Spaces:
Running
on
Zero
Running
on
Zero
fixing app.py
Browse files
app.py
CHANGED
@@ -2,12 +2,12 @@ import os
|
|
2 |
import warnings
|
3 |
import torch
|
4 |
import gc
|
5 |
-
from transformers import
|
6 |
from PIL import Image
|
7 |
import gradio as gr
|
8 |
from huggingface_hub import login
|
9 |
|
10 |
-
#
|
11 |
warnings.filterwarnings('ignore')
|
12 |
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
13 |
|
@@ -15,7 +15,7 @@ os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
|
15 |
model = None
|
16 |
processor = None
|
17 |
|
18 |
-
#
|
19 |
if torch.cuda.is_available():
|
20 |
torch.cuda.empty_cache()
|
21 |
gc.collect()
|
@@ -34,24 +34,35 @@ def load_model_and_processor():
|
|
34 |
print("กำลังโหลดโมเดลและ processor...")
|
35 |
|
36 |
try:
|
37 |
-
#
|
|
|
38 |
hub_model_path = "Aekanun/thai-handwriting-llm"
|
39 |
|
40 |
-
#
|
41 |
-
|
42 |
-
|
43 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
44 |
)
|
45 |
|
|
|
46 |
print("กำลังโหลดโมเดลจาก Hub...")
|
47 |
-
model =
|
48 |
hub_model_path,
|
49 |
device_map="auto",
|
50 |
torch_dtype=torch.bfloat16,
|
|
|
51 |
trust_remote_code=True,
|
52 |
-
|
53 |
)
|
54 |
-
print("
|
55 |
|
56 |
return True
|
57 |
except Exception as e:
|
@@ -73,12 +84,12 @@ def process_handwriting(image):
|
|
73 |
# Convert to RGB if needed
|
74 |
if image.mode != "RGB":
|
75 |
image = image.convert("RGB")
|
76 |
-
|
77 |
-
#
|
78 |
prompt = """Transcribe the Thai handwritten text from the provided image.
|
79 |
Only return the transcription in Thai language."""
|
80 |
|
81 |
-
#
|
82 |
messages = [
|
83 |
{
|
84 |
"role": "user",
|
@@ -89,12 +100,12 @@ Only return the transcription in Thai language."""
|
|
89 |
}
|
90 |
]
|
91 |
|
92 |
-
#
|
93 |
text = processor.apply_chat_template(messages, tokenize=False)
|
94 |
inputs = processor(text=text, images=image, return_tensors="pt")
|
95 |
inputs = {k: v.to(model.device) for k, v in inputs.items()}
|
96 |
|
97 |
-
#
|
98 |
with torch.no_grad():
|
99 |
outputs = model.generate(
|
100 |
**inputs,
|
@@ -103,7 +114,7 @@ Only return the transcription in Thai language."""
|
|
103 |
pad_token_id=processor.tokenizer.pad_token_id
|
104 |
)
|
105 |
|
106 |
-
#
|
107 |
transcription = processor.decode(outputs[0], skip_special_tokens=True)
|
108 |
return transcription.strip()
|
109 |
|
@@ -126,5 +137,4 @@ if load_model_and_processor():
|
|
126 |
if __name__ == "__main__":
|
127 |
demo.launch()
|
128 |
else:
|
129 |
-
print("ไม่สามารถเริ่มต้นแอปพลิเคชันได้")
|
130 |
-
|
|
|
2 |
import warnings
|
3 |
import torch
|
4 |
import gc
|
5 |
+
from transformers import AutoModelForVision2Seq, AutoProcessor, BitsAndBytesConfig
|
6 |
from PIL import Image
|
7 |
import gradio as gr
|
8 |
from huggingface_hub import login
|
9 |
|
10 |
+
# ตั้งค่าพื้นฐาน
|
11 |
warnings.filterwarnings('ignore')
|
12 |
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
13 |
|
|
|
15 |
model = None
|
16 |
processor = None
|
17 |
|
18 |
+
# เคลียร์ CUDA cache
|
19 |
if torch.cuda.is_available():
|
20 |
torch.cuda.empty_cache()
|
21 |
gc.collect()
|
|
|
34 |
print("กำลังโหลดโมเดลและ processor...")
|
35 |
|
36 |
try:
|
37 |
+
# กำหนด paths
|
38 |
+
base_model_path = "meta-llama/Llama-3.2-11B-Vision-Instruct"
|
39 |
hub_model_path = "Aekanun/thai-handwriting-llm"
|
40 |
|
41 |
+
# ตั้งค่า BitsAndBytes
|
42 |
+
bnb_config = BitsAndBytesConfig(
|
43 |
+
load_in_4bit=True,
|
44 |
+
bnb_4bit_use_double_quant=True,
|
45 |
+
bnb_4bit_quant_type="nf4",
|
46 |
+
bnb_4bit_compute_dtype=torch.bfloat16
|
47 |
+
)
|
48 |
+
|
49 |
+
# โหลด processor จาก base model
|
50 |
+
processor = AutoProcessor.from_pretrained(
|
51 |
+
base_model_path,
|
52 |
+
use_auth_token=os.environ.get('HUGGING_FACE_HUB_TOKEN')
|
53 |
)
|
54 |
|
55 |
+
# โหลดโมเดลจาก Hub
|
56 |
print("กำลังโหลดโมเดลจาก Hub...")
|
57 |
+
model = AutoModelForVision2Seq.from_pretrained(
|
58 |
hub_model_path,
|
59 |
device_map="auto",
|
60 |
torch_dtype=torch.bfloat16,
|
61 |
+
quantization_config=bnb_config,
|
62 |
trust_remote_code=True,
|
63 |
+
use_auth_token=os.environ.get('HUGGING_FACE_HUB_TOKEN')
|
64 |
)
|
65 |
+
print("โหลดโมเดลจาก Hub สำเร็จ!")
|
66 |
|
67 |
return True
|
68 |
except Exception as e:
|
|
|
84 |
# Convert to RGB if needed
|
85 |
if image.mode != "RGB":
|
86 |
image = image.convert("RGB")
|
87 |
+
|
88 |
+
# สร้าง prompt สำหรับการถอดความ
|
89 |
prompt = """Transcribe the Thai handwritten text from the provided image.
|
90 |
Only return the transcription in Thai language."""
|
91 |
|
92 |
+
# สร้าง input สำหรับโมเดล
|
93 |
messages = [
|
94 |
{
|
95 |
"role": "user",
|
|
|
100 |
}
|
101 |
]
|
102 |
|
103 |
+
# สร้าง inputs โดยตรงจาก processor
|
104 |
text = processor.apply_chat_template(messages, tokenize=False)
|
105 |
inputs = processor(text=text, images=image, return_tensors="pt")
|
106 |
inputs = {k: v.to(model.device) for k, v in inputs.items()}
|
107 |
|
108 |
+
# ทำนาย
|
109 |
with torch.no_grad():
|
110 |
outputs = model.generate(
|
111 |
**inputs,
|
|
|
114 |
pad_token_id=processor.tokenizer.pad_token_id
|
115 |
)
|
116 |
|
117 |
+
# แปลงผลลัพธ์
|
118 |
transcription = processor.decode(outputs[0], skip_special_tokens=True)
|
119 |
return transcription.strip()
|
120 |
|
|
|
137 |
if __name__ == "__main__":
|
138 |
demo.launch()
|
139 |
else:
|
140 |
+
print("ไม่สามารถเริ่มต้นแอปพลิเคชันได้")
|
|