Spaces:
Sleeping
Sleeping
vidhiparikh
commited on
Commit
•
fe4fe9d
1
Parent(s):
746d530
Update app.py
Browse files
app.py
CHANGED
@@ -74,9 +74,22 @@ callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
|
|
74 |
|
75 |
# Function to create a conversational chain
|
76 |
def create_conversational_chain(database):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
77 |
llama_llm = LlamaCpp(
|
78 |
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system. llama-2-7b-chat.Q8_0.gguf
|
79 |
-
model_path =
|
80 |
temperature=0.75,
|
81 |
max_tokens=200,
|
82 |
top_p=1,
|
|
|
74 |
|
75 |
# Function to create a conversational chain
|
76 |
def create_conversational_chain(database):
|
77 |
+
model_name = “llama-7b-hf”
|
78 |
+
model_directory = “models”
|
79 |
+
#Check if the model file exists in the specified directory
|
80 |
+
model_file = os.path.join(model_directory, model_name)
|
81 |
+
if os.path.exists(model_file):
|
82 |
+
model_path = model_file
|
83 |
+
print(“Model file found in the directory. Using the local model file.”)
|
84 |
+
else:
|
85 |
+
model_path = model_name
|
86 |
+
print(“Model file not found in the directory. Downloading the model from the repository.”)
|
87 |
+
#Load the model
|
88 |
+
model = AutoModelForCausalLM.from_pretrained(model_path)
|
89 |
+
print(model_path)
|
90 |
llama_llm = LlamaCpp(
|
91 |
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system. llama-2-7b-chat.Q8_0.gguf
|
92 |
+
model_path = model_path,
|
93 |
temperature=0.75,
|
94 |
max_tokens=200,
|
95 |
top_p=1,
|