Skip to content

Commit 17d7549

Browse files
committed
Update decoder_only.cpp
1 parent 3936fa9 commit 17d7549

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

src/models/decoder_only.cpp

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -29,7 +29,7 @@ void DecoderOnly_State::SetExtraInputs(const std::vector<ExtraInput>& extra_inpu
2929

3030
DeviceSpan<float> DecoderOnly_State::Run(int total_length, DeviceSpan<int32_t>& next_tokens, DeviceSpan<int32_t> next_indices) {
3131
size_t num_tokens = next_tokens.size();
32-
const size_t chunk_size = 15;
32+
const size_t chunk_size = 1024; // Experimental value
3333

3434
if (num_tokens > chunk_size) {
3535
// Chunking logic for context phase - process in chunks of 512 tokens

0 commit comments

Comments
 (0)