Scene_Foundry_Demo / EVALUATION_RESULTS.md
Chunteng's picture
Initial commit (Fresh Start)
a03fc9e

A newer version of the Gradio SDK is available: 6.4.0

Upgrade

Evaluation Results

Summary Statistics

  • Total Test Cases: 20
  • Average Score: 90.8%
  • Score Range: 70.0% - 100.0%
  • Average Generation Time: 271.7s (~4.5 minutes)
  • Time Range: 259.0s - 289.7s

Score Distribution

Category Count Percentage
🏆 Excellent (≥90%) 13 65.0%
✓ Good (80-89%) 3 15.0%
⚠️ Acceptable (70-79%) 4 20.0%
✗ Poor (<70%) 0 0.0%

Component-wise Average Scores (Before Weighting)

Component Average Score Weight
Room Counts 75.0% 30%
Room Presence 97.4% 20%
Adjacency 97.5% 30%
Constraints 97.7% 20%

Complete Test Results

Test ID Score Gen Time Room Counts Room Presence Adjacency Constraints
🏆 two_bedroom_apt_01 100.0% 262.2s 100.0% 100.0% 100.0% 100.0%
🏆 open_plan_loft_01 100.0% 274.7s 100.0% 100.0% 100.0% 100.0%
🏆 family_home_01 100.0% 276.7s 100.0% 100.0% 100.0% 100.0%
🏆 master_suite_home_01 100.0% 265.3s 100.0% 100.0% 100.0% 100.0%
🏆 four_bedroom_house_01 100.0% 270.8s 100.0% 100.0% 100.0% 100.0%
🏆 guest_suite_home_01 100.0% 282.2s 100.0% 100.0% 100.0% 100.0%
🏆 dual_master_suite_01 100.0% 269.5s 100.0% 100.0% 100.0% 100.0%
🏆 balcony_apartment_01 100.0% 285.2s 100.0% 100.0% 100.0% 100.0%
🏆 multigenerational_home_01 100.0% 289.7s 100.0% 100.0% 100.0% 100.0%
🏆 single_floor_accessible_01 95.0% 262.5s 83.3% 100.0% 100.0% 100.0%
🏆 three_bed_family_townhouse_01 95.0% 268.3s 83.3% 100.0% 100.0% 100.0%
🏆 entertainment_home_01 90.8% 274.9s 83.3% 100.0% 100.0% 79.2%
🏆 basic_studio_01 90.0% 259.0s 66.7% 100.0% 100.0% 100.0%
✓ luxury_penthouse_01 87.5% 270.8s 83.3% 87.5% 100.0% 75.0%
✓ one_bedroom_apt_01 85.0% 270.0s 50.0% 100.0% 100.0% 100.0%
✓ separated_zones_01 85.0% 285.7s 100.0% 100.0% 50.0% 100.0%
⚠️ home_office_layout_01 73.5% 267.8s 25.0% 80.0% 100.0% 100.0%
⚠️ work_from_home_layout_01 73.5% 263.0s 25.0% 80.0% 100.0% 100.0%
⚠️ compact_efficiency_01 70.0% 269.7s 0.0% 100.0% 100.0% 100.0%
⚠️ student_apartment_01 70.0% 266.0s 0.0% 100.0% 100.0% 100.0%

Detailed Results

🏆 two_bedroom_apt_01 - 100.0%

Description: Standard 2BR family apartment

Prompt: Create a spacious two-bedroom apartment (100m² total) with living room (35m²), kitchen (18m²), dining room (15m²), two bedrooms (25m² and 20m²), and two bathrooms. Kitchen should connect to dining room.

Generation Time: 262.2s (~4.4 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 open_plan_loft_01 - 100.0%

Description: Modern loft with open concept

Prompt: Design a modern open-plan loft (85m² total) with spacious combined living/dining area (40m²), open kitchen (18m²), one bedroom (22m²) with ensuite bathroom (5m²). Contemporary style.

Generation Time: 274.7s (~4.6 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 family_home_01 - 100.0%

Description: Family home with multiple bedrooms

Prompt: Create a large family home (140m² total) with living room (40m²), dining room (20m²), kitchen (20m²), three bedrooms (30m², 25m², 20m²), and two bathrooms (8m² each). Kitchen adjacent to dining room.

Generation Time: 276.7s (~4.6 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 master_suite_home_01 - 100.0%

Description: Home with master suite

Prompt: Design a comfortable home (110m² total) with living room (35m²), kitchen (18m²), spacious master bedroom (35m²) with ensuite bathroom (8m²), and one additional bedroom (20m²). Master bedroom must be adjacent to its bathroom.

Generation Time: 265.3s (~4.4 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 four_bedroom_house_01 - 100.0%

Description: Large family house

Prompt: Create a large family house (180m² total) with spacious living room (45m²), dining room (25m²), kitchen (22m²), four bedrooms (35m², 30m², 25m², 20m²), and three bathrooms (10m², 8m², 6m²). Living room and dining room should be adjacent. Kitchen adjacent to dining room. Traditional family layout.

Generation Time: 270.8s (~4.5 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 guest_suite_home_01 - 100.0%

Description: Home with separated guest suite

Prompt: Create a luxury home (150m² total) with main living areas (living room 40m², kitchen 20m², dining room 20m²), spacious master bedroom (35m²) with bathroom (10m²), and a separate guest suite with bedroom (25m²) and its own bathroom (8m²). Master and guest bedrooms each adjacent to their respective bathrooms.

Generation Time: 282.2s (~4.7 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 dual_master_suite_01 - 100.0%

Description: Dual master suite configuration

Prompt: Design a luxury apartment (160m² total) with two master bedrooms (40m² and 35m²), each with its own ensuite bathroom (12m² and 10m²). Include spacious living room (45m²), elegant dining room (22m²), and gourmet kitchen (25m²). High-end finishes throughout.

Generation Time: 269.5s (~4.5 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 balcony_apartment_01 - 100.0%

Description: Apartment with outdoor space

Prompt: Design a modern apartment (80m² total) with living room (32m²), kitchen (14m²), bedroom (22m²), bathroom (6m²), and a balcony (6m²) accessible from living room. Indoor-outdoor living style.

Generation Time: 285.2s (~4.8 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 multigenerational_home_01 - 100.0%

Description: Tests complex multi-suite layouts

Prompt: Design a spacious multigenerational home (200m² total) with main living room (45m²), kitchen (22m²), dining room (20m²), master bedroom (35m²) with bathroom (10m²), two additional bedrooms (25m² each), shared bathroom (8m²), and a separate accessible suite for elderly parents with bedroom (30m²) and bathroom (10m²) adjacent to each other. Three-generation family layout.

Generation Time: 289.7s (~4.8 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 single_floor_accessible_01 - 95.0%

Description: Tests accessibility-friendly layouts

Prompt: Design an accessible barrier-free single-floor home (130m² total) with open living (38m²) and dining area (18m²), kitchen (20m²) adjacent to dining, spacious master bedroom (32m²) with accessible ensuite bathroom (10m²), one additional bedroom (25m²), and another accessible bathroom (8m²). Wide hallways and easy navigation for wheelchair access.

Generation Time: 262.5s (~4.4 minutes)

Component Scores:

  • Room Counts: 83.3%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 three_bed_family_townhouse_01 - 95.0%

Description: Tests standard family townhouse layout

Prompt: Design a three-bedroom family townhouse (145m² total) with living room (40m²), separate dining room (18m²), kitchen (20m²) near dining room, three bedrooms (30m², 25m², 22m²), and two bathrooms (10m² and 7m²). Traditional townhouse layout for growing family.

Generation Time: 268.3s (~4.5 minutes)

Component Scores:

  • Room Counts: 83.3%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

🏆 entertainment_home_01 - 90.8%

Description: Tests entertainment-focused layouts

Prompt: Create an entertainer's home (150m² total) with large open living (40m²) and dining area (22m²) open to each other, spacious gourmet kitchen (25m²) adjacent to dining, dedicated media/theater room (20m²), two bedrooms (25m² each), and two bathrooms (8m² each). Perfect for hosting guests.

Generation Time: 274.9s (~4.6 minutes)

Component Scores:

  • Room Counts: 83.3%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 79.2%

🏆 basic_studio_01 - 90.0%

Description: Most basic layout - should score very high

Prompt: Create a compact studio apartment (45m² total) with cozy living area (20m²), small kitchen (10m²), and one bathroom (5m²). Minimalist layout.

Generation Time: 259.0s (~4.3 minutes)

Component Scores:

  • Room Counts: 66.7%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

✓ luxury_penthouse_01 - 87.5%

Description: Tests luxury/high-end layouts with special rooms

Prompt: Create an ultra-luxury penthouse (220m² total) with expansive living room (60m²), formal dining room (28m²), gourmet kitchen (30m²) adjacent to dining, spacious master bedroom (45m²) with walk-in closet and luxurious ensuite bathroom (15m²), elegant study room (18m²), and comfortable guest bedroom (28m²) with guest bathroom (10m²). Premium materials and high ceilings throughout.

Generation Time: 270.8s (~4.5 minutes)

Component Scores:

  • Room Counts: 83.3%
  • Room Presence: 87.5%
  • Adjacency: 100.0%
  • Constraints: 75.0%

✓ one_bedroom_apt_01 - 85.0%

Description: Classic 1BR layout

Prompt: Design a comfortable one-bedroom apartment (70m² total) with living room (30m²), kitchen (15m²), bedroom (20m²), and bathroom (5m²). Open floor plan with living room adjacent to kitchen.

Generation Time: 270.0s (~4.5 minutes)

Component Scores:

  • Room Counts: 50.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

✓ separated_zones_01 - 85.0%

Description: Functional separation of spaces

Prompt: Design a modern apartment (95m² total) with separate day and night zones: day zone has living room (35m²) and kitchen (15m²) adjacent to each other, night zone has two bedrooms (25m² and 20m²) and bathroom (8m²). Bedrooms should NOT be adjacent to kitchen. Contemporary functional layout.

Generation Time: 285.7s (~4.8 minutes)

Component Scores:

  • Room Counts: 100.0%
  • Room Presence: 100.0%
  • Adjacency: 50.0%
  • Constraints: 100.0%

⚠️ home_office_layout_01 - 73.5%

Description: Work-from-home layout

Prompt: Create a work-from-home apartment (90m² total) with living room (30m²), kitchen (15m²), bedroom (25m²), bathroom (6m²), and a dedicated home office room (14m²) for remote work. Modern professional layout.

Generation Time: 267.8s (~4.5 minutes)

Component Scores:

  • Room Counts: 25.0%
  • Room Presence: 80.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

⚠️ work_from_home_layout_01 - 73.5%

Description: Tests home office integration and zoning

Prompt: Create a professional work-from-home apartment (92m² total) with living room (30m²), kitchen (15m²), bedroom (24m²), bathroom (7m²), and dedicated home office (16m²) with good sound separation from living areas and bedroom. Quiet workspace prioritized.

Generation Time: 263.0s (~4.4 minutes)

Component Scores:

  • Room Counts: 25.0%
  • Room Presence: 80.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

⚠️ compact_efficiency_01 - 70.0%

Description: Tests compact/minimal space layouts

Prompt: Create a highly efficient compact apartment (48m² total) with small living room (18m²), compact galley kitchen (10m²) adjacent to living room, cozy bedroom (15m²), and efficient bathroom (5m²). Maximize every square meter.

Generation Time: 269.7s (~4.5 minutes)

Component Scores:

  • Room Counts: 0.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%

⚠️ student_apartment_01 - 70.0%

Description: Tests minimal/basic student housing

Prompt: Design a budget-friendly student apartment (42m² total) with compact living area (15m² max), small kitchenette (10m² max), small bedroom (15m²), and basic bathroom (5m²). Affordable and functional layout for students.

Generation Time: 266.0s (~4.4 minutes)

Component Scores:

  • Room Counts: 0.0%
  • Room Presence: 100.0%
  • Adjacency: 100.0%
  • Constraints: 100.0%