From 7cecd48d150e58e1e5376de9fee1c05a7b9d9672 Mon Sep 17 00:00:00 2001
From: Neta Zmora <neta.zmora@intel.com>
Date: Wed, 5 Dec 2018 00:12:47 +0200
Subject: [PATCH] AMC: adjust to latest channel/filter pruning APIs
---
examples/automated_deep_compression/ADC.py | 22 ++++++++++++----------
1 file changed, 12 insertions(+), 10 deletions(-)
diff --git a/examples/automated_deep_compression/ADC.py b/examples/automated_deep_compression/ADC.py
index 8003341..f4ebe20 100755
--- a/examples/automated_deep_compression/ADC.py
+++ b/examples/automated_deep_compression/ADC.py
@@ -66,6 +66,8 @@ PERFORM_THINNING = True
#reward = -1 * math.log(total_macs)
#reward = -1 * vloss
+steps_per_episode = 13 # TODO: this should not be hard-coded
+
def do_adc(model, dataset, arch, data_loader, validate_fn, save_checkpoint_fn):
np.random.seed()
@@ -88,6 +90,7 @@ def coach_adc(model, dataset, arch, data_loader, validate_fn, save_checkpoint_fn
exploration_noise = 0.5
#exploration_noise = 0.25
exploitation_decay = 0.996
+ # These parameters are passed to the Distiller environment
graph_manager.env_params.additional_simulator_parameters = {
'model': model,
'dataset': dataset,
@@ -125,8 +128,6 @@ def coach_adc(model, dataset, arch, data_loader, validate_fn, save_checkpoint_fn
#'reward_fn': lambda top1, total_macs: min(top1/100, 0.75)
}
- #msglogger.debug('Experiment configuarion:\n' + json.dumps(graph_manager.env_params.additional_simulator_parameters, indent=2))
- steps_per_episode = 13
agent_params.exploration.noise_percentage_schedule = PieceWiseSchedule([(ConstantSchedule(exploration_noise),
EnvironmentSteps(100*steps_per_episode)),
(ExponentialSchedule(exploration_noise, 0, exploitation_decay),
@@ -172,7 +173,7 @@ class CNNEnvironment(gym.Env):
self.desired_reduction = desired_reduction
self.STATE_EMBEDDING_LEN = len(Observation._fields)
if self.onehot_encoding:
- self.STATE_EMBEDDING_LEN += 12
+ self.STATE_EMBEDDING_LEN += (steps_per_episode - 1)
self.observation_space = spaces.Box(0, float("inf"), shape=(self.STATE_EMBEDDING_LEN,))
def reset(self, init_only=False):
@@ -202,12 +203,12 @@ class CNNEnvironment(gym.Env):
return self.conv_layers[idx]
except KeyError:
return None
+
def episode_is_done(self):
return self.current_layer_id == self.num_layers()
def remaining_macs(self):
- """Return the amount of MACs remaining in the model's unprocessed
- Convolution layers.
+ """Return the amount of MACs remaining in the model's unprocessed Convolution layers.
This is normalized to the range 0..1
"""
#return 1 - self.sum_list_macs(self.unprocessed_layers) / self.dense_model_macs
@@ -228,7 +229,7 @@ class CNNEnvironment(gym.Env):
# return total_macs
def render(self, mode, close):
- """Provide some feedback to the user about what's going on
+ """Provide some feedback to the user about what's going on.
This is invoked by the Agent.
"""
if self.current_layer_id == 0:
@@ -281,7 +282,7 @@ class CNNEnvironment(gym.Env):
def step(self, action):
"""Take a step, given an action.
- The action represents the desired sparsity.
+ The action represents the desired density.
This function is invoked by the Agent.
"""
msglogger.info("env.step - current_layer_id={} action={}".format(self.current_layer_id, action))
@@ -423,16 +424,17 @@ class CNNEnvironment(gym.Env):
if prune_what == "channels":
calculate_sparsity = distiller.sparsity_ch
- reg_regims = {conv_pname: [fraction_to_prune, "Channels"]}
remove_structures = distiller.remove_channels
+ group_type = "Channels"
elif prune_what == "filters":
calculate_sparsity = distiller.sparsity_3D
- reg_regims = {conv_pname: [fraction_to_prune, "3D"]}
+ group_type = "Filters"
remove_structures = distiller.remove_filters
else:
raise ValueError("unsupported structure {}".format(prune_what))
# Create a channel-ranking pruner
- pruner = distiller.pruning.L1RankedStructureParameterPruner("adc_pruner", reg_regims)
+ pruner = distiller.pruning.L1RankedStructureParameterPruner("adc_pruner", group_type,
+ fraction_to_prune, conv_pname)
pruner.set_param_mask(conv_p, conv_pname, self.zeros_mask_dict, meta=None)
if (self.zeros_mask_dict[conv_pname].mask is None or
--
GitLab