Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Embedding] Refine the code of exporting EV. #1002

Open
wants to merge 2 commits into
base: deeprec2206
Choose a base branch
from
Open

[Embedding] Refine the code of exporting EV. #1002

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c02681c
[Embedding] Refine the code of exporting EV.
JackMoriarty Jul 4, 2024
bcebba8
[opt] optimize embedding save.
candyzone Sep 5, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions tensorflow/core/framework/embedding/lockless_hash_map.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,6 +83,8 @@ class LocklessHashMap : public KVInterface<K, V> {
std::pair<std::pair<const K, ValuePtr<V>*>*, long unsigned int> it = hash_map_.GetSnapshot();
hash_map_dump = it.first;
bucket_count = it.second;
key_list->reserve(bucket_count);
value_ptr_list->reserve(bucket_count);
for (int64 j = 0; j < bucket_count; j++) {
if (hash_map_dump[j].first != LocklessHashMap<K, V>::EMPTY_KEY_
&& hash_map_dump[j].first != LocklessHashMap<K, V>::DELETED_KEY_) {
Expand Down
97 changes: 67 additions & 30 deletions tensorflow/core/kernels/kv_variable_ops.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,22 @@ namespace {
const int kSavedPartitionNum = 1000;
}

template<class T>
std::vector<T> concatenateVectors(const std::vector<std::vector<T> >& vectors) {
size_t totalSize = 0;
for (const auto& vec : vectors) {
totalSize += vec.size();
}

std::vector<T> result(totalSize);
auto it = result.begin();
for (const auto& vec : vectors) {
it = std::copy(vec.begin(), vec.end(), it);
}

return result;
}

template<class T>
class EVKeyDumpIterator: public DumpIterator<T> {
public:
Expand Down Expand Up @@ -226,44 +242,47 @@ Status DumpEmbeddingValues(EmbeddingVar<K, V>* ev,
// so that we can dynamically load ev with changed partition number
int64 filter_freq = ev->MinFreq();
for (size_t i = 0; i < tot_key_list.size(); i++) {
for (int partid = 0; partid < kSavedPartitionNum; partid++) {
if (tot_key_list[i] % kSavedPartitionNum == partid) {
if (tot_valueptr_list[i] == reinterpret_cast<V*>(-1)) {
// only forward, no backward, bypass
} else if (tot_valueptr_list[i] == nullptr) {
key_filter_list_parts[partid].push_back(tot_key_list[i]);
} else {
key_list_parts[partid].push_back(tot_key_list[i]);
valueptr_list_parts[partid].push_back(tot_valueptr_list[i]);
}
break;
}
if (tot_key_list[i] < 0) {
LOG(WARNING) << "find negative key [" << tot_key_list[i] << "] in EV ("
<< tensor_key << ")";
continue;
}
int partid = tot_key_list[i] % kSavedPartitionNum;
if (tot_valueptr_list[i] == reinterpret_cast<V*>(-1)) {
// only forward, no backward, bypass
} else if (tot_valueptr_list[i] == nullptr) {
key_filter_list_parts[partid].push_back(tot_key_list[i]);
} else {
key_list_parts[partid].push_back(tot_key_list[i]);
valueptr_list_parts[partid].push_back(tot_valueptr_list[i]);
}
}

for (size_t i = 0; i < tot_version_list.size(); i++) {
for (int partid = 0; partid < kSavedPartitionNum; partid++) {
if (tot_key_list[i] % kSavedPartitionNum == partid) {
if (tot_valueptr_list[i] == nullptr) {
version_filter_list_parts[partid].push_back(tot_version_list[i]);
} else {
version_list_parts[partid].push_back(tot_version_list[i]);
}
break;
}
if (tot_key_list[i] < 0) {
LOG(WARNING) << "find negative key [" << tot_key_list[i] << "] in EV("
<< tensor_key << ")";
continue;
}
int partid = tot_key_list[i] % kSavedPartitionNum;
if (tot_valueptr_list[i] == nullptr) {
version_filter_list_parts[partid].push_back(tot_version_list[i]);
} else {
version_list_parts[partid].push_back(tot_version_list[i]);
}
}

for (size_t i = 0; i < tot_freq_list.size(); i++) {
for (int partid = 0; partid < kSavedPartitionNum; partid++) {
if (tot_key_list[i] % kSavedPartitionNum == partid) {
if (tot_valueptr_list[i] == nullptr) {
freq_filter_list_parts[partid].push_back(tot_freq_list[i]);
} else {
freq_list_parts[partid].push_back(tot_freq_list[i]);
}
break;
}
if (tot_key_list[i] < 0) {
LOG(WARNING) << "find negative key [" << tot_key_list[i] << "] in EV("
<< tensor_key << ")";
continue;
}
int partid = tot_key_list[i] % kSavedPartitionNum;
if (tot_valueptr_list[i] == nullptr) {
freq_filter_list_parts[partid].push_back(tot_freq_list[i]);
} else {
freq_list_parts[partid].push_back(tot_freq_list[i]);
}
}
// LOG(INFO) << "EV:" << tensor_key << ", key_list_parts:" << key_list_parts.size();
Expand All @@ -272,6 +291,23 @@ Status DumpEmbeddingValues(EmbeddingVar<K, V>* ev,
part_offset_flat(0) = 0;
part_filter_offset[0] = 0;
int ptsize = 0;

for (int partid = 0; partid < kSavedPartitionNum; partid++) {
std::vector<K>& key_list = key_list_parts[partid];
std::vector<K>& key_filter_list = key_filter_list_parts[partid];
ptsize += key_list.size();
part_offset_flat(partid + 1) = part_offset_flat(partid) + key_list.size();
part_filter_offset[partid + 1] = part_filter_offset[partid] + key_filter_list.size();
}
partitioned_tot_key_list=concatenateVectors(key_list_parts);
partitioned_tot_valueptr_list=concatenateVectors(valueptr_list_parts);
partitioned_tot_version_list=concatenateVectors(version_list_parts);
partitioned_tot_freq_list=concatenateVectors(freq_list_parts);
partitioned_tot_key_filter_list=concatenateVectors(key_filter_list_parts);
partitioned_tot_version_filter_list=concatenateVectors(version_filter_list_parts);
partitioned_tot_freq_filter_list=concatenateVectors(freq_filter_list_parts);

/*
for (int partid = 0; partid < kSavedPartitionNum; partid++) {
std::vector<K>& key_list = key_list_parts[partid];
std::vector<V* >& valueptr_list = valueptr_list_parts[partid];
Expand Down Expand Up @@ -306,6 +342,7 @@ Status DumpEmbeddingValues(EmbeddingVar<K, V>* ev,
part_offset_flat(partid + 1) = part_offset_flat(partid) + key_list.size();
part_filter_offset[partid + 1] = part_filter_offset[partid] + key_filter_list.size();
}
*/
// TODO: DB iterator not support partition_offset
writer->Add(tensor_key + "-partition_offset", *part_offset_tensor);
for(int i = 0; i < kSavedPartitionNum + 1; i++) {
Expand Down