Robotics
PyTorch
Cosmos
xperience10m_task_baseline_suite
embodied-ai
multimodal
xperience-10m
baseline
evaluation
qwen3-omni
Instructions to use cy0307/ropedia-xperience-10m-task-baselines with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use cy0307/ropedia-xperience-10m-task-baselines with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 3,932 Bytes
596ac86 16a39bb d9e465e 16a39bb 01f57c3 16a39bb d9e465e 596ac86 d9e465e 596ac86 01f57c3 596ac86 d9e465e 16a39bb d9e465e 16a39bb d9e465e 16a39bb d9e465e 16a39bb d9e465e 16a39bb d9e465e 16a39bb d9e465e 01f57c3 d9e465e 16a39bb d9e465e 596ac86 d9e465e 16a39bb d9e465e 16a39bb d9e465e 596ac86 d9e465e 596ac86 d9e465e 596ac86 d9e465e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 | #!/usr/bin/env python3
"""Build the three foundation-pipeline slide diagrams.
The public foundation-direction visuals intentionally use the direction-slide
sources provided by the project owner, not generated concept art. Clean slide
PNGs are used directly when available; older photo sources are restored only as
fallbacks. The output asset names stay stable for the website, README, and HF
mirrors.
"""
from __future__ import annotations
from dataclasses import dataclass
from pathlib import Path
from PIL import Image, ImageEnhance, ImageFilter, ImageOps
ROOT = Path(__file__).resolve().parents[1]
OUT_DIR = ROOT / "docs/assets/foundation-pipelines"
SOURCE_DIR = OUT_DIR / "source-photos"
SOURCE_SLIDE_DIR = OUT_DIR / "source-slides"
TARGET_WIDTH = 2560
@dataclass(frozen=True)
class PhotoAsset:
source: str
slide_source: str | None
output: str
title: str
brightness: float
contrast: float
color: float
sharpness: float
PHOTOS = [
PhotoAsset(
source="spatial-intelligence-source.jpg",
slide_source="spatial-intelligence-slide.png",
output="spatial-intelligence-pipeline.png",
title="Spatial intelligence slide diagram",
brightness=1.04,
contrast=1.18,
color=1.08,
sharpness=1.36,
),
PhotoAsset(
source="human-video-world-model-source.jpg",
slide_source="human-video-world-model-slide.png",
output="human-video-world-model-pipeline.png",
title="Human-video world-model slide diagram",
brightness=1.05,
contrast=1.20,
color=1.08,
sharpness=1.34,
),
PhotoAsset(
source="vision-language-action-source.jpg",
slide_source="vision-language-action-slide.png",
output="vision-language-action-pipeline.png",
title="Vision-language-action slide diagram",
brightness=1.06,
contrast=1.18,
color=1.09,
sharpness=1.34,
),
]
def enhance(asset: PhotoAsset) -> Image.Image:
if asset.slide_source:
slide_path = SOURCE_SLIDE_DIR / asset.slide_source
if slide_path.is_file():
img = Image.open(slide_path).convert("RGB")
img = ImageOps.exif_transpose(img)
if img.width != TARGET_WIDTH:
scale = TARGET_WIDTH / img.width
target_size = (TARGET_WIDTH, round(img.height * scale))
img = img.resize(target_size, Image.Resampling.LANCZOS)
return img
source_path = SOURCE_DIR / asset.source
if not source_path.is_file():
raise FileNotFoundError(f"Missing source slide/photo for {asset.output}: {source_path}")
img = Image.open(source_path).convert("RGB")
img = ImageOps.exif_transpose(img)
img = ImageOps.autocontrast(img, cutoff=0.35)
img = ImageEnhance.Brightness(img).enhance(asset.brightness)
img = ImageEnhance.Contrast(img).enhance(asset.contrast)
img = ImageEnhance.Color(img).enhance(asset.color)
if img.width != TARGET_WIDTH:
scale = TARGET_WIDTH / img.width
target_size = (TARGET_WIDTH, round(img.height * scale))
img = img.resize(target_size, Image.Resampling.LANCZOS)
# Gentle deblur/edge recovery without hallucinating slide text.
smooth = img.filter(ImageFilter.GaussianBlur(radius=0.55))
img = Image.blend(smooth, img, 0.68)
img = ImageEnhance.Sharpness(img).enhance(asset.sharpness)
img = img.filter(ImageFilter.UnsharpMask(radius=1.15, percent=135, threshold=3))
return img
def main() -> int:
OUT_DIR.mkdir(parents=True, exist_ok=True)
for asset in PHOTOS:
output = OUT_DIR / asset.output
image = enhance(asset)
image.save(output, optimize=True, compress_level=9)
print(f"{asset.title}: {output} {image.width}x{image.height} {output.stat().st_size} bytes")
return 0
if __name__ == "__main__":
raise SystemExit(main())
|