Update app.py
Browse files
app.py
CHANGED
|
@@ -12,55 +12,16 @@ math_samples = []
|
|
| 12 |
loading_status = {"loaded": False, "error": None}
|
| 13 |
|
| 14 |
def load_sample_problems():
|
| 15 |
-
"""Load sample problems
|
| 16 |
global math_samples, loading_status
|
| 17 |
-
samples = []
|
| 18 |
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
if i >= 30:
|
| 24 |
-
break
|
| 25 |
-
samples.append(item["question"])
|
| 26 |
-
|
| 27 |
-
# Fineweb-edu
|
| 28 |
-
fw = load_dataset("HuggingFaceFW/fineweb-edu", name="sample-10BT", split="train", streaming=True)
|
| 29 |
-
fw_count = 0
|
| 30 |
-
for item in fw:
|
| 31 |
-
if fw_count >= 15:
|
| 32 |
-
break
|
| 33 |
-
text_lower = item['text'].lower()
|
| 34 |
-
if any(w in text_lower for w in ['math', 'calculate', 'solve', 'equation']):
|
| 35 |
-
q = item['text'][:120].strip()
|
| 36 |
-
if len(q) > 20:
|
| 37 |
-
samples.append(q + "...")
|
| 38 |
-
fw_count += 1
|
| 39 |
-
|
| 40 |
-
# Ultrachat
|
| 41 |
-
ds = load_dataset("HuggingFaceH4/ultrachat_200k", streaming=True, split="train_sft")
|
| 42 |
-
ds_count = 0
|
| 43 |
-
for item in ds:
|
| 44 |
-
if ds_count >= 15:
|
| 45 |
-
break
|
| 46 |
-
if len(item['messages']) > 0:
|
| 47 |
-
content = item['messages'][0]['content'].lower()
|
| 48 |
-
if any(w in content for w in ['math', 'solve', 'equation', 'calculate']):
|
| 49 |
-
samples.append(item['messages'][0]['content'])
|
| 50 |
-
ds_count += 1
|
| 51 |
-
|
| 52 |
-
math_samples = samples if samples else get_fallback_samples()
|
| 53 |
-
loading_status["loaded"] = True
|
| 54 |
-
print(f"✅ Loaded {len(math_samples)} samples")
|
| 55 |
-
|
| 56 |
-
except Exception as e:
|
| 57 |
-
print(f"⚠️ Dataset error: {e}")
|
| 58 |
-
math_samples = get_fallback_samples()
|
| 59 |
-
loading_status["error"] = str(e)
|
| 60 |
-
loading_status["loaded"] = True
|
| 61 |
|
| 62 |
def get_fallback_samples():
|
| 63 |
-
"""Extended fallback problems"""
|
| 64 |
return [
|
| 65 |
"Find the derivative of f(x) = 3x² + 2x - 1",
|
| 66 |
"A triangle has sides 5, 12, and 13. What is its area?",
|
|
@@ -77,6 +38,16 @@ def get_fallback_samples():
|
|
| 77 |
"Solve x² - 4x + 4 = 0",
|
| 78 |
"Find tan(π/4)",
|
| 79 |
"What is 15% of 240?",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
]
|
| 81 |
|
| 82 |
# Start background loading
|
|
|
|
| 12 |
loading_status = {"loaded": False, "error": None}
|
| 13 |
|
| 14 |
def load_sample_problems():
|
| 15 |
+
"""Load sample problems - using fallback only to avoid storage limits"""
|
| 16 |
global math_samples, loading_status
|
|
|
|
| 17 |
|
| 18 |
+
# Skip dataset loading to avoid 50GB storage limit
|
| 19 |
+
math_samples = get_fallback_samples()
|
| 20 |
+
loading_status["loaded"] = True
|
| 21 |
+
print(f"✅ Loaded {len(math_samples)} fallback samples")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
def get_fallback_samples():
|
| 24 |
+
"""Extended fallback problems covering diverse topics"""
|
| 25 |
return [
|
| 26 |
"Find the derivative of f(x) = 3x² + 2x - 1",
|
| 27 |
"A triangle has sides 5, 12, and 13. What is its area?",
|
|
|
|
| 38 |
"Solve x² - 4x + 4 = 0",
|
| 39 |
"Find tan(π/4)",
|
| 40 |
"What is 15% of 240?",
|
| 41 |
+
"Evaluate ∫x³ dx from 1 to 3",
|
| 42 |
+
"Find the slope of the line through (2,3) and (5,9)",
|
| 43 |
+
"Convert 45° to radians",
|
| 44 |
+
"Solve |2x - 5| = 7",
|
| 45 |
+
"Find the vertex of y = x² - 6x + 8",
|
| 46 |
+
"Calculate the discriminant of 2x² + 3x - 5 = 0",
|
| 47 |
+
"What is the standard deviation of [2, 4, 6, 8, 10]?",
|
| 48 |
+
"Simplify √(72)",
|
| 49 |
+
"Find cos(60°)",
|
| 50 |
+
"Solve the inequality 3x - 7 > 11",
|
| 51 |
]
|
| 52 |
|
| 53 |
# Start background loading
|