CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

examples : replace llama_kv_cache_seq_* with llama_past_seq_*

Browse Source
This commit is contained in:
Francis Couture-Harpin
2024-06-10 14:44:42 -04:00
parent 372482dffe
commit 43d8d4bf9e
23 changed files with 125 additions and 112 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
examples/llama.swiftui/llama.cpp.swift/LibLlama.swift
View File

@@ -214,7 +214,7 @@ actor LlamaContext {
}
batch.logits[Int(batch.n_tokens) - 1] = 1 // true
llama_kv_cache_clear(context)
llama_past_clear(context)
let t_pp_start = ggml_time_us()
@@ -227,7 +227,7 @@ actor LlamaContext {
// bench text generation
llama_kv_cache_clear(context)
llama_past_clear(context)
let t_tg_start = ggml_time_us()
@@ -246,7 +246,7 @@ actor LlamaContext {
let t_tg_end = ggml_time_us()
llama_kv_cache_clear(context)
llama_past_clear(context)
let t_pp = Double(t_pp_end - t_pp_start) / 1000000.0
let t_tg = Double(t_tg_end - t_tg_start) / 1000000.0
@@ -296,7 +296,7 @@ actor LlamaContext {
func clear() {
tokens_list.removeAll()
temporary_invalid_cchars.removeAll()
llama_kv_cache_clear(context)
llama_past_clear(context)
}
private func tokenize(text: String, add_bos: Bool) -> [llama_token] {