From e042bf86248f1ecff9d95fae49812750c27ce758 Mon Sep 17 00:00:00 2001
From: cal066 <60696996+cal066@users.noreply.github.com>
Date: Tue, 22 Aug 2023 19:51:34 +0000
Subject: [PATCH] ctransformers: add mlock and no-mmap options (#3649)

---
 modules/ctransformers_model.py | 4 +++-
 modules/loaders.py             | 4 +++-
 requirements.txt               | 2 +-
 3 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/modules/ctransformers_model.py b/modules/ctransformers_model.py
index 5e0f347c..8b8b5c4d 100644
--- a/modules/ctransformers_model.py
+++ b/modules/ctransformers_model.py
@@ -19,7 +19,9 @@ class CtransformersModel:
             gpu_layers=shared.args.n_gpu_layers,
             batch_size=shared.args.n_batch,
             context_length=shared.args.n_ctx,
-            stream=True
+            stream=True,
+            mmap=not shared.args.no_mmap,
+            mlock=shared.args.mlock
         )
 
         self.model = AutoModelForCausalLM.from_pretrained(
diff --git a/modules/loaders.py b/modules/loaders.py
index 7444555f..472e8ddb 100644
--- a/modules/loaders.py
+++ b/modules/loaders.py
@@ -101,7 +101,9 @@ loaders_and_params = OrderedDict({
         'n_gpu_layers',
         'n_batch',
         'threads',
-        'model_type'
+        'model_type',
+        'no_mmap',
+        'mlock'
     ]
 })
 
diff --git a/requirements.txt b/requirements.txt
index 05a3ec0e..25c953ee 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -42,4 +42,4 @@ https://github.com/jllllll/GPTQ-for-LLaMa-CUDA/releases/download/0.1.0/gptq_for_
 https://github.com/jllllll/GPTQ-for-LLaMa-CUDA/releases/download/0.1.0/gptq_for_llama-0.1.0+cu117-cp310-cp310-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
 
 # ctransformers
-https://github.com/jllllll/ctransformers-cuBLAS-wheels/releases/download/AVX2/ctransformers-0.2.22+cu117-py3-none-any.whl
+https://github.com/jllllll/ctransformers-cuBLAS-wheels/releases/download/AVX2/ctransformers-0.2.23+cu117-py3-none-any.whl