Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -2,11 +2,18 @@ import gradio as gr
|
|
| 2 |
from huggingface_hub import hf_hub_download
|
| 3 |
from llama_cpp import Llama
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
# 1. Download the high-performance CPU file from your new repository
|
| 6 |
print("📥 Fetching optimized GGUF weights...")
|
| 7 |
model_path = hf_hub_download(
|
| 8 |
repo_id="UH32/Storyteller-Llama3-GGUF",
|
| 9 |
-
filename="
|
| 10 |
)
|
| 11 |
|
| 12 |
# 2. Initialize the model inside the CPU memory footprint
|
|
|
|
| 2 |
from huggingface_hub import hf_hub_download
|
| 3 |
from llama_cpp import Llama
|
| 4 |
|
| 5 |
+
# # 1. Download the high-performance CPU file from your new repository
|
| 6 |
+
# print("📥 Fetching optimized GGUF weights...")
|
| 7 |
+
# model_path = hf_hub_download(
|
| 8 |
+
# repo_id="UH32/Storyteller-Llama3-GGUF",
|
| 9 |
+
# filename="unsloth.Q4_K_M.gguf" # This is the standard output name generated by Unsloth
|
| 10 |
+
# )
|
| 11 |
+
|
| 12 |
# 1. Download the high-performance CPU file from your new repository
|
| 13 |
print("📥 Fetching optimized GGUF weights...")
|
| 14 |
model_path = hf_hub_download(
|
| 15 |
repo_id="UH32/Storyteller-Llama3-GGUF",
|
| 16 |
+
filename="llama-3.1-8b.Q4_K_M.gguf" # <-- CHANGED THIS LINE
|
| 17 |
)
|
| 18 |
|
| 19 |
# 2. Initialize the model inside the CPU memory footprint
|