Skip to content

Fix crashes during multi-turn conversation due to KV cache serialization #9

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

Merged
JohannesGaessler merged 1 commit into from
Feb 28, 2026
Merged

Fix crashes during multi-turn conversation due to KV cache serialization #9

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
8 changes: 4 additions & 4 deletions ggml/src/ggml-backend-meta.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -496,7 +496,6 @@ static enum ggml_status ggml_backend_meta_buffer_init_tensor(ggml_backend_buffer

static void ggml_backend_meta_buffer_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) {
const size_t n_bufs = ggml_backend_meta_buffer_n_bufs(buffer);
GGML_ASSERT(offset == 0);
GGML_ASSERT(ggml_is_contiguous(tensor));

const ggml_backend_meta_split_state split_state = ggml_backend_meta_get_split_state(tensor, /*assume_sync =*/ false);
Expand All @@ -515,7 +514,8 @@ static void ggml_backend_meta_buffer_set_tensor(ggml_backend_buffer_t buffer, gg
for (size_t j = 0; j < n_bufs; j++){
ggml_tensor * simple_tensor = ggml_backend_meta_buffer_simple_tensor(tensor, j);
const size_t chunk_size_j = simple_tensor->nb[split_state.axis + 1];
ggml_backend_tensor_set_2d(simple_tensor, (const char *) data + offset_j, offset, chunk_size_j, i_stop - i_start, chunk_size_j, chunk_size_full);
const size_t simple_offset = i_start * chunk_size_j;
ggml_backend_tensor_set_2d(simple_tensor, (const char *) data + offset_j, simple_offset, chunk_size_j, i_stop - i_start, chunk_size_j, chunk_size_full);
offset_j += chunk_size_j;
}
GGML_ASSERT(offset_j == chunk_size_full);
Expand All @@ -534,7 +534,6 @@ static void ggml_backend_meta_buffer_set_tensor(ggml_backend_buffer_t buffer, gg

static void ggml_backend_meta_buffer_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) {
const size_t n_bufs = ggml_backend_meta_buffer_n_bufs(buffer);
GGML_ASSERT(offset == 0);
GGML_ASSERT(ggml_is_contiguous(tensor));

const ggml_backend_meta_split_state split_state = ggml_backend_meta_get_split_state(tensor, /*assume_sync =*/ false);
Expand All @@ -553,7 +552,8 @@ static void ggml_backend_meta_buffer_get_tensor(ggml_backend_buffer_t buffer, co
for (size_t j = 0; j < n_bufs; j++){
const ggml_tensor * simple_tensor = ggml_backend_meta_buffer_simple_tensor(tensor, j);
const size_t chunk_size_j = simple_tensor->nb[split_state.axis + 1];
ggml_backend_tensor_get_2d(simple_tensor, (char *) data + offset_j, offset, chunk_size_j, i_stop - i_start, chunk_size_j, chunk_size_full);
const size_t simple_offset = i_start * chunk_size_j;
ggml_backend_tensor_get_2d(simple_tensor, (char *) data + offset_j, simple_offset, chunk_size_j, i_stop - i_start, chunk_size_j, chunk_size_full);
offset_j += chunk_size_j;
}
GGML_ASSERT(offset_j == chunk_size_full);
Expand Down