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

add more logging

Browse source
This commit is contained in:
ngxson 2024-07-23 20:01:23 +02:00
parent 793a1cdc6d
commit 1f965737f0
1 changed files with 15 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

17
examples/export-lora/export-lora.cpp
View file

@ -27,6 +27,17 @@ static void zeros(std::ofstream & file, size_t n) {
} }
} }
static std::string ggml_ne_string(const ggml_tensor * t) {
std::string str;
for (int i = 0; i < GGML_MAX_DIMS; ++i) {
str += std::to_string(t->ne[i]);
if (i + 1 < GGML_MAX_DIMS) {
str += ", ";
}
}
return str;
}
static struct gguf_context * load_gguf(std::string & fname, struct ggml_context ** ctx_ggml) { static struct gguf_context * load_gguf(std::string & fname, struct ggml_context ** ctx_ggml) {
struct gguf_init_params params = { struct gguf_init_params params = {
/*.no_alloc = */ true, /*.no_alloc = */ true,
@ -242,7 +253,7 @@ struct lora_merge_ctx {
} }
void copy_tensor(struct ggml_tensor * base) { void copy_tensor(struct ggml_tensor * base) {
printf("%s : %s\n", __func__, base->name); printf("%s : %s [%s]\n", __func__, base->name, ggml_ne_string(base).c_str());
size_t len = ggml_nbytes(base); size_t len = ggml_nbytes(base);
base_model.read_tensor_data(base->name, read_buf); base_model.read_tensor_data(base->name, read_buf);
fout.write((char* )read_buf.data(), len); fout.write((char* )read_buf.data(), len);
@ -254,7 +265,7 @@ struct lora_merge_ctx {
std::string name_lora_a = name_base + ".lora_a"; std::string name_lora_a = name_base + ".lora_a";
std::string name_lora_b = name_base + ".lora_b"; std::string name_lora_b = name_base + ".lora_b";
printf("%s : %s\n", __func__, base->name); printf("%s : %s [%s]\n", __func__, base->name, ggml_ne_string(base).c_str());
// context for input tensor // context for input tensor
std::vector<struct ggml_tensor *> inp_a(adapters.size()); std::vector<struct ggml_tensor *> inp_a(adapters.size());
@ -308,6 +319,8 @@ struct lora_merge_ctx {
const float scale = alpha ? adapters[i]->scale * alpha / rank : adapters[i]->scale; const float scale = alpha ? adapters[i]->scale * alpha / rank : adapters[i]->scale;
delta = ggml_scale(ctx0, delta, scale); delta = ggml_scale(ctx0, delta, scale);
cur = ggml_add(ctx0, cur, delta); cur = ggml_add(ctx0, cur, delta);
printf("%s : merging from adapter[%ld]\n", __func__, i);
printf("%s : input_scale=%f calculated_scale=%f rank=%d\n", __func__, adapters[i]->scale, scale, (int) inp_b[i]->ne[0]);
} }
cur = ggml_cast(ctx0, cur, get_out_tensor_type(base)); cur = ggml_cast(ctx0, cur, get_out_tensor_type(base));
ggml_build_forward_expand(gf, cur); ggml_build_forward_expand(gf, cur);