Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -2,11 +2,9 @@ from fastapi import FastAPI, HTTPException
|
|
2 |
from pydantic import BaseModel
|
3 |
from typing import Optional
|
4 |
import torch
|
5 |
-
from transformers import AutoTokenizer, AutoModelForCausalLM,
|
6 |
import logging
|
7 |
-
import os
|
8 |
|
9 |
-
# ตั้งค่า logging
|
10 |
logging.basicConfig(level=logging.INFO)
|
11 |
logger = logging.getLogger(__name__)
|
12 |
|
@@ -16,32 +14,21 @@ try:
|
|
16 |
model_name = "scb10x/llama-3-typhoon-v1.5-8b-instruct"
|
17 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
18 |
|
19 |
-
|
20 |
-
|
21 |
-
logger.info("GPU is available. Using CUDA.")
|
22 |
-
device = "cuda"
|
23 |
-
else:
|
24 |
-
logger.info("No GPU found. Using CPU.")
|
25 |
-
device = "cpu"
|
26 |
|
27 |
-
#
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
if device == "cuda":
|
34 |
-
from transformers import BitsAndBytesConfig
|
35 |
-
model_kwargs["quantization_config"] = BitsAndBytesConfig(load_in_8bit=True)
|
36 |
|
37 |
-
# โหลดโมเดล
|
38 |
model = AutoModelForCausalLM.from_pretrained(
|
39 |
model_name,
|
40 |
-
|
41 |
-
|
|
|
42 |
)
|
43 |
-
|
44 |
-
model.to(device)
|
45 |
logger.info(f"Model loaded successfully on {device}")
|
46 |
except Exception as e:
|
47 |
logger.error(f"Error loading model: {str(e)}")
|
|
|
2 |
from pydantic import BaseModel
|
3 |
from typing import Optional
|
4 |
import torch
|
5 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
|
6 |
import logging
|
|
|
7 |
|
|
|
8 |
logging.basicConfig(level=logging.INFO)
|
9 |
logger = logging.getLogger(__name__)
|
10 |
|
|
|
14 |
model_name = "scb10x/llama-3-typhoon-v1.5-8b-instruct"
|
15 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
16 |
|
17 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
18 |
+
logger.info(f"Using device: {device}")
|
|
|
|
|
|
|
|
|
|
|
19 |
|
20 |
+
# 4-bit quantization configuration
|
21 |
+
quantization_config = BitsAndBytesConfig(
|
22 |
+
load_in_4bit=True,
|
23 |
+
bnb_4bit_compute_dtype=torch.float16
|
24 |
+
)
|
|
|
|
|
|
|
|
|
25 |
|
|
|
26 |
model = AutoModelForCausalLM.from_pretrained(
|
27 |
model_name,
|
28 |
+
quantization_config=quantization_config,
|
29 |
+
device_map="auto",
|
30 |
+
low_cpu_mem_usage=True,
|
31 |
)
|
|
|
|
|
32 |
logger.info(f"Model loaded successfully on {device}")
|
33 |
except Exception as e:
|
34 |
logger.error(f"Error loading model: {str(e)}")
|