app.py

"""Sapiens2 body-part segmentation Gradio Space.

Image → 29-class semantic segmentation. Renders an AnnotatedImage so the user
can hover over each predicted body part to highlight it with its class name.
"""

import sys
import os
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

import tempfile
from typing import List, Tuple

import cv2
import gradio as gr
import numpy as np
import spaces
import torch
import torch.nn.functional as F
from PIL import Image

from huggingface_hub import hf_hub_download
from sapiens.dense.models import SegEstimator, init_model  # registers SegEstimator
from sapiens.dense.src.datasets.seg.seg_utils import DOME_CLASSES_29
_ = SegEstimator


# -----------------------------------------------------------------------------
# Config

ASSETS_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "assets")
CONFIGS_DIR = os.path.join(ASSETS_DIR, "configs")

SEG_MODELS = {
    "0.4B": {
        "repo": "facebook/sapiens2-seg-0.4b",
        "filename": "sapiens2_0.4b_seg.safetensors",
        "config": os.path.join(CONFIGS_DIR, "sapiens2_0.4b_seg_shutterstock_goliath-1024x768.py"),
    },
    "1B": {
        "repo": "facebook/sapiens2-seg-1b",
        "filename": "sapiens2_1b_seg.safetensors",
        "config": os.path.join(CONFIGS_DIR, "sapiens2_1b_seg_shutterstock_goliath-1024x768.py"),
    },
}
DEFAULT_SIZE = "1B"

DEVICE = "cuda" if torch.cuda.is_available() else "cpu"

# All 29 named classes (Background is class 0).
_CLASS_LABELS = {cid: meta["name"].replace("_", " ") for cid, meta in DOME_CLASSES_29.items()}
_CLASS_COLORS_RGB = {cid: meta["color"] for cid, meta in DOME_CLASSES_29.items()}
_CLASS_COLORS_HEX = {
    _CLASS_LABELS[cid]: "#{:02x}{:02x}{:02x}".format(*meta["color"])
    for cid, meta in DOME_CLASSES_29.items()
}


# -----------------------------------------------------------------------------
# Model cache

_seg_model_cache: dict = {}


def _get_seg_model(size: str):
    if size not in _seg_model_cache:
        spec = SEG_MODELS[size]
        ckpt = hf_hub_download(repo_id=spec["repo"], filename=spec["filename"])
        model = init_model(spec["config"], ckpt, device=DEVICE)
        _seg_model_cache[size] = model
    return _seg_model_cache[size]


print("[startup] pre-loading all seg sizes ...")
for _size in SEG_MODELS:
    _get_seg_model(_size)
print("[startup] ready.")


# -----------------------------------------------------------------------------
# Inference (mirrors sapiens/dense/tools/vis/vis_seg.py)

def _segment(image_bgr: np.ndarray, model) -> np.ndarray:
    h0, w0 = image_bgr.shape[:2]
    data = model.pipeline(dict(img=image_bgr))         # resize + pad
    data = model.data_preprocessor(data)               # normalize + batch
    inputs = data["inputs"]                            # already (B, 3, H, W)

    with torch.no_grad():
        logits = model(inputs)                         # (1, 29, H, W)

    logits = F.interpolate(logits, size=(h0, w0), mode="bilinear", align_corners=False)
    return logits.argmax(dim=1).squeeze(0).cpu().numpy().astype(np.int32)  # (H, W)


def _label_map_to_annotations(label_map: np.ndarray) -> List[Tuple[np.ndarray, str]]:
    """Convert (H, W) class-id map → AnnotatedImage's [(bool_mask, label), ...] list.

    Includes all 29 named classes (Background as well) so the legend is complete.
    """
    annotations: List[Tuple[np.ndarray, str]] = []
    for cid in np.unique(label_map):
        cid = int(cid)
        if cid not in _CLASS_LABELS:
            continue
        mask = (label_map == cid)
        if not mask.any():
            continue
        annotations.append((mask, _CLASS_LABELS[cid]))
    return annotations


def _label_map_to_overlay(image_bgr: np.ndarray, label_map: np.ndarray, opacity: float) -> np.ndarray:
    """Static color overlay (BGR) — colors from the DOME palette, alpha-blended."""
    palette = np.zeros((256, 3), dtype=np.uint8)
    for cid, rgb in _CLASS_COLORS_RGB.items():
        palette[cid] = rgb[::-1]  # RGB → BGR for cv2
    color_mask = palette[label_map]
    return cv2.addWeighted(image_bgr, 1.0 - opacity, color_mask, opacity, 0)


# -----------------------------------------------------------------------------
# Gradio handler

@spaces.GPU(duration=120)
def predict(image: Image.Image, size: str):
    if image is None:
        return None, None, None

    image_pil = image.convert("RGB")
    image_rgb = np.array(image_pil)
    image_bgr = cv2.cvtColor(image_rgb, cv2.COLOR_RGB2BGR)

    model = _get_seg_model(size)
    label_map = _segment(image_bgr, model)               # (H, W)

    annotations = _label_map_to_annotations(label_map)
    annotated = (image_pil, annotations)

    overlay_bgr = _label_map_to_overlay(image_bgr, label_map, 0.5)
    overlay_rgb = cv2.cvtColor(overlay_bgr, cv2.COLOR_BGR2RGB)

    npy_path = tempfile.NamedTemporaryFile(delete=False, suffix=".npy").name
    np.save(npy_path, label_map.astype(np.uint8))

    return annotated, Image.fromarray(overlay_rgb), npy_path


# -----------------------------------------------------------------------------
# UI

EXAMPLES = sorted(
    os.path.join(ASSETS_DIR, "images", n)
    for n in os.listdir(os.path.join(ASSETS_DIR, "images"))
    if n.lower().endswith((".jpg", ".jpeg", ".png"))
)

CUSTOM_CSS = """
:root, body, .gradio-container, button, input, select, textarea,
.gradio-container *:not(code):not(pre) {
    font-family: "Helvetica Neue", Helvetica, Arial, sans-serif !important;
    -webkit-font-smoothing: antialiased;
    -moz-osx-font-smoothing: grayscale;
}

#title { text-align: center; font-size: 44px; font-weight: 700;
         letter-spacing: -0.01em; margin: 28px 0 4px;
         background: linear-gradient(90deg, #1d4ed8 0%, #6d28d9 50%, #be185d 100%);
         -webkit-background-clip: text; -webkit-text-fill-color: transparent;
         background-clip: text; }
#subtitle { text-align: center; font-size: 12px; color: #64748b;
            letter-spacing: 0.18em; margin: 0 0 14px; text-transform: uppercase;
            font-weight: 500; }
#tagline { text-align: center; font-size: 15px; color: #475569;
           max-width: 680px; margin: 4px auto 22px; line-height: 1.55;
           font-weight: 400; }
#badges { display: flex; justify-content: center; flex-wrap: wrap;
          gap: 8px; margin: 0 0 32px; }
.pill { display: inline-flex; align-items: center; gap: 6px;
        padding: 7px 14px; border-radius: 999px;
        background: #f1f5f9; color: #0f172a !important;
        font-size: 13px; font-weight: 500; letter-spacing: 0.01em;
        text-decoration: none !important; border: 1px solid #e2e8f0;
        transition: background 150ms ease, transform 150ms ease, border-color 150ms ease; }
.pill:hover { background: #0f172a; color: #f8fafc !important;
              border-color: #0f172a; transform: translateY(-1px); }
.pill svg { width: 14px; height: 14px; }

/* AnnotatedImage hover behavior:
   - Default: every mask at 55% (RGB shows through)
   - Hover legend item OR mask region → that mask pops to 75% with a coloured glow,
     every OTHER mask fades down to 10% (dim, not invisible).
   Gradio toggles `.active` on the hovered mask and `.inactive` on the rest, so we
   just over-style those. */
#seg-out .mask { opacity: 0.55 !important;
                 transition: opacity 200ms ease, filter 200ms ease; }
#seg-out .mask.active { opacity: 0.78 !important;
                        filter: brightness(1.18)
                                drop-shadow(0 0 10px rgba(255,255,255,0.55))
                                drop-shadow(0 4px 16px rgba(0,0,0,0.40)) !important; }
#seg-out .mask.inactive { opacity: 0.10 !important; }
#seg-out .legend-item { cursor: pointer; }

/* Legend: vertical column on the right of the image instead of horizontal below. */
#seg-out .container { flex-direction: row !important; align-items: stretch !important; gap: 12px; }
#seg-out .image-container { flex: 1 1 auto; min-width: 0; }
#seg-out .legend {
    flex: 0 0 180px; flex-direction: column !important; flex-wrap: nowrap !important;
    align-items: stretch; justify-content: flex-start;
    gap: 4px; padding: 8px 4px;
    max-height: 640px; overflow-y: auto;
    border-left: 1px solid var(--border-color-primary, #e2e8f0);
}
#seg-out .legend-item { font-size: 12px; font-weight: 500;
                        padding: 4px 10px; width: 100%; cursor: pointer;
                        border-radius: 6px;
                        transition: transform 140ms ease, background 140ms ease; }
#seg-out .legend-item:hover { transform: translateX(2px); }
"""

HEADER_HTML = """
<div id="title">Sapiens2: Segmentation</div>
<div id="subtitle">ICLR 2026</div>
<div id="badges">
  <a class="pill" href="https://github.com/facebookresearch/sapiens2" target="_blank" rel="noopener">
    <svg viewBox="0 0 24 24" fill="currentColor"><path d="M12 .3a12 12 0 0 0-3.8 23.4c.6.1.8-.3.8-.6v-2c-3.3.7-4-1.6-4-1.6-.6-1.4-1.4-1.8-1.4-1.8-1.1-.7.1-.7.1-.7 1.3.1 2 1.3 2 1.3 1.1 1.9 3 1.4 3.7 1 .1-.8.4-1.4.8-1.7-2.7-.3-5.5-1.3-5.5-5.9 0-1.3.5-2.4 1.3-3.2-.1-.4-.6-1.6.1-3.2 0 0 1-.3 3.3 1.2a11.5 11.5 0 0 1 6 0c2.3-1.5 3.3-1.2 3.3-1.2.7 1.6.2 2.8.1 3.2.8.8 1.3 1.9 1.3 3.2 0 4.6-2.8 5.6-5.5 5.9.4.4.8 1.1.8 2.2v3.3c0 .3.2.7.8.6A12 12 0 0 0 12 .3"/></svg>
    Code
  </a>
  <a class="pill" href="https://huggingface.co/facebook/sapiens2" target="_blank" rel="noopener">
    🤗 Models
  </a>
  <a class="pill" href="https://arxiv.org/pdf/2604.21681" target="_blank" rel="noopener">
    <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M14 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V8z"/><polyline points="14 2 14 8 20 8"/><line x1="9" y1="13" x2="15" y2="13"/><line x1="9" y1="17" x2="15" y2="17"/></svg>
    Paper
  </a>
  <a class="pill" href="https://rawalkhirodkar.github.io/sapiens2" target="_blank" rel="noopener">
    <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><circle cx="12" cy="12" r="10"/><line x1="2" y1="12" x2="22" y2="12"/><path d="M12 2a15.3 15.3 0 0 1 4 10 15.3 15.3 0 0 1-4 10 15.3 15.3 0 0 1-4-10 15.3 15.3 0 0 1 4-10z"/></svg>
    Project
  </a>
</div>
"""

with gr.Blocks(title="Sapiens2 Seg", theme=gr.themes.Soft(), css=CUSTOM_CSS) as demo:
    gr.HTML(HEADER_HTML)

    with gr.Row(equal_height=True):
        inp = gr.Image(label="Input", type="pil", height=640)
        out_annot = gr.AnnotatedImage(
            label="Outputs (29 classes)",
            color_map=_CLASS_COLORS_HEX,
            height=640,
            show_legend=True,
            elem_id="seg-out",
        )

    with gr.Row():
        size = gr.Radio(
            choices=list(SEG_MODELS.keys()),
            value=DEFAULT_SIZE,
            label="Model",
            scale=4,
        )
        run = gr.Button("Run", variant="primary", size="lg", scale=1)

    gr.Examples(examples=EXAMPLES, inputs=inp, examples_per_page=16)

    with gr.Accordion("Original Res + Raw Labels", open=False):
        out_img = gr.Image(label="Color overlay (PNG)", type="pil")
        out_npy = gr.File(label="Raw labels (.npy uint8, class indices 0–28)")

    run.click(predict, inputs=[inp, size], outputs=[out_annot, out_img, out_npy])


if __name__ == "__main__":
    if torch.cuda.is_available():
        torch.backends.cuda.matmul.allow_tf32 = True
        torch.backends.cudnn.allow_tf32 = True
    demo.launch(share=False)