diff --git a/convert-hf-to-gguf-update.py b/convert-hf-to-gguf-update.py
index f43b15760e1b2..fbf1e1ea3de37 100755
--- a/convert-hf-to-gguf-update.py
+++ b/convert-hf-to-gguf-update.py
@@ -83,6 +83,7 @@ class TOKENIZER_TYPE(IntEnum):
     {"name": "jina-v2-es",     "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-es", },
     {"name": "jina-v2-de",     "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-de", },
     {"name": "smaug-bpe",      "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/abacusai/Smaug-Llama-3-70B-Instruct", },
+    {"name": "poro-chat",      "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Poro-34B-chat", },
     {"name": "jina-v2-code",   "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-code", },
 ]
 
diff --git a/convert-hf-to-gguf.py b/convert-hf-to-gguf.py
index a86864f04861b..b2d8e406bdf55 100755
--- a/convert-hf-to-gguf.py
+++ b/convert-hf-to-gguf.py
@@ -475,6 +475,9 @@ def get_vocab_base_pre(self, tokenizer) -> str:
         if chkhsh == "c136ed14d01c2745d4f60a9596ae66800e2b61fa45643e72436041855ad4089d":
             # ref: https://huggingface.co/abacusai/Smaug-Llama-3-70B-Instruct
             res = "smaug-bpe"
+        if chkhsh == "c7ea5862a53e4272c035c8238367063e2b270d51faa48c0f09e9d5b54746c360":
+            # ref: https://huggingface.co/LumiOpen/Poro-34B-chat
+            res = "poro-chat"
         if chkhsh == "7967bfa498ade6b757b064f31e964dddbb80f8f9a4d68d4ba7998fcf281c531a":
             # ref: https://huggingface.co/jinaai/jina-embeddings-v2-base-code
             res = "jina-v2-code"
diff --git a/llama.cpp b/llama.cpp
index 32264a0082a90..88bf4359c0885 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -4705,6 +4705,9 @@ static void llm_load_vocab(
             } else if (
                 tokenizer_pre == "smaug-bpe") {
                 vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_SMAUG;
+            } else if (
+                tokenizer_pre == "poro-chat") {
+                vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_PORO;
             } else {
                 throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
             }
@@ -13025,6 +13028,11 @@ struct llm_tokenizer_bpe {
                             "(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
                         });
                         break;
+                    case LLAMA_VOCAB_PRE_TYPE_PORO:
+                        word_collection = unicode_regex_split(text, {
+                            " ?[^(\\s|.,!?…。，、।۔،)]+",
+                        });
+                        break;
                     default:
                         // default regex for BPE tokenization pre-processing
                         word_collection = unicode_regex_split(text, {
diff --git a/llama.h b/llama.h
index 62908261f2791..da310ffaf9ad9 100644
--- a/llama.h
+++ b/llama.h
@@ -86,6 +86,7 @@ extern "C" {
         LLAMA_VOCAB_PRE_TYPE_OLMO           = 12,
         LLAMA_VOCAB_PRE_TYPE_DBRX           = 13,
         LLAMA_VOCAB_PRE_TYPE_SMAUG          = 14,
+        LLAMA_VOCAB_PRE_TYPE_PORO           = 15,
     };
 
     // note: these values should be synchronized with ggml_rope