Hugging Face's logo

Datasets:
TsienDragon
/
qwen-image-finetune-test-resources

Dataset card Files
xet
Community
Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

Test Resources for Qwen Image Finetune

This repository contains test resources for the qwen-image-finetune project.

Directory Structure 

test_resources_organized/
├── flux_models/              # Flux model related test data
│   └── transformer/
│       └── input/            # Transformer input test data
│           └── flux_input.pth (19MB)
│
├── flux_training/            # Training process test data
│   └── face_segmentation/    # Face segmentation training samples
│       ├── sample1/          # Training sample 1 (11MB)
│       │   ├── sample_control_ids.pt
│       │   ├── sample_control_latents.pt
│       │   ├── sample_guidance.pt
│       │   ├── sample_image_latents.pt
│       │   ├── sample_latent_ids.pt
│       │   ├── sample_latent_model_input.pt
│       │   ├── sample_loss.pt
│       │   ├── sample_model_pred.pt
│       │   ├── sample_noise.pt
│       │   ├── sample_pooled_prompt_embeds.pt
│       │   ├── sample_prompt_embeds.pt
│       │   ├── sample_t.pt
│       │   └── sample_text_ids.pt
│       └── sample2/          # Training sample 2 (12MB)
│           └── (same structure as sample1)
│
├── flux_sampling/            # Sampling and generation test data
│   └── embeddings/           # Embedding tensors for sampling (7.3MB total)
│       ├── aaa_sample_control_image.pt
│       ├── sample_control_ids.pt
│       ├── sample_control_image.pt
│       ├── sample_control_latents.pt
│       ├── sample_latent_ids.pt
│       ├── sample_latents.pt
│       ├── sample_pooled_prompt_embeds.pt
│       ├── sample_prompt_embeds.pt
│       └── sample_text_ids.pt
│
└── reference_outputs/        # Reference output images for comparison
    └── images/               # Generated test images (688KB total)
        ├── test_flux_kontext_output.png
        └── test_flux_kontext_output_unit_test.png

Total Size

Approximately 50MB total

Usage

These resources are automatically downloaded by the test suite when needed. See tests/resources_config.yaml for resource group definitions and test dependencies.

Resource Groups

  • flux_input: Flux transformer input test data
  • flux_training_sample1: Training process test sample 1
  • flux_training_sample2: Training process test sample 2
  • flux_sampling: Sampling and generation embeddings
  • reference_outputs: Reference output images

Adding New Resources

  1. Add files to appropriate directories following the structure above
  2. Update tests/resources_config.yaml in the main repository
  3. Upload to HuggingFace Hub using the upload script
Downloads last month
40