Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -1,63 +1,28 @@
|
|
1 |
import gradio as gr
|
2 |
-
|
3 |
-
from
|
|
|
|
|
4 |
import torch
|
5 |
-
import
|
6 |
|
7 |
-
# טוענים דאטאסט
|
8 |
dataset = load_dataset("imagefolder", data_dir=".", split={"train": "train[:80%]", "test": "train[80%:]"})
|
9 |
|
10 |
-
|
|
|
11 |
checkpoint = "facebook/deit-tiny-patch16-224"
|
12 |
processor = AutoImageProcessor.from_pretrained(checkpoint)
|
|
|
|
|
13 |
model = AutoModelForImageClassification.from_pretrained(
|
14 |
checkpoint,
|
15 |
num_labels=3,
|
16 |
id2label={0: "rock", 1: "paper", 2: "scissors"},
|
17 |
-
label2id={"rock": 0, "paper": 1, "scissors": 2}
|
18 |
-
|
19 |
-
|
20 |
-
# פונקציה לעיבוד התמונות
|
21 |
-
def preprocess(examples):
|
22 |
-
images = [x.convert("RGB") for x in examples["image"]]
|
23 |
-
inputs = processor(images=images, return_tensors="pt")
|
24 |
-
inputs["labels"] = examples["label"]
|
25 |
-
return inputs
|
26 |
-
|
27 |
-
dataset = dataset.map(preprocess, batched=True)
|
28 |
-
|
29 |
-
# הגדרות אימון מהיר
|
30 |
-
training_args = TrainingArguments(
|
31 |
-
output_dir="./results",
|
32 |
-
evaluation_strategy="epoch",
|
33 |
-
save_strategy="epoch",
|
34 |
-
per_device_train_batch_size=4,
|
35 |
-
per_device_eval_batch_size=4,
|
36 |
-
num_train_epochs=2, # ✅ אפוקים מהירים: רק 2!
|
37 |
-
load_best_model_at_end=True,
|
38 |
-
logging_dir='./logs',
|
39 |
-
logging_steps=5,
|
40 |
-
)
|
41 |
-
|
42 |
-
trainer = Trainer(
|
43 |
-
model=model,
|
44 |
-
args=training_args,
|
45 |
-
train_dataset=dataset["train"],
|
46 |
-
eval_dataset=dataset["test"],
|
47 |
)
|
48 |
|
49 |
-
# אימון
|
50 |
-
|
51 |
-
|
52 |
-
# פונקציה להרצת חיזוי על תמונה חדשה
|
53 |
-
def predict(image):
|
54 |
-
inputs = processor(images=image, return_tensors="pt")
|
55 |
-
outputs = model(**inputs)
|
56 |
-
logits = outputs.logits
|
57 |
-
predicted_class_idx = logits.argmax(-1).item()
|
58 |
-
label = model.config.id2label[predicted_class_idx]
|
59 |
-
return label
|
60 |
-
|
61 |
-
# בניית אפליקציה
|
62 |
-
demo = gr.Interface(fn=predict, inputs="image", outputs="text")
|
63 |
-
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
import random
|
3 |
+
from PIL import Image
|
4 |
+
import time
|
5 |
+
from transformers import AutoImageProcessor, AutoModelForImageClassification
|
6 |
import torch
|
7 |
+
from datasets import load_dataset
|
8 |
|
9 |
+
# טוענים דאטאסט
|
10 |
dataset = load_dataset("imagefolder", data_dir=".", split={"train": "train[:80%]", "test": "train[80%:]"})
|
11 |
|
12 |
+
|
13 |
+
# טוענים מודל בסיס
|
14 |
checkpoint = "facebook/deit-tiny-patch16-224"
|
15 |
processor = AutoImageProcessor.from_pretrained(checkpoint)
|
16 |
+
|
17 |
+
# במקום להטעין מודל סופי → טוענים ואז משנים את הראש
|
18 |
model = AutoModelForImageClassification.from_pretrained(
|
19 |
checkpoint,
|
20 |
num_labels=3,
|
21 |
id2label={0: "rock", 1: "paper", 2: "scissors"},
|
22 |
+
label2id={"rock": 0, "paper": 1, "scissors": 2},
|
23 |
+
ignore_mismatched_sizes=True, # 🔥 קריטי!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
)
|
25 |
|
26 |
+
# פונקציית אימון קטנה (נוסיף עוד מעט)
|
27 |
+
# פונקציית משחק
|
28 |
+
# אפליקציה Gradio
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|