From df9362eeea09a2f1e2649c3ad658673b6ad159d4 Mon Sep 17 00:00:00 2001
From: Andrei Betlen <abetlen@gmail.com>
Date: Fri, 3 Nov 2023 11:34:50 -0400
Subject: [PATCH] Update llama.cpp

---
 llama_cpp/llama.py      |  2 +-
 llama_cpp/llama_cpp.py  | 10 +++++-----
 llama_cpp/server/app.py |  2 +-
 vendor/llama.cpp        |  2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/llama_cpp/llama.py b/llama_cpp/llama.py
index 3e8cf58..8c69210 100644
--- a/llama_cpp/llama.py
+++ b/llama_cpp/llama.py
@@ -233,7 +233,7 @@ class Llama:
         rope_scaling_type: Optional[int] = llama_cpp.LLAMA_ROPE_SCALING_UNSPECIFIED,
         rope_freq_base: float = 0.0,
         rope_freq_scale: float = 0.0,
-        yarn_ext_factor: float = float("nan"),
+        yarn_ext_factor: float = -1.0,
         yarn_attn_factor: float = 1.0,
         yarn_beta_fast: float = 32.0,
         yarn_beta_slow: float = 1.0,
diff --git a/llama_cpp/llama_cpp.py b/llama_cpp/llama_cpp.py
index b6216a5..48958e5 100644
--- a/llama_cpp/llama_cpp.py
+++ b/llama_cpp/llama_cpp.py
@@ -315,11 +315,11 @@ class llama_model_params(Structure):
 
 
 # struct llama_context_params {
-#     uint32_t seed;            // RNG seed, -1 for random
-#     uint32_t n_ctx;           // text context, 0 = from model
-#     uint32_t n_batch;         // prompt processing maximum batch size
-#     uint32_t n_threads;       // number of threads to use for generation
-#     uint32_t n_threads_batch; // number of threads to use for batch processing
+#     uint32_t seed;              // RNG seed, -1 for random
+#     uint32_t n_ctx;             // text context, 0 = from model
+#     uint32_t n_batch;           // prompt processing maximum batch size
+#     uint32_t n_threads;         // number of threads to use for generation
+#     uint32_t n_threads_batch;   // number of threads to use for batch processing
 #     int8_t   rope_scaling_type; // RoPE scaling type, from `enum llama_rope_scaling_type`
 
 #     // ref: https://github.com/ggerganov/llama.cpp/pull/2054
diff --git a/llama_cpp/server/app.py b/llama_cpp/server/app.py
index bec9561..93afc3e 100644
--- a/llama_cpp/server/app.py
+++ b/llama_cpp/server/app.py
@@ -93,7 +93,7 @@ class Settings(BaseSettings):
         default=0.0, description="RoPE frequency scaling factor"
     )
     yarn_ext_factor: float = Field(
-        default=float("nan")
+        default=-1.0
     )
     yarn_attn_factor: float = Field(
         default=1.0
diff --git a/vendor/llama.cpp b/vendor/llama.cpp
index 629f917..abb77e7 160000
--- a/vendor/llama.cpp
+++ b/vendor/llama.cpp
@@ -1 +1 @@
-Subproject commit 629f917cd6b96ba1274c49a8aab163b1b189229d
+Subproject commit abb77e7319aabc0b5cfb7c22da690a692489b6b7