Feature/local processor for labels

[hardyliao85]

Description

Enhance Vision Service for label generation with configurable languages (en and zh_TW).

Changes:

1. Prevent model loading timeouts with --timeout 120 in Dockerfile.
2. Updated requirements.txt to specify torch==2.8.0+cu126 (CUDA version) for GPU inference support. Requires GPU driver: Linux >=525.60.13, Windows >=527.41.
3. Added 5 Hugging Face image classification models for hardware-based selection.
4. LABELS_LOCALE configurable via Docker Compose.

Below is my test configuration
compose:

# Only show key parts
services:
  photoprism-vision:
    image: photoprism/vision:develop
    ports:
      - "5000:5000"
    environment:
      NVIDIA_VISIBLE_DEVICES: "all"
      NVIDIA_DRIVER_CAPABILITIES: "compute,utility"
      LABELS_LOCALE: "zh_TW"
    volumes:
      - ./photoprism_vision_models:/app/models
      - ./venv:/app/venv
    deploy:
      resources:
        reservations:
          devices:
            - driver: "nvidia"
              capabilities: [ gpu ]
              count: "all"

vision.yml:

- Type: labels
  Name: convnextv2_huge.fcmae_ft_in22k_in1k_384
  Resolution: 384
  Service:
    Uri: http://IP:5000/api/v1/vision/labels
    FileScheme: data
    RequestFormat: vision
    ResponseFormat: vision

Related Issues

Partially addresses #19 (main functionality implemented; language support pending)

Acceptance Criteria

• New features or enhancements are fully implemented and do not break existing functionality, so that they can be released at any time without requiring additional work
• Automated unit and/or acceptance tests are included to ensure that changes work as expected and to reduce repetitive manual work
• Documentation has been / will be updated, especially as it relates to new configuration options or potentially disruptive changes

Note: Documentation updated only in README; no plans to update other official documents.
Note: I don't have much experience with PRs. I apologize if anything is unclear or not done properly. Please feel free to give me feedback, and I am happy to make any necessary corrections. Thank you very much.