Skip to content

Commit d25473a

Browse files
authored
Update decoder_only.cpp
1 parent 9673ff4 commit d25473a

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
@@ -26,7 +26,7 @@ DecoderOnly_State::DecoderOnly_State(const DecoderOnly_Model& model, DeviceSpan<
2626

2727
DeviceSpan<float> DecoderOnly_State::Run(int total_length, DeviceSpan<int32_t>& next_tokens, DeviceSpan<int32_t> next_indices) {
2828
size_t num_tokens = next_tokens.size();
29-
const size_t chunk_size = 15;
29+
const size_t chunk_size = 1024; // Experimental value
3030

3131
if (num_tokens > chunk_size) {
3232
// Chunking logic for context phase - process in chunks of 512 tokens

0 commit comments

Comments
 (0)