LLM: simplify condition

Author: ilya-lavrenov

src/cpp/src/continuous_batching_impl.cpp

@@ -385,7 +385,7 @@ void ContinuousBatchingPipeline::ContinuousBatchingImpl::_free_non_running_reque
     std::vector<SequenceGroup::Ptr>::iterator requests_iterator = m_requests.begin();
     while (requests_iterator != m_requests.end()) {
         const auto& request = *requests_iterator;
-        if(request->has_finished() || request->out_of_memory() || request->handle_dropped()) {
+        if(request->has_finished()) {
             for (const auto& sequence: request->get_sequences()) {
                 if (m_scheduler->has_block_table(sequence->get_id())) {
                     m_scheduler->free_sequence(sequence->get_id());

src/cpp/src/lm_encoding.cpp

@@ -101,7 +101,7 @@ std::pair<EncodedResults, std::optional<int64_t>> get_lm_encoded_results(
     auto free_non_running_requests = [&streamer_ptr, &generations, &active_sequence_groups]() {
         auto removed_it = std::remove_if(active_sequence_groups.begin(), active_sequence_groups.end(),
             [](SequenceGroup::Ptr sg) -> bool {
-                return sg->has_finished() || sg->out_of_memory() || sg->handle_dropped();
+                return sg->has_finished();
             });
         active_sequence_groups.erase(removed_it, active_sequence_groups.end());
     };

src/cpp/src/sequence_group.hpp

@@ -356,8 +356,9 @@ class SequenceGroup  : public std::enable_shared_from_this<SequenceGroup> {
 
     std::vector<Sequence::Ptr> get_running_sequences() {
         std::vector<Sequence::Ptr> running_seqs;
+
         for (size_t seq_id = 0; seq_id < m_sequences.size(); ++seq_id) {
-            if (m_sequences[seq_id]->is_running()) {
+            if (m_sequences[seq_id]->is_running() && !handle_dropped()) {
                 running_seqs.emplace_back(m_sequences[seq_id]);
             }
         }
@@ -366,20 +367,22 @@ class SequenceGroup  : public std::enable_shared_from_this<SequenceGroup> {
     }
 
     std::vector<Sequence::Ptr> get_not_finished_sequences() {
-        std::vector<Sequence::Ptr> running_seqs;
+        std::vector<Sequence::Ptr> not_finished_sequences;
+
         for (size_t seq_id = 0; seq_id < m_sequences.size(); ++seq_id) {
-            if (!m_sequences[seq_id]->has_finished()) {
-                running_seqs.emplace_back(m_sequences[seq_id]);
+            if (!m_sequences[seq_id]->has_finished() && !handle_dropped()) {
+                not_finished_sequences.emplace_back(m_sequences[seq_id]);
             }
         }
 
-        return running_seqs;
+        return not_finished_sequences;
     }
 
     std::vector<Sequence::CPtr> get_running_sequences() const {
         std::vector<Sequence::CPtr> running_seqs;
+
         for (size_t seq_id = 0; seq_id < m_sequences.size(); ++seq_id) {
-            if (m_sequences[seq_id]->is_running()) {
+            if (m_sequences[seq_id]->is_running() && !handle_dropped()) {
                 running_seqs.emplace_back(m_sequences[seq_id]);
             }
         }