dm + ".onnx"] + tinygrad_files, cmd)
thneed_lib = env.SharedLibrary('thneed', thneed_src, LIBS=[gpucommon, common, 'OpenCL', 'dl'])
thneedmodel_lib = env.Library('thneedmodel', ['runners/thneedmodel.cc'])
lenvCython.Program('runners/thneedmodel_pyx.so', 'runners/thneedmodel_pyx.pyx', LIBS=envCython["LIBS"]+[thneedmodel_lib, thneed_lib, gpucommon, common, 'dl', 'OpenCL'])