matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-11 08:17:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: update to commit 71e90e88 (#10192)

Browse Source
This commit is contained in:
Jeffrey Morgan
2025-04-16 18:14:01 -04:00
committed by GitHub
parent 369de832cd
commit 943464ccb8
160 changed files with 42219 additions and 33080 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
llama/patches/0002-pretokenizer.patch
View File

@@ -3,15 +3,17 @@ From: Michael Yang <mxyng@pm.me>
Date: Mon, 16 Sep 2024 15:53:13 -0700
Subject: [PATCH] pretokenizer
allow for an unset pretokenizer with a warning in the
logs instead of throwing an error
---
src/llama-vocab.cpp | 14 +++-----------
1 file changed, 3 insertions(+), 11 deletions(-)
diff --git a/src/llama-vocab.cpp b/src/llama-vocab.cpp
index ad9ffe66..a4eee9b8 100644
index 464ff01e..0125ee53 100644
--- a/src/llama-vocab.cpp
+++ b/src/llama-vocab.cpp
@@ -1468,16 +1468,7 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
@@ -1491,16 +1491,7 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
if (type == LLAMA_VOCAB_TYPE_BPE) {
add_space_prefix = false;
clean_spaces = true;
@@ -29,9 +31,9 @@ index ad9ffe66..a4eee9b8 100644
pre_type = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
} else if (
tokenizer_pre == "llama3" ||
@@ -1593,7 +1584,8 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
tokenizer_pre == "megrez") {
pre_type = LLAMA_VOCAB_PRE_TYPE_QWEN2;
@@ -1634,7 +1625,8 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
pre_type = LLAMA_VOCAB_PRE_TYPE_BAILINGMOE;
clean_spaces = false;
} else {
- throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
+ LLAMA_LOG_WARN("%s: missing or unrecognized pre-tokenizer type, using: 'default'\n", __func__);