Skip to content

UPSTREAM PR #19280: fix: only reset LoRa configs when they have changed from previous batch #1142

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

Open
loci-dev wants to merge 1 commit into
base: main
Choose a base branch
from
Open

UPSTREAM PR #19280: fix: only reset LoRa configs when they have changed from previous batch #1142

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
11 changes: 9 additions & 2 deletions tools/server/server-context.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,6 +136,7 @@ struct server_slot {
}

std::vector<common_adapter_lora_info> lora;
std::vector<common_adapter_lora_info> previous_lora;
int32_t alora_invocation_start = -1;

// sampling
Expand Down Expand Up @@ -191,6 +192,9 @@ struct server_slot {

// clear alora start
alora_invocation_start = -1;

lora.clear();
previous_lora.clear();
}

void init_sampler() const {
Expand Down Expand Up @@ -2576,8 +2580,11 @@ struct server_context_impl {
SRV_DBG("decoding batch, n_tokens = %d\n", batch.n_tokens);

if (slot_batched) {
// apply lora, only need to do it once per batch
common_set_adapter_lora(ctx, slot_batched->lora);
if (!are_lora_equal(slot_batched->previous_lora, slot_batched->lora)) {
// apply lora, only need to do it once per batch
common_set_adapter_lora(ctx, slot_batched->lora);
slot_batched->previous_lora = slot_batched->lora;
}

// if the lora is temporarily disabled for an alora, re-enable it
// for next time
Expand Down