Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -1,17 +1,17 @@
|
|
1 |
import gradio as gr
|
2 |
import re
|
3 |
import torch
|
4 |
-
import sqlite3
|
|
|
5 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
6 |
from peft import PeftModel, PeftConfig
|
7 |
-
import torch
|
8 |
-
import os
|
9 |
-
os.makedirs("offload", exist_ok=True)
|
10 |
|
11 |
-
# β
|
12 |
-
|
13 |
-
mistral_model_path = "srishtirai/mistral-sql-finetuned" # Upload to HF Model Hub
|
14 |
|
|
|
|
|
|
|
15 |
|
16 |
def load_model(model_path):
|
17 |
tokenizer = AutoTokenizer.from_pretrained(model_path)
|
@@ -21,18 +21,24 @@ def load_model(model_path):
|
|
21 |
peft_config = PeftConfig.from_pretrained(model_path)
|
22 |
base_model_name = peft_config.base_model_name_or_path
|
23 |
|
|
|
24 |
base_model = AutoModelForCausalLM.from_pretrained(
|
25 |
base_model_name,
|
26 |
-
torch_dtype=torch.float16, # Use FP16 to
|
27 |
-
device_map="auto", # Automatically
|
28 |
-
offload_folder="offload" # β
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
)
|
30 |
|
31 |
-
model = PeftModel.from_pretrained(base_model, model_path)
|
32 |
model.eval()
|
33 |
return model, tokenizer
|
34 |
|
35 |
-
|
36 |
# β
Load both models from Hugging Face
|
37 |
codellama_model, codellama_tokenizer = load_model(codellama_model_path)
|
38 |
mistral_model, mistral_tokenizer = load_model(mistral_model_path)
|
@@ -93,14 +99,14 @@ def generate_sql_with_explanation(model_choice, schema, question, max_new_tokens
|
|
93 |
"full_response": full_response
|
94 |
}
|
95 |
|
96 |
-
# β
Function to execute SQL query (
|
97 |
def execute_sql_query(sql_query):
|
98 |
"""
|
99 |
Runs the generated SQL query on a sample SQLite database.
|
100 |
-
(
|
101 |
"""
|
102 |
try:
|
103 |
-
conn = sqlite3.connect(":memory:") # Temporary SQLite DB
|
104 |
cursor = conn.cursor()
|
105 |
cursor.execute(sql_query)
|
106 |
result = cursor.fetchall()
|
|
|
1 |
import gradio as gr
|
2 |
import re
|
3 |
import torch
|
4 |
+
import sqlite3 # Can be replaced with other DB connections
|
5 |
+
import os
|
6 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
7 |
from peft import PeftModel, PeftConfig
|
|
|
|
|
|
|
8 |
|
9 |
+
# β
Ensure offload directory exists
|
10 |
+
os.makedirs("offload", exist_ok=True)
|
|
|
11 |
|
12 |
+
# β
Load fine-tuned models from Hugging Face Model Hub
|
13 |
+
codellama_model_path = "srishtirai/codellama-sql-finetuned"
|
14 |
+
mistral_model_path = "srishtirai/mistral-sql-finetuned"
|
15 |
|
16 |
def load_model(model_path):
|
17 |
tokenizer = AutoTokenizer.from_pretrained(model_path)
|
|
|
21 |
peft_config = PeftConfig.from_pretrained(model_path)
|
22 |
base_model_name = peft_config.base_model_name_or_path
|
23 |
|
24 |
+
# β
Load base model with offloading & low-memory optimization
|
25 |
base_model = AutoModelForCausalLM.from_pretrained(
|
26 |
base_model_name,
|
27 |
+
torch_dtype=torch.float16, # Use FP16 to reduce memory usage
|
28 |
+
device_map="auto", # Automatically distribute across CPU/GPU
|
29 |
+
offload_folder="offload" # β
Prevents memory crashes
|
30 |
+
)
|
31 |
+
|
32 |
+
# β
Load LoRA adapter with `is_trainable=False`
|
33 |
+
model = PeftModel.from_pretrained(
|
34 |
+
base_model,
|
35 |
+
model_path,
|
36 |
+
is_trainable=False # β
Fixes LoRA adapter loading issues
|
37 |
)
|
38 |
|
|
|
39 |
model.eval()
|
40 |
return model, tokenizer
|
41 |
|
|
|
42 |
# β
Load both models from Hugging Face
|
43 |
codellama_model, codellama_tokenizer = load_model(codellama_model_path)
|
44 |
mistral_model, mistral_tokenizer = load_model(mistral_model_path)
|
|
|
99 |
"full_response": full_response
|
100 |
}
|
101 |
|
102 |
+
# β
Function to execute SQL query (if database connection is available)
|
103 |
def execute_sql_query(sql_query):
|
104 |
"""
|
105 |
Runs the generated SQL query on a sample SQLite database.
|
106 |
+
(You can replace SQLite with a connection to a real database)
|
107 |
"""
|
108 |
try:
|
109 |
+
conn = sqlite3.connect(":memory:") # Temporary SQLite DB (Replace with actual DB connection)
|
110 |
cursor = conn.cursor()
|
111 |
cursor.execute(sql_query)
|
112 |
result = cursor.fetchall()
|