Update app.py
Browse files
app.py
CHANGED
@@ -1,70 +1,69 @@
|
|
1 |
-
import
|
|
|
2 |
from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
|
3 |
import torch
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
# Load model dan tokenizer
|
7 |
model_name = "facebook/m2m100_418M"
|
8 |
try:
|
9 |
tokenizer = M2M100Tokenizer.from_pretrained(model_name)
|
10 |
model = M2M100ForConditionalGeneration.from_pretrained(model_name)
|
|
|
11 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
12 |
model.to(device)
|
13 |
model.eval()
|
14 |
except Exception as e:
|
15 |
raise Exception(f"Gagal memuat model: {str(e)}")
|
16 |
|
17 |
-
#
|
18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
try:
|
20 |
-
#
|
21 |
-
if not
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
source_lang = detected_lang
|
27 |
-
except LangDetectException:
|
28 |
-
return "Gagal mendeteksi bahasa. Harap masukkan kode bahasa sumber.", None
|
29 |
-
else:
|
30 |
-
if source_lang not in tokenizer.supported_languages:
|
31 |
-
return f"Kode bahasa '{source_lang}' tidak didukung.", None
|
32 |
|
33 |
# Set bahasa sumber
|
34 |
-
tokenizer.src_lang = source_lang
|
|
|
|
|
|
|
35 |
|
36 |
-
#
|
37 |
-
encoded = tokenizer(text, return_tensors="pt", padding=True, truncation=True).to(device)
|
38 |
generated_tokens = model.generate(
|
39 |
**encoded,
|
40 |
forced_bos_token_id=tokenizer.get_lang_id("en")
|
41 |
)
|
|
|
|
|
42 |
translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
|
43 |
|
44 |
-
return translated_text
|
45 |
|
46 |
except Exception as e:
|
47 |
-
|
48 |
-
|
49 |
-
# Buat antarmuka Gradio
|
50 |
-
iface = gr.Interface(
|
51 |
-
fn=translate_text,
|
52 |
-
inputs=[
|
53 |
-
gr.Textbox(label="Teks untuk Diterjemahkan"),
|
54 |
-
gr.Dropdown(
|
55 |
-
choices=["id", "fr", "es", "de", "ja", ""], # Tambahkan lebih banyak kode bahasa jika perlu
|
56 |
-
label="Bahasa Sumber (kosongkan untuk autodeteksi)",
|
57 |
-
value=""
|
58 |
-
)
|
59 |
-
],
|
60 |
-
outputs=[
|
61 |
-
gr.Textbox(label="Terjemahan ke Bahasa Inggris"),
|
62 |
-
gr.Textbox(label="Bahasa Sumber Terdeteksi")
|
63 |
-
],
|
64 |
-
title="M2M100 Translation to English",
|
65 |
-
description="Masukkan teks untuk diterjemahkan ke bahasa Inggris. Biarkan bahasa sumber kosong untuk autodeteksi."
|
66 |
-
)
|
67 |
|
68 |
-
|
69 |
-
|
70 |
-
|
|
|
|
|
|
|
|
1 |
+
from fastapi import FastAPI, HTTPException
|
2 |
+
from pydantic import BaseModel
|
3 |
from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
|
4 |
import torch
|
5 |
+
|
6 |
+
# Inisialisasi FastAPI
|
7 |
+
app = FastAPI(
|
8 |
+
title="M2M100 Translation API",
|
9 |
+
description="API untuk menerjemahkan teks dari berbagai bahasa ke bahasa Inggris menggunakan facebook/m2m100_418M.",
|
10 |
+
version="1.0.0"
|
11 |
+
)
|
12 |
|
13 |
# Load model dan tokenizer
|
14 |
model_name = "facebook/m2m100_418M"
|
15 |
try:
|
16 |
tokenizer = M2M100Tokenizer.from_pretrained(model_name)
|
17 |
model = M2M100ForConditionalGeneration.from_pretrained(model_name)
|
18 |
+
# Pindahkan model ke GPU jika tersedia
|
19 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
20 |
model.to(device)
|
21 |
model.eval()
|
22 |
except Exception as e:
|
23 |
raise Exception(f"Gagal memuat model: {str(e)}")
|
24 |
|
25 |
+
# Definisikan request body
|
26 |
+
class TranslationRequest(BaseModel):
|
27 |
+
text: str
|
28 |
+
source_lang: str # Kode bahasa sumber, misalnya "id" untuk Indonesia, "fr" untuk Prancis
|
29 |
+
|
30 |
+
# Definisikan response body
|
31 |
+
class TranslationResponse(BaseModel):
|
32 |
+
translated_text: str
|
33 |
+
|
34 |
+
@app.post("/translate", response_model=TranslationResponse)
|
35 |
+
async def translate_text(request: TranslationRequest):
|
36 |
try:
|
37 |
+
# Validasi kode bahasa
|
38 |
+
if request.source_lang not in tokenizer.supported_languages:
|
39 |
+
raise HTTPException(
|
40 |
+
status_code=400,
|
41 |
+
detail=f"Kode bahasa '{request.source_lang}' tidak didukung. Gunakan kode seperti 'id', 'fr', 'es', dll."
|
42 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
|
44 |
# Set bahasa sumber
|
45 |
+
tokenizer.src_lang = request.source_lang
|
46 |
+
|
47 |
+
# Encode input teks
|
48 |
+
encoded = tokenizer(request.text, return_tensors="pt", padding=True, truncation=True).to(device)
|
49 |
|
50 |
+
# Generate terjemahan (target: English, "en")
|
|
|
51 |
generated_tokens = model.generate(
|
52 |
**encoded,
|
53 |
forced_bos_token_id=tokenizer.get_lang_id("en")
|
54 |
)
|
55 |
+
|
56 |
+
# Decode hasil terjemahan
|
57 |
translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
|
58 |
|
59 |
+
return TranslationResponse(translated_text=translated_text)
|
60 |
|
61 |
except Exception as e:
|
62 |
+
raise HTTPException(status_code=500, detail=f"Error saat menerjemahkan: {str(e)}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
63 |
|
64 |
+
@app.get("/")
|
65 |
+
async def root():
|
66 |
+
return {
|
67 |
+
"message": "Selamat datang di M2M100 Translation API! Gunakan endpoint /translate untuk menerjemahkan teks ke bahasa Inggris.",
|
68 |
+
"docs": "/docs"
|
69 |
+
}
|