vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

better check

Browse source
This commit is contained in:
ngxson 2024-07-23 20:48:44 +02:00
parent 0ec2b5818a
commit 4bd55ec02c
1 changed files with 7 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

11
examples/export-lora/export-lora.cpp
View file

@ -197,12 +197,15 @@ struct lora_merge_ctx {
// check if all lora adapters have the same tensors
// TODO: remove this when we can support merging subset of adapters. Ref: https://github.com/ggerganov/llama.cpp/pull/8607#discussion_r1686027777
static const char * err_no_subset_adapter = "Input adapters do not have the same list of tensors. This is not yet supported. Please merge the adapter one-by-one instead of merging all at once.";
if (adapters.size() > 1) {
auto & base_adapter = adapters[0];
for (size_t i = 1; i < adapters.size(); ++i) {
for (auto & it : base_adapter->tensors) {
if (base_adapter->get_tensor(it.first) == nullptr) {
throw std::runtime_error("Input adapters do not have the same list of tensors. This is not yet supported. Please merge the adapter one-by-one instead of merging all at once.");
if (adapters[0]->tensors.size() != adapters[i]->tensors.size()) {
throw std::runtime_error(err_no_subset_adapter);
}
for (auto & it : adapters[0]->tensors) {
if (adapters[0]->get_tensor(it.first) == nullptr) {
throw std::runtime_error(err_no_subset_adapter);
}
}
}