Skip to content

Commit

Permalink
formatting
Browse files Browse the repository at this point in the history
  • Loading branch information
@byronxu99
byronxu99 committed Dec 1, 2023
1 parent bebf7c4 commit 33f8423
Show file tree
Hide file tree
Showing 5 changed files with 71 additions and 70 deletions.
5 changes: 1 addition & 4 deletions vowpalwabbit/core/include/vw/core/array_parameters_dense.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -102,10 +102,7 @@ class dense_parameters
dense_parameters(dense_parameters&&) noexcept;

bool not_null();
VW::weight* first()
{
return _begin.get();
} // TODO: Temporary fix for allreduce.
VW::weight* first() { return _begin.get(); } // TODO: Temporary fix for allreduce.

VW::weight* data() { return _begin.get(); }

Expand Down
2 changes: 1 addition & 1 deletion vowpalwabbit/core/include/vw/core/reductions/ftrl.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -47,5 +47,5 @@ size_t write_model_field(io_buf&, VW::reductions::ftrl&, const std::string&, boo
} // namespace model_utils

std::shared_ptr<VW::LEARNER::learner> ftrl_setup(VW::setup_base_i& stack_builder);
}
} // namespace reductions
} // namespace VW
12 changes: 6 additions & 6 deletions vowpalwabbit/core/include/vw/core/reductions/search/search.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,12 +12,12 @@
// (going to clog [which in turn goes to err, with some differences])
// We may want to create/use some macro-based loggers (which will wrap the spdlog ones)
// to mimic this behavior.
#define cdbg std::clog
#undef cdbg
#define cdbg \
if (1) {} \
else \
std::clog
# define cdbg std::clog
# undef cdbg
# define cdbg \
if (1) {} \
else \
std::clog
// comment the previous two lines if you want loads of debug output :)

using action = uint32_t;
Expand Down
111 changes: 57 additions & 54 deletions vowpalwabbit/core/src/reductions/search/search.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,7 @@ class action_repr
action a = 0;
std::shared_ptr<VW::features> repr = nullptr;
action_repr() = default;
//action_repr(action _a, std::shared_ptr<VW::features> _repr) : a(_a), repr(std::move(_repr)) {}
// action_repr(action _a, std::shared_ptr<VW::features> _repr) : a(_a), repr(std::move(_repr)) {}
action_repr(action _a, VW::features* _repr) : a(_a)
{
// Copy construct the features object
Expand Down Expand Up @@ -191,7 +191,7 @@ class search_private
auto_condition_settings acset; // settings for auto-conditioning
size_t history_length = 0; // value of --search_history_length, used by some tasks, default 1

size_t A = 0; // NOLINT total number of actions, [1..A]; 0 means ldf
size_t A = 0; // NOLINT total number of actions, [1..A]; 0 means ldf
size_t feature_width = 0; // total number of learners;
bool cb_learner = false; // do contextual bandit learning on action (was "! rollout_all_actions" which was confusing)
search_state state; // current state of learning
Expand Down Expand Up @@ -315,22 +315,13 @@ class search_private

~search_private()
{
if (all)
{
clear_memo_foreach_action(*this);
}
if (all) { clear_memo_foreach_action(*this); }
}
};

void clear_memo_foreach_action(search_private& priv)
{
priv.memo_foreach_action.clear();
}
void clear_memo_foreach_action(search_private& priv) { priv.memo_foreach_action.clear(); }

search::search()
{
priv = std::make_shared<search_private>();
}
search::search() { priv = std::make_shared<search_private>(); }

std::string audit_feature_space("conditional");
uint64_t conditional_constant = 8290743;
Expand Down Expand Up @@ -817,7 +808,8 @@ void add_example_conditioning(search_private& priv, VW::example& ec, size_t cond
{
auto old_ft_index_offset = ec.ft_offset;
ec.ft_offset = 0;
auto restore_ft_index_offset = VW::scope_exit([&ec, old_ft_index_offset] { ec.ft_offset = old_ft_index_offset; });
auto restore_ft_index_offset =
VW::scope_exit([&ec, old_ft_index_offset] { ec.ft_offset = old_ft_index_offset; });
VW::foreach_feature<search_private, uint64_t, add_new_feature>(*priv.all, ec, priv);
}
}
Expand All @@ -835,7 +827,8 @@ void add_example_conditioning(search_private& priv, VW::example& ec, size_t cond
{
if ((fs.values[k] > 1e-10) || (fs.values[k] < -1e-10))
{
uint64_t fid = 84913 + 48371803 * (extra_offset + 8392817 * name) + 840137 * (4891 + fs.indices[k] / multiplier);
uint64_t fid =
84913 + 48371803 * (extra_offset + 8392817 * name) + 840137 * (4891 + fs.indices[k] / multiplier);
if (priv.all->output_config.audit)
{
priv.dat_new_feature_audit_ss.str("");
Expand Down Expand Up @@ -1330,11 +1323,13 @@ action single_prediction_ldf(search_private& priv, VW::example* ecs, size_t ec_c

VW::polylabel old_label = ecs[a].l;
uint64_t old_offset = ecs[a].ft_offset;
auto restore_example = VW::scope_exit([&ecs, a, start_K, &old_label, old_offset] {
ecs[a].l = old_label;
ecs[a].ft_offset = old_offset;
if (start_K > 0) { VW::details::truncate_example_namespaces_from_example(ecs[a], ecs[0]); }
});
auto restore_example = VW::scope_exit(
[&ecs, a, start_K, &old_label, old_offset]
{
ecs[a].l = old_label;
ecs[a].ft_offset = old_offset;
if (start_K > 0) { VW::details::truncate_example_namespaces_from_example(ecs[a], ecs[0]); }
});

if (start_K > 0) { VW::details::append_example_namespaces_from_example(ecs[a], ecs[0]); }
ecs[a].l.cs = priv.ldf_test_label;
Expand Down Expand Up @@ -1520,10 +1515,12 @@ void generate_training_example(search_private& priv, VW::polylabel& losses, floa

VW::example& ec = priv.learn_ec_ref[0];
VW::polylabel old_label = ec.l;
auto restore_example = VW::scope_exit([&priv, &ec, &old_label, add_conditioning] {
if (add_conditioning) { del_example_conditioning(priv, ec); }
ec.l = old_label;
});
auto restore_example = VW::scope_exit(
[&priv, &ec, &old_label, add_conditioning]
{
if (add_conditioning) { del_example_conditioning(priv, ec); }
ec.l = old_label;
});

ec.l = losses; // labels;
if (add_conditioning)
Expand All @@ -1546,17 +1543,19 @@ void generate_training_example(search_private& priv, VW::polylabel& losses, floa
{
assert(cs_get_costs_size(priv.cb_learner, losses) == priv.learn_ec_ref_cnt);
size_t start_K = (priv.is_ldf && VW::is_cs_example_header(priv.learn_ec_ref[0])) ? 1 : 0; // NOLINT

auto restore_example = VW::scope_exit([&priv, start_K, add_conditioning] {
if (add_conditioning)
{
for (action a = static_cast<uint32_t>(start_K); a < priv.learn_ec_ref_cnt; a++)

auto restore_example = VW::scope_exit(
[&priv, start_K, add_conditioning]
{
VW::example& ec = priv.learn_ec_ref[a];
del_example_conditioning(priv, ec);
}
}
});
if (add_conditioning)
{
for (action a = static_cast<uint32_t>(start_K); a < priv.learn_ec_ref_cnt; a++)
{
VW::example& ec = priv.learn_ec_ref[a];
del_example_conditioning(priv, ec);
}
}
});

// TODO: weight
if (add_conditioning)
Expand Down Expand Up @@ -1874,12 +1873,14 @@ action search_predict(search_private& priv, VW::example* ecs, size_t ec_cnt, pta
size_t start_K = (priv.is_ldf && VW::is_cs_example_header(ecs[0])) ? 1 : 0; // NOLINT
priv.last_action_repr.clear();

auto restore_example = VW::scope_exit([&priv, start_K, ec_cnt, ecs] {
if (priv.auto_condition_features)
{
for (size_t n = start_K; n < ec_cnt; n++) { del_example_conditioning(priv, ecs[n]); }
}
});
auto restore_example = VW::scope_exit(
[&priv, start_K, ec_cnt, ecs]
{
if (priv.auto_condition_features)
{
for (size_t n = start_K; n < ec_cnt; n++) { del_example_conditioning(priv, ecs[n]); }
}
});
if (priv.auto_condition_features)
{
for (size_t n = start_K; n < ec_cnt; n++)
Expand Down Expand Up @@ -1986,10 +1987,10 @@ void hoopla_permute(size_t* B, size_t* end)
std::sort(B, end, cmp_size_t);
// make some temporary space
std::vector<size_t> A((N + 1) * 2, 0); // NOLINT
A[N] = B[0]; // arbitrarily choose the maximum in the middle
A[N + 1] = B[N - 1]; // so the maximum goes next to it
size_t lo = N, hi = N + 1; // which parts of A have we filled in? [lo,hi]
size_t i = 0, j = N - 1; // which parts of B have we already covered? [0,i] and [j,N-1]
A[N] = B[0]; // arbitrarily choose the maximum in the middle
A[N + 1] = B[N - 1]; // so the maximum goes next to it
size_t lo = N, hi = N + 1; // which parts of A have we filled in? [lo,hi]
size_t i = 0, j = N - 1; // which parts of B have we already covered? [0,i] and [j,N-1]
while (i + 1 < j)
{
// there are four options depending on where things get placed
Expand Down Expand Up @@ -2085,15 +2086,17 @@ void BaseTask::Run()
float old_test_loss = priv.test_loss;
// float old_learn_loss = priv.learn_loss;
float old_train_loss = priv.train_loss;
auto restore_priv = VW::scope_exit([this, &priv, old_should_produce_string, old_test_loss, old_train_loss] {
priv.should_produce_string = old_should_produce_string;
if (!this->_final_run)
{
priv.test_loss = old_test_loss;
// priv.learn_loss = old_learn_loss;
priv.train_loss = old_train_loss;
}
});
auto restore_priv = VW::scope_exit(
[this, &priv, old_should_produce_string, old_test_loss, old_train_loss]
{
priv.should_produce_string = old_should_produce_string;
if (!this->_final_run)
{
priv.test_loss = old_test_loss;
// priv.learn_loss = old_learn_loss;
priv.train_loss = old_train_loss;
}
});

if (!_final_run && !_with_output_string) { priv.should_produce_string = false; }
priv.learn_loss *= 0.5;
Expand Down
11 changes: 6 additions & 5 deletions vowpalwabbit/core/src/reductions/search/search_meta.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -75,14 +75,14 @@ class task_data
public:
size_t max_branches, kbest;
std::vector<branch> branches;
std::vector<std::pair<branch, std::unique_ptr<std::string>> > final;
std::vector<std::pair<branch, std::unique_ptr<std::string>>> final;
path trajectory;
float total_cost;
size_t cur_branch;
std::unique_ptr<std::string> output_string = nullptr;
std::unique_ptr<std::stringstream> kbest_out = nullptr;

task_data(size_t mb, size_t kb) : max_branches(mb), kbest(kb) { }
task_data(size_t mb, size_t kb) : max_branches(mb), kbest(kb) {}
};

void initialize(Search::search& sch, size_t& /*num_actions*/, options_i& options)
Expand All @@ -99,7 +99,8 @@ void initialize(Search::search& sch, size_t& /*num_actions*/, options_i& options
.help("Number of best items to output (0=just like non-selectional-branching, default)"));
options.add_and_parse(new_options);

auto d = std::make_shared<task_data>(VW::cast_to_smaller_type<size_t>(max_branches), VW::cast_to_smaller_type<size_t>(kbest));
auto d = std::make_shared<task_data>(
VW::cast_to_smaller_type<size_t>(max_branches), VW::cast_to_smaller_type<size_t>(kbest));
sch.set_metatask_data(std::move(d));
}

Expand Down Expand Up @@ -204,8 +205,8 @@ void run(Search::search& sch, VW::multi_ex& ec)

// sort the finals by cost
stable_sort(d.final.begin(), d.final.end(),
[](const std::pair<branch, std::unique_ptr<std::string>>& a, const std::pair<branch, std::unique_ptr<std::string>>& b) -> bool
{ return a.first.first < b.first.first; });
[](const std::pair<branch, std::unique_ptr<std::string>>& a,
const std::pair<branch, std::unique_ptr<std::string>>& b) -> bool { return a.first.first < b.first.first; });

d.kbest_out.reset();
if (d.output_string && (d.kbest > 0))
Expand Down

0 comments on commit 33f8423

Please sign in to comment.