def ranked_filter_pruning(config, ratio_to_prune, is_parallel):
"""Test L1 ranking and pruning of filters.
First we rank and prune the filters of a Convolutional layer using
a L1RankedStructureParameterPruner. Then we physically remove the
filters from the model (via "thining" process).
"""
model, zeros_mask_dict = common.setup_test(config.arch, config.dataset,
is_parallel)
for pair in config.module_pairs:
# Test that we can access the weights tensor of the first convolution in layer 1
conv1_p = distiller.model_find_param(model, pair[0] + ".weight")
assert conv1_p is not None
num_filters = conv1_p.size(0)
# Test that there are no zero-filters
assert distiller.sparsity_3D(conv1_p) == 0.0
# Create a filter-ranking pruner
pruner = distiller.pruning.L1RankedStructureParameterPruner(
"filter_pruner",
group_type="Filters",
desired_sparsity=ratio_to_prune,
weights=pair[0] + ".weight")
pruner.set_param_mask(conv1_p,
pair[0] + ".weight",
zeros_mask_dict,
meta=None)
conv1 = common.find_module_by_name(model, pair[0])
assert conv1 is not None
# Test that the mask has the correct fraction of filters pruned.
# We asked for 10%, but there are only 16 filters, so we have to settle for 1/16 filters
expected_cnt_removed_filters = int(ratio_to_prune * conv1.out_channels)
expected_pruning = expected_cnt_removed_filters / conv1.out_channels
masker = zeros_mask_dict[pair[0] + ".weight"]
assert masker is not None
assert distiller.sparsity_3D(masker.mask) == expected_pruning
# Use the mask to prune
assert distiller.sparsity_3D(conv1_p) == 0
masker.apply_mask(conv1_p)
assert distiller.sparsity_3D(conv1_p) == expected_pruning
# Remove filters
conv2 = common.find_module_by_name(model, pair[1])
assert conv2 is not None
assert conv1.out_channels == num_filters
assert conv2.in_channels == num_filters
# Test thinning
distiller.remove_filters(model,
zeros_mask_dict,
config.arch,
config.dataset,
optimizer=None)
assert conv1.out_channels == num_filters - expected_cnt_removed_filters
assert conv2.in_channels == num_filters - expected_cnt_removed_filters
return model, zeros_mask_dict
def rank_and_prune_filters(self, fraction_to_prune, param, param_name, zeros_mask_dict, model, binary_map=None):
assert param.dim() == 4, "This pruning is only supported for 4D weights"
if param.grad is None:
msglogger.info("Skipping gradient pruning of %s because it does not have a gradient yet", param_name)
return
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters", 100*fraction_to_prune)
return
# Compute the multiplication of the filters times the filter_gradienrs
view_filters = param.view(param.size(0), -1)
view_filter_grads = param.grad.view(param.size(0), -1)
weighted_gradients = view_filter_grads * view_filters
weighted_gradients = weighted_gradients.sum(dim=1)
# Sort from high to low, and remove the bottom 'num_filters_to_prune' filters
filters_ordered_by_gradient = np.argsort(-weighted_gradients.detach().cpu().numpy())[:-num_filters_to_prune]
mask, binary_map = _mask_from_filter_order(filters_ordered_by_gradient, param, num_filters, binary_map)
zeros_mask_dict[param_name].mask = mask
msglogger.info("GradientRankedFilterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
return binary_map
def rank_and_prune_filters(self, fraction_to_prune, param, param_name, zeros_mask_dict, model, binary_map=None):
assert param.dim() == 4, "This pruning is only supported for 4D weights"
# Use the parameter name to locate the module that has the activation sparsity statistics
fq_name = param_name.replace(".conv", ".relu")[:-len(".weight")]
module = distiller.find_module_by_fq_name(model, fq_name)
if module is None:
raise ValueError("Could not find a layer named %s in the model."
"\nMake sure to use assign_layer_fq_names()" % fq_name)
if not hasattr(module, self.activation_rank_criterion):
raise ValueError("Could not find attribute \"{}\" in module %s"
"\nMake sure to use SummaryActivationStatsCollector(\"{}\")".
format(self.activation_rank_criterion, fq_name, self.activation_rank_criterion))
quality_criterion, std = getattr(module, self.activation_rank_criterion).value()
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters", 100*fraction_to_prune)
return
# Sort from low to high, and remove the bottom 'num_filters_to_prune' filters
filters_ordered_by_criterion = np.argsort(quality_criterion)[:-num_filters_to_prune]
mask, binary_map = _mask_from_filter_order(filters_ordered_by_criterion, param, num_filters, binary_map)
zeros_mask_dict[param_name].mask = mask
msglogger.info("ActivationL1RankedStructureParameterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
return binary_map
def rank_and_prune_filters(self,
fraction_to_prune,
param,
param_name,
zeros_mask_dict,
model,
binary_map=None):
assert param.dim() == 4, "This pruner is only supported for 4D weights"
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
keep_prob = 1 - fraction_to_prune
if binary_map is None:
binary_map = torch.bernoulli(
torch.as_tensor([keep_prob] * num_filters))
mask, _ = _mask_from_filter_order(None, param, num_filters, binary_map)
mask = mask.to(param.device)
# Compensate for dropping filters
pruning_factor = binary_map.sum() / num_filters
mask.div_(pruning_factor)
zeros_mask_dict[param_name].mask = mask
msglogger.debug(
"BernoulliFilterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
return binary_map
def rank_and_prune_filters(self,
fraction_to_prune,
param,
param_name,
zeros_mask_dict,
model,
binary_map=None):
assert param.dim(
) == 4, "This pruning is only supported for 4D weights"
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters",
100 * fraction_to_prune)
return
filters_ordered_randomly = np.random.permutation(
num_filters)[:-num_filters_to_prune]
mask, binary_map = _mask_from_filter_order(filters_ordered_randomly,
param, num_filters,
binary_map)
zeros_mask_dict[param_name].mask = mask
msglogger.info(
"RandomRankedFilterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
return binary_map
def rank_prune_filters(fraction_to_prune, param, param_name, zeros_mask_dict, model):
assert param.dim() == 4, "This thresholding is only supported for 4D weights"
# Use the parameter name to locate the module that has the activation sparsity statistics
fq_name = param_name.replace(".conv", ".relu")[:-len(".weight")]
module = distiller.find_module_by_fq_name(model, fq_name)
if module is None:
raise ValueError("Could not find a layer named %s in the model."
"\nMake sure to use assign_layer_fq_names()" % fq_name)
if not hasattr(module, 'apoz_channels'):
raise ValueError("Could not find attribute \'apoz_channels\' in module %s."
"\nMake sure to use SummaryActivationStatsCollector(\"apoz_channels\")" % fq_name)
apoz, std = module.apoz_channels.value()
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters", 100*fraction_to_prune)
return
# Sort from high to low, and remove the bottom 'num_filters_to_prune' filters
filters_ordered_by_apoz = np.argsort(-apoz)[:-num_filters_to_prune]
zeros_mask_dict[param_name].mask = RankedFiltersParameterPruner.mask_from_filter_order(filters_ordered_by_apoz,
param, num_filters)
msglogger.info("ActivationL1RankedStructureParameterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
def test_sparsity():
zeros = torch.zeros(2, 3, 5, 6)
print(distiller.sparsity(zeros))
assert distiller.sparsity(zeros) == 1.0
assert distiller.sparsity_3D(zeros) == 1.0
assert distiller.density_3D(zeros) == 0.0
ones = torch.zeros(12, 43, 4, 6)
ones.fill_(1)
assert distiller.sparsity(ones) == 0.0
def test_ranked_filter_pruning():
model, zeros_mask_dict = setup_test("resnet20_cifar", "cifar10")
# Test that we can access the weights tensor of the first convolution in layer 1
conv1_p = distiller.model_find_param(model, "layer1.0.conv1.weight")
assert conv1_p is not None
# Test that there are no zero-channels
assert distiller.sparsity_3D(conv1_p) == 0.0
# Create a filter-ranking pruner
reg_regims = {"layer1.0.conv1.weight": [0.1, "3D"]}
pruner = distiller.pruning.L1RankedStructureParameterPruner(
"filter_pruner", reg_regims)
pruner.set_param_mask(conv1_p,
"layer1.0.conv1.weight",
zeros_mask_dict,
meta=None)
conv1 = find_module_by_name(model, "layer1.0.conv1")
assert conv1 is not None
# Test that the mask has the correct fraction of filters pruned.
# We asked for 10%, but there are only 16 filters, so we have to settle for 1/16 filters
expected_pruning = int(0.1 * conv1.out_channels) / conv1.out_channels
assert distiller.sparsity_3D(
zeros_mask_dict["layer1.0.conv1.weight"].mask) == expected_pruning
# Use the mask to prune
assert distiller.sparsity_3D(conv1_p) == 0
zeros_mask_dict["layer1.0.conv1.weight"].apply_mask(conv1_p)
assert distiller.sparsity_3D(conv1_p) == expected_pruning
# Remove filters
conv2 = find_module_by_name(model, "layer1.0.conv2")
assert conv2 is not None
assert conv1.out_channels == 16
assert conv2.in_channels == 16
# Test thinning
distiller.remove_filters(model, zeros_mask_dict, "resnet20_cifar",
"cifar10")
assert conv1.out_channels == 15
assert conv2.in_channels == 15
def rank_and_prune_filters(self,
fraction_to_prune,
param,
param_name,
zeros_mask_dict,
model,
binary_map=None):
assert param.dim(
) == 4, "This pruning is only supported for 4D weights"
# Use the parameter name to locate the module that has the activation sparsity statistics
fq_name = param_name.replace(".conv", ".relu")[:-len(".weight")]
distiller.assign_layer_fq_names(model)
module = distiller.find_module_by_fq_name(model, fq_name)
assert module is not None
if not hasattr(module, self.activation_rank_criterion):
raise ValueError(
"Could not find attribute \"%s\" in module %s\n"
"\tThis is pruner uses activation statistics collected during forward-"
"passes of the network.\n"
"\tThis error is an indication that these statistics "
"have not been collected yet.\n"
"\tMake sure to use SummaryActivationStatsCollector(\"%s\")\n"
"\tFor more info see issue #444 (https://github.com/NervanaSystems/distiller/issues/444)"
% (self.activation_rank_criterion, fq_name,
self.activation_rank_criterion))
quality_criterion, std = getattr(
module, self.activation_rank_criterion).value()
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters",
100 * fraction_to_prune)
return
# Sort from low to high, and remove the bottom 'num_filters_to_prune' filters
filters_ordered_by_criterion = np.argsort(
quality_criterion)[:-num_filters_to_prune]
mask, binary_map = _mask_from_filter_order(
filters_ordered_by_criterion, param, num_filters, binary_map)
zeros_mask_dict[param_name].mask = mask
msglogger.info(
"ActivationL1RankedStructureParameterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
return binary_map
def prune_tensor(param, param_name, fraction_to_prune, zeros_mask_dict):
"""Prune filters from a parameter tensor.
Returns the filter-sparsity of the tensor.
"""
# Create a filter-ranking pruner
pruner = distiller.pruning.L1RankedStructureParameterPruner(
name=None,
group_type="Filters",
desired_sparsity=fraction_to_prune,
weights=param_name)
pruner.set_param_mask(param, param_name, zeros_mask_dict, meta=None)
# Use the mask to prune
zeros_mask_dict[param_name].apply_mask(param)
return distiller.sparsity_3D(param)
def weights_sparsity_summary(model,
return_total_sparsity=False,
param_dims=[2, 4]):
df = pd.DataFrame(columns=[
'Name', 'Shape', 'NNZ (dense)', 'NNZ (sparse)', 'Cols (%)', 'Rows (%)',
'Ch (%)', '2D (%)', '3D (%)', 'Fine (%)', 'Std', 'Mean', 'Abs-Mean'
])
pd.set_option('precision', 2)
params_size = 0
sparse_params_size = 0
summary_param_types = ['weight', 'bias']
for name, param in model.state_dict().items():
# Extract just the actual parameter's name, which in this context we treat as its "type"
curr_param_type = name.split('.')[-1]
if param.dim(
) in param_dims and curr_param_type in summary_param_types:
_density = distiller.density(param)
params_size += torch.numel(param)
sparse_params_size += param.numel() * _density
df.loc[len(df.index)] = ([
name,
distiller.size_to_str(param.size()),
torch.numel(param),
int(_density * param.numel()),
distiller.sparsity_cols(param) * 100,
distiller.sparsity_rows(param) * 100,
distiller.sparsity_ch(param) * 100,
distiller.sparsity_2D(param) * 100,
distiller.sparsity_3D(param) * 100, (1 - _density) * 100,
param.std().item(),
param.mean().item(),
param.abs().mean().item()
])
total_sparsity = (1 - sparse_params_size / params_size) * 100
df.loc[len(df.index)] = ([
'Total sparsity:', '-', params_size,
int(sparse_params_size), 0, 0, 0, 0, 0, total_sparsity, 0, 0, 0
])
if return_total_sparsity:
return df, total_sparsity
return df
def test_sparsity():
zeros = torch.zeros(2, 3, 5, 6)
print(distiller.sparsity(zeros))
assert distiller.sparsity(zeros) == 1.0
assert distiller.sparsity_3D(zeros) == 1.0
assert distiller.density_3D(zeros) == 0.0
ones = torch.ones(12, 43, 4, 6)
assert distiller.sparsity(ones) == 0.0
x = torch.tensor([[1., 2., 0, 4., 0], [1., 2., 0, 4., 0]])
assert distiller.density(x) == 0.6
assert distiller.density_cols(x, transposed=False) == 0.6
assert distiller.sparsity_rows(x, transposed=False) == 0
x = torch.tensor([[0., 0., 0], [1., 4., 0], [1., 2., 0], [0., 0., 0]])
assert distiller.density(x) == 4 / 12
assert distiller.sparsity_rows(x, transposed=False) == 0.5
assert common.almost_equal(distiller.sparsity_cols(x, transposed=False),
1 / 3)
assert common.almost_equal(distiller.sparsity_rows(x), 1 / 3)
def weights_sparsity_summary(model, return_total_sparsity=False, param_dims=[2,4]):
df = pd.DataFrame(columns=['Name', 'Shape', 'NNZ (dense)', 'NNZ (sparse)',
'Cols (%)','Rows (%)', 'Ch (%)', '2D (%)', '3D (%)',
'Fine (%)', 'Std', 'Mean', 'Abs-Mean'])
pd.set_option('precision', 2)
params_size = 0
sparse_params_size = 0
for name, param in model.state_dict().items():
if (param.dim() in param_dims) and any(type in name for type in ['weight', 'bias']):
_density = distiller.density(param)
params_size += torch.numel(param)
sparse_params_size += param.numel() * _density
df.loc[len(df.index)] = ([
name,
distiller.size_to_str(param.size()),
torch.numel(param),
int(_density * param.numel()),
distiller.sparsity_cols(param)*100,
distiller.sparsity_rows(param)*100,
distiller.sparsity_ch(param)*100,
distiller.sparsity_2D(param)*100,
distiller.sparsity_3D(param)*100,
(1-_density)*100,
param.std().item(),
param.mean().item(),
param.abs().mean().item()
])
total_sparsity = (1 - sparse_params_size/params_size)*100
df.loc[len(df.index)] = ([
'Total sparsity:',
'-',
params_size,
int(sparse_params_size),
0, 0, 0, 0, 0,
total_sparsity,
0, 0, 0])
if return_total_sparsity:
return df, total_sparsity
return df
def rank_prune_filters(fraction_to_prune, param, param_name, zeros_mask_dict, model):
assert param.dim() == 4, "This thresholding is only supported for 4D weights"
num_filters = param.size(0)
num_filters_to_prune = int(fraction_to_prune * num_filters)
if num_filters_to_prune == 0:
msglogger.info("Too few filters - can't prune %.1f%% filters", 100*fraction_to_prune)
return
# Compute the multiplicatipn of the filters times the filter_gradienrs
view_filters = param.view(param.size(0), -1)
view_filter_grads = param.grad.view(param.size(0), -1)
weighted_gradients = view_filter_grads * view_filters
weighted_gradients = weighted_gradients.sum(dim=1)
# Sort from high to low, and remove the bottom 'num_filters_to_prune' filters
filters_ordered_by_gradient = np.argsort(-weighted_gradients.detach().cpu().numpy())[:-num_filters_to_prune]
zeros_mask_dict[param_name].mask = RankedFiltersParameterPruner.mask_from_filter_order(filters_ordered_by_gradient,
param, num_filters)
msglogger.info("GradientRankedFilterPruner - param: %s pruned=%.3f goal=%.3f (%d/%d)",
param_name,
distiller.sparsity_3D(zeros_mask_dict[param_name].mask),
fraction_to_prune, num_filters_to_prune, num_filters)
def weights_sparsity_summary(model,
return_total_sparsity=False,
param_dims=[2, 4]):
df = pd.DataFrame(columns=[
"Name",
"Shape",
"NNZ (dense)",
"NNZ (sparse)",
"Cols (%)",
"Rows (%)",
"Ch (%)",
"2D (%)",
"3D (%)",
"Fine (%)",
"Std",
"Mean",
"Abs-Mean",
])
pd.set_option("precision", 2)
params_size = 0
sparse_params_size = 0
for name, param in model.state_dict().items():
# Extract just the actual parameter's name, which in this context we treat as its "type"
if param.dim() in param_dims and any(type in name
for type in ["weight", "bias"]):
_density = distiller.density(param)
params_size += torch.numel(param)
sparse_params_size += param.numel() * _density
df.loc[len(df.index)] = [
name,
distiller.size_to_str(param.size()),
torch.numel(param),
int(_density * param.numel()),
distiller.sparsity_cols(param) * 100,
distiller.sparsity_rows(param) * 100,
distiller.sparsity_ch(param) * 100,
distiller.sparsity_2D(param) * 100,
distiller.sparsity_3D(param) * 100,
(1 - _density) * 100,
param.std().item(),
param.mean().item(),
param.abs().mean().item(),
]
total_sparsity = (1 - sparse_params_size / params_size) * 100
df.loc[len(df.index)] = [
"Total sparsity:",
"-",
params_size,
int(sparse_params_size),
0,
0,
0,
0,
0,
total_sparsity,
0,
0,
0,
]
if return_total_sparsity:
return df, total_sparsity
return df
def ranked_filter_pruning(config,
ratio_to_prune,
is_parallel,
rounding_fn=math.floor):
"""Test L1 ranking and pruning of filters.
First we rank and prune the filters of a Convolutional layer using
a L1RankedStructureParameterPruner. Then we physically remove the
filters from the model (via "thining" process).
"""
logger.info("executing: %s (invoked by %s)" %
(inspect.currentframe().f_code.co_name,
inspect.currentframe().f_back.f_code.co_name))
model, zeros_mask_dict = common.setup_test(config.arch, config.dataset,
is_parallel)
for pair in config.module_pairs:
# Test that we can access the weights tensor of the first convolution in layer 1
conv1_p = distiller.model_find_param(model, pair[0] + ".weight")
assert conv1_p is not None
num_filters = conv1_p.size(0)
# Test that there are no zero-filters
assert distiller.sparsity_3D(conv1_p) == 0.0
# Create a filter-ranking pruner
pruner = distiller.pruning.L1RankedStructureParameterPruner(
"filter_pruner",
group_type="Filters",
desired_sparsity=ratio_to_prune,
weights=pair[0] + ".weight",
rounding_fn=rounding_fn)
pruner.set_param_mask(conv1_p,
pair[0] + ".weight",
zeros_mask_dict,
meta=None)
conv1 = common.find_module_by_name(model, pair[0])
assert conv1 is not None
# Test that the mask has the correct fraction of filters pruned.
# We asked for 10%, but there are only 16 filters, so we have to settle for 1/16 filters
expected_cnt_removed_filters = int(ratio_to_prune * conv1.out_channels)
expected_pruning = expected_cnt_removed_filters / conv1.out_channels
masker = zeros_mask_dict[pair[0] + ".weight"]
assert masker is not None
assert distiller.sparsity_3D(masker.mask) == expected_pruning
# Use the mask to prune
assert distiller.sparsity_3D(conv1_p) == 0
masker.apply_mask(conv1_p)
assert distiller.sparsity_3D(conv1_p) == expected_pruning
# Remove filters
conv2 = common.find_module_by_name(model, pair[1])
assert conv2 is not None
assert conv1.out_channels == num_filters
assert conv2.in_channels == num_filters
# Test thinning
distiller.remove_filters(model,
zeros_mask_dict,
config.arch,
config.dataset,
optimizer=None)
assert conv1.out_channels == num_filters - expected_cnt_removed_filters
assert conv2.in_channels == num_filters - expected_cnt_removed_filters
# Test the thinned model
dummy_input = distiller.get_dummy_input(config.dataset,
distiller.model_device(model))
optimizer = torch.optim.SGD(model.parameters(),
lr=0.01,
momentum=0.9,
weight_decay=0.1)
run_forward_backward(model, optimizer, dummy_input)
return model, zeros_mask_dict
