Skip to content
Other

Depth Pro 2-focal: Apple's focal length estimator

2 min readupdated

The DepthPro2-focal model is a specialized component from Apple's new Depth Pro 2 lineup, designed to predict focal length in pixels for individual photos and video frames. Its compact size of 23.5M parameters allows for quick integration of focus estimation into image and video processing pipelines. The model helps recover metric depth and scene geometry even when the original frame metadata is lost. The tool is useful for computer vision developers, researchers, and multimedia creators working with spatial analysis.

Key features

  • Compact size: Only 23.5 million parameters for fast performance.
  • Specialization: Precise determination of focal length in pixels for photos and videos.
  • Easy integration: Simple launch via the official library with PyTorch and CUDA support.

What you can do

  • Frame analysis: Embed focus estimation into video pipelines to process clips frame by frame.
  • Photo processing: Recover focal length for images without EXIF data.
  • Research projects: Use the component in computer vision and depth recovery tasks.

Example

from functools import partial
import torch
from depth_pro_2 import load_rgb
from depth_pro_2.model import to_normalized_tensor
from depth_pro_2.network.fov import FOVEstimator

fov = FOVEstimator.from_pretrained("apple/DepthPro2-focal").to("cuda").eval()
transform = partial(to_normalized_tensor, device=torch.device("cuda"))

image, _ = load_rgb("photo.jpg")
with torch.no_grad():
    f_px = fov.focal_length_px(transform(image)[None])
Read moreDetails
Parameters: 23.5M
Library: depth-pro-2
License: Apple Machine Learning Research Model license (non-commercial scientific and academic research only)
Discuss Telegram 20 370

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.1 GB · ready files0.6 GB
  • BF16weights 0.0 GB0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # apple/DepthPro2-focal
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download apple/DepthPro2-focal --local-dir ~/models/DepthPro2-focal
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to DepthPro2-focal
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of DepthPro2-focal (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

apple/DepthPro2-focal
Author
apple
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

37020

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments