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

llama : support NUL bytes in tokens

Browse source
This commit is contained in:
Francis Couture-Harpin 2024-08-11 21:00:03 -04:00
parent 4134999e01
commit faaac59d16
7 changed files with 28 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml/include/ggml.h
View file

@ -2313,10 +2313,12 @@ extern "C" {
GGML_API double gguf_get_val_f64 (const struct gguf_context * ctx, int key_id);
GGML_API bool gguf_get_val_bool(const struct gguf_context * ctx, int key_id);
GGML_API const char * gguf_get_val_str (const struct gguf_context * ctx, int key_id);
GGML_API int gguf_get_val_str_n(const struct gguf_context * ctx, int key_id);
GGML_API const void * gguf_get_val_data(const struct gguf_context * ctx, int key_id);
GGML_API int gguf_get_arr_n (const struct gguf_context * ctx, int key_id);
GGML_API const void * gguf_get_arr_data(const struct gguf_context * ctx, int key_id);
GGML_API const char * gguf_get_arr_str (const struct gguf_context * ctx, int key_id, int i);
GGML_API int gguf_get_arr_str_n(const struct gguf_context * ctx, int key_id, int i);
GGML_API int gguf_get_n_tensors (const struct gguf_context * ctx);
GGML_API int gguf_find_tensor (const struct gguf_context * ctx, const char * name);

14
ggml/src/ggml.c
View file

@ -21335,6 +21335,14 @@ const char * gguf_get_arr_str(const struct gguf_context * ctx, int key_id, int i
return str->data;
}
int gguf_get_arr_str_n(const struct gguf_context * ctx, int key_id, int i) {
GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx));
GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY);
struct gguf_kv * kv = &ctx->kv[key_id];
struct gguf_str * str = &((struct gguf_str *) kv->value.arr.data)[i];
return str->n;
}
int gguf_get_arr_n(const struct gguf_context * ctx, int key_id) {
GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx));
GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_ARRAY);
@ -21413,6 +21421,12 @@ const char * gguf_get_val_str(const struct gguf_context * ctx, int key_id) {
return ctx->kv[key_id].value.str.data;
}
int gguf_get_val_str_n(const struct gguf_context * ctx, int key_id) {
GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx));
GGML_ASSERT(ctx->kv[key_id].type == GGUF_TYPE_STRING);
return ctx->kv[key_id].value.str.n;
}
const void * gguf_get_val_data(const struct gguf_context * ctx, int key_id) {
GGML_ASSERT(key_id >= 0 && key_id < gguf_get_n_kv(ctx));
GGML_ASSERT(ctx->kv[key_id].type != GGUF_TYPE_ARRAY);