extern/whisper.cpp
1
0
Fork 1
mirror of https://github.com/ggerganov/whisper.cpp.git synced 2025-08-14 12:49:13 +02:00
Code Issues Packages Projects Releases Wiki Activity

talk-llama : sync llama.cpp

Browse Source
This commit is contained in:
Georgi Gerganov
2024-10-02 15:14:46 +03:00
parent 162a455402
commit ccc2547210
5 changed files with 588 additions and 133 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
examples/talk-llama/llama.h
View File

@ -102,6 +102,7 @@ extern "C" {
LLAMA_VOCAB_PRE_TYPE_BLOOM = 23,
LLAMA_VOCAB_PRE_TYPE_GPT3_FINNISH = 24,
LLAMA_VOCAB_PRE_TYPE_EXAONE = 25,
LLAMA_VOCAB_PRE_TYPE_CHAMELEON = 26,
};
enum llama_rope_type {
@ -192,6 +193,7 @@ extern "C" {
LLAMA_POOLING_TYPE_MEAN = 1,
LLAMA_POOLING_TYPE_CLS = 2,
LLAMA_POOLING_TYPE_LAST = 3,
LLAMA_POOLING_TYPE_RANK = 4, // used by reranking models to attach the classification head to the graph
};
enum llama_attention_type {
@ -201,9 +203,9 @@ extern "C" {
};
enum llama_split_mode {
LLAMA_SPLIT_MODE_NONE = 0, // single GPU
LLAMA_SPLIT_MODE_LAYER = 1, // split layers and KV across GPUs
LLAMA_SPLIT_MODE_ROW = 2, // split rows across GPUs
LLAMA_SPLIT_MODE_NONE = 0, // single GPU
LLAMA_SPLIT_MODE_LAYER = 1, // split layers and KV across GPUs
LLAMA_SPLIT_MODE_ROW = 2, // split rows across GPUs
};
// TODO: simplify (https://github.com/ggerganov/llama.cpp/pull/9294#pullrequestreview-2286561979)
@ -871,7 +873,8 @@ extern "C" {
// Get the embeddings for a sequence id
// Returns NULL if pooling_type is LLAMA_POOLING_TYPE_NONE
// shape: [n_embd] (1-dimensional)
// when pooling_type == LLAMA_POOLING_TYPE_RANK, returns float[1] with the rank of the sequence
// otherwise: float[n_embd] (1-dimensional)
LLAMA_API float * llama_get_embeddings_seq(struct llama_context * ctx, llama_seq_id seq_id);
//
@ -910,6 +913,8 @@ extern "C" {
//
// Tokenization
//
// The API is thread-safe.
//
/// @details Convert the provided text into tokens.
/// @param tokens The tokens pointer must be large enough to hold the resulting tokens.