Skip to content

Commit

Permalink
[Doc][AutoTVM] Fix bugs that override n_trials (#4842)
Browse files Browse the repository at this point in the history
  • Loading branch information
@comaniac
comaniac authored Feb 7, 2020
1 parent 0186ca4 commit e33de24
Show file tree
Hide file tree
Showing 4 changed files with 59 additions and 48 deletions.
9 changes: 5 additions & 4 deletions tutorials/autotvm/tune_relay_arm.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -299,13 +299,14 @@ def tune_tasks(tasks,
tuner_obj.load_history(autotvm.record.load_from_file(tmp_log_file))

# do tuning
n_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=n_trial,
tsk_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=tsk_trial,
early_stopping=early_stopping,
measure_option=measure_option,
callbacks=[
autotvm.callback.progress_bar(n_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)])
autotvm.callback.progress_bar(tsk_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)
])

# pick best records to a cache file
autotvm.record.pick_best(tmp_log_file, log_filename)
Expand Down
9 changes: 5 additions & 4 deletions tutorials/autotvm/tune_relay_cuda.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -201,13 +201,14 @@ def tune_tasks(tasks,
tuner_obj.load_history(autotvm.record.load_from_file(tmp_log_file))

# do tuning
n_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=n_trial,
tsk_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=tsk_trial,
early_stopping=early_stopping,
measure_option=measure_option,
callbacks=[
autotvm.callback.progress_bar(n_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)])
autotvm.callback.progress_bar(tsk_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)
])

# pick best records to a cache file
autotvm.record.pick_best(tmp_log_file, log_filename)
Expand Down
9 changes: 5 additions & 4 deletions tutorials/autotvm/tune_relay_mobile_gpu.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -283,13 +283,14 @@ def tune_tasks(tasks,
tuner_obj.load_history(autotvm.record.load_from_file(tmp_log_file))

# do tuning
n_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=n_trial,
tsk_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=tsk_trial,
early_stopping=early_stopping,
measure_option=measure_option,
callbacks=[
autotvm.callback.progress_bar(n_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)])
autotvm.callback.progress_bar(tsk_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)
])

# pick best records to a cache file
autotvm.record.pick_best(tmp_log_file, log_filename)
Expand Down
80 changes: 44 additions & 36 deletions vta/tutorials/autotvm/tune_relay_vta.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -74,6 +74,7 @@
# ---------------
# Perform vta-specific compilation with Relay from a Gluon model


def compile_network(env, target, model, start_pack, stop_pack):

# Populate the shape and data type dictionary
Expand All @@ -91,20 +92,18 @@ def compile_network(env, target, model, start_pack, stop_pack):
# Perform quantization in Relay
# Note: We set opt_level to 3 in order to fold batch norm
with relay.build_config(opt_level=3):
with relay.quantize.qconfig(global_scale=8.0,
skip_conv_layers=[0]):
with relay.quantize.qconfig(global_scale=8.0, skip_conv_layers=[0]):
mod = relay.quantize.quantize(mod, params=params)

# Perform graph packing and constant folding for VTA target
if target.device_name == "vta":
assert env.BLOCK_IN == env.BLOCK_OUT
relay_prog = graph_pack(
mod["main"],
env.BATCH,
env.BLOCK_OUT,
env.WGT_WIDTH,
start_name=start_pack,
stop_name=stop_pack)
relay_prog = graph_pack(mod["main"],
env.BATCH,
env.BLOCK_OUT,
env.WGT_WIDTH,
start_name=start_pack,
stop_name=stop_pack)

return relay_prog, params

Expand Down Expand Up @@ -195,8 +194,8 @@ def compile_network(env, target, model, start_pack, stop_pack):
# to start and end the graph packing relay pass: in other words
# where to start and finish offloading to VTA.
network = "resnet18_v1"
start_pack="nn.max_pool2d"
stop_pack="nn.global_avg_pool2d"
start_pack = "nn.max_pool2d"
stop_pack = "nn.global_avg_pool2d"

# Tuning option
log_file = "%s.%s.log" % (device, network)
Expand All @@ -209,12 +208,12 @@ def compile_network(env, target, model, start_pack, stop_pack):

'measure_option': autotvm.measure_option(
builder=autotvm.LocalBuilder(),
runner=autotvm.RPCRunner(
env.TARGET, host=tracker_host, port=tracker_port,
number=5,
timeout=60,
check_correctness=True
),
runner=autotvm.RPCRunner(env.TARGET,
host=tracker_host,
port=tracker_port,
number=5,
timeout=60,
check_correctness=True),
),
}

Expand All @@ -240,6 +239,7 @@ def compile_network(env, target, model, start_pack, stop_pack):
# Given that the tuning will be done on Pynq FPGA boards, make sure that
# the ```TARGET`` entry in the ``vta_config.json`` file is set to ``pynq``.


# You can skip the implementation of this function for this tutorial.
def tune_tasks(tasks,
measure_option,
Expand All @@ -255,7 +255,7 @@ def tune_tasks(tasks,
os.remove(tmp_log_file)

for i, tsk in enumerate(reversed(tasks)):
prefix = "[Task %2d/%2d] " % (i+1, len(tasks))
prefix = "[Task %2d/%2d] " % (i + 1, len(tasks))

# create tuner
if tuner == 'xgb' or tuner == 'xgb-rank':
Expand All @@ -276,23 +276,24 @@ def tune_tasks(tasks,
tuner_obj.load_history(autotvm.record.load_from_file(tmp_log_file))

# do tuning
n_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=n_trial,
tsk_trial = min(n_trial, len(tsk.config_space))
tuner_obj.tune(n_trial=tsk_trial,
early_stopping=early_stopping,
measure_option=measure_option,
callbacks=[
autotvm.callback.progress_bar(n_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)])
autotvm.callback.progress_bar(tsk_trial, prefix=prefix),
autotvm.callback.log_to_file(tmp_log_file)
])

# pick best records to a cache file
autotvm.record.pick_best(tmp_log_file, log_filename)
os.remove(tmp_log_file)



########################################################################
# Register VTA-specific tuning tasks


def register_vta_tuning_tasks():
from tvm.autotvm.task.topi_integration import TaskExtractEnv, deserialize_args

Expand Down Expand Up @@ -330,11 +331,15 @@ def _topi_nn_conv2d(*args, **kwargs):
########################################################################
# Finally, we launch tuning jobs and evaluate the end-to-end performance.


def tune_and_evaluate(tuning_opt):

if env.TARGET != "sim":
# Get remote from fleet node
remote = autotvm.measure.request_remote(env.TARGET, tracker_host, tracker_port, timeout=10000)
remote = autotvm.measure.request_remote(env.TARGET,
tracker_host,
tracker_port,
timeout=10000)
# Reconfigure the JIT runtime and FPGA.
vta.reconfig_runtime(remote)
vta.program_fpga(remote, bitstream=None)
Expand All @@ -351,7 +356,7 @@ def tune_and_evaluate(tuning_opt):
mod = relay.Module.from_expr(relay_prog)
tasks = autotvm.task.extract_from_program(mod,
params=params,
ops=(tvm.relay.op.nn.conv2d,),
ops=(tvm.relay.op.nn.conv2d, ),
target=target,
target_host=env.target_host)

Expand All @@ -361,17 +366,16 @@ def tune_and_evaluate(tuning_opt):
for tsk in tasks:
inp = tsk.args[0][1]
wgt = tsk.args[1][1]
batch = inp[0]*inp[4]
in_filter = inp[1]*inp[5]
out_filter = wgt[0]*wgt[4]
batch = inp[0] * inp[4]
in_filter = inp[1] * inp[5]
out_filter = wgt[0] * wgt[4]
height, width = inp[2], inp[3]
hkernel, wkernel = wgt[2], wgt[3]
hstride, wstride = tsk.args[2][0], tsk.args[2][1]
hpad, wpad = tsk.args[3][0], tsk.args[3][1]
print("({}, {}, {}, {}, {}, {}, {}, {}, {}, {}, {})".format(
batch, height, width, in_filter, out_filter,
hkernel, wkernel, hpad, wpad, hstride, wstride
))
batch, height, width, in_filter, out_filter, hkernel, wkernel,
hpad, wpad, hstride, wstride))

# We do not run the tuning in our webpage server since it takes too long.
# Comment the following line to run it by yourself.
Expand All @@ -387,14 +391,17 @@ def tune_and_evaluate(tuning_opt):
print("Compile...")
with relay.build_config(opt_level=3, disabled_pass={"AlterOpLayout"}):
if target.device_name != "vta":
graph, lib, params = relay.build(
relay_prog, target=target,
params=params, target_host=env.target_host)
graph, lib, params = relay.build(relay_prog,
target=target,
params=params,
target_host=env.target_host)
else:
with vta.build_config():
graph, lib, params = relay.build(
relay_prog, target=target,
params=params, target_host=env.target_host)
relay_prog,
target=target,
params=params,
target_host=env.target_host)

# Export library
print("Upload...")
Expand All @@ -421,6 +428,7 @@ def tune_and_evaluate(tuning_opt):
print("Mean inference time (std dev): %.2f ms (%.2f ms)" %
(np.mean(prof_res), np.std(prof_res)))


# Run the tuning and evaluate the results
tune_and_evaluate(tuning_option)

Expand Down

0 comments on commit e33de24

Please sign in to comment.