Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
@@ -20,12 +20,7 @@ from pathlib import Path
|
|
20 |
|
21 |
model_subdir = Path.home() / ".cache" / "huggingface" / "hub" / "models--unsloth--Llama-3.2-3B"
|
22 |
|
23 |
-
|
24 |
-
print(f"🧹 Deleting model cache at {model_subdir}")
|
25 |
-
shutil.rmtree(model_subdir)
|
26 |
-
print("✅ Model removed.")
|
27 |
-
else:
|
28 |
-
print("⚠️ Specified model not found in cache.")
|
29 |
|
30 |
# Enable persistent caching on Hugging Face Spaces (if persistent storage is enabled)
|
31 |
os.environ["TRANSFORMERS_CACHE"] = "/data/models"
|
@@ -41,8 +36,8 @@ MAX_FILE_SIZE_MB = 20
|
|
41 |
MAX_FILE_SIZE_BYTES = MAX_FILE_SIZE_MB * 1024 * 1024
|
42 |
|
43 |
MODEL_ID = "meta-llama/Meta-Llama-3-8B"# unsloth/Llama-3.2-3B"#meta-llama/Meta-Llama-3-8B" #"unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF"# unsloth/Qwen2.5-1.5B" #unsloth/Llama-3.2-3B" #unsloth/Llama-3.2-1B"
|
44 |
-
glotoken=Tokentest
|
45 |
-
# Global logging system -
|
46 |
logs = []
|
47 |
|
48 |
def add_log(message):
|
@@ -50,7 +45,7 @@ def add_log(message):
|
|
50 |
logs.append(f"[{time.strftime('%H:%M:%S')}] {message}")
|
51 |
print(message)
|
52 |
|
53 |
-
# Initialize model with comprehensive error handling
|
54 |
model = None
|
55 |
tokenizer = None
|
56 |
generation_config = None
|
|
|
20 |
|
21 |
model_subdir = Path.home() / ".cache" / "huggingface" / "hub" / "models--unsloth--Llama-3.2-3B"
|
22 |
|
23 |
+
|
|
|
|
|
|
|
|
|
|
|
24 |
|
25 |
# Enable persistent caching on Hugging Face Spaces (if persistent storage is enabled)
|
26 |
os.environ["TRANSFORMERS_CACHE"] = "/data/models"
|
|
|
36 |
MAX_FILE_SIZE_BYTES = MAX_FILE_SIZE_MB * 1024 * 1024
|
37 |
|
38 |
MODEL_ID = "meta-llama/Meta-Llama-3-8B"# unsloth/Llama-3.2-3B"#meta-llama/Meta-Llama-3-8B" #"unsloth/Llama-4-Scout-17B-16E-Instruct-GGUF"# unsloth/Qwen2.5-1.5B" #unsloth/Llama-3.2-3B" #unsloth/Llama-3.2-1B"
|
39 |
+
glotoken = os.environ.get("Tokentest")
|
40 |
+
# Global logging system -
|
41 |
logs = []
|
42 |
|
43 |
def add_log(message):
|
|
|
45 |
logs.append(f"[{time.strftime('%H:%M:%S')}] {message}")
|
46 |
print(message)
|
47 |
|
48 |
+
# Initialize model with comprehensive error handling
|
49 |
model = None
|
50 |
tokenizer = None
|
51 |
generation_config = None
|