| --- |
| library_name: diffusers |
| base_model: Qwen/Qwen-Image-Edit-2511 |
| pipeline_tag: image-to-image |
| tags: |
| - pathology |
| - microscopy |
| - image-segmentation |
| - instance-segmentation |
| - lora |
| - qwen-image-edit |
| --- |
| |
| # Spatial Banana biomedical segmentation adapters |
|
|
| This repository stores Patho-Banana task LoRAs for adapting |
| Qwen-Image-Edit-2511 to direct biomedical instance-mask generation. The |
| adapters are not standalone pipelines: inference requires the frozen Qwen base, |
| the official four-step Qwen-Image-Edit-2511 Lightning LoRA, task-specific |
| prompts, aspect-aware geometry, and deterministic RGB-to-instance decoding. |
|
|
| ## Recommended release |
|
|
| Use **Mix704 step-8,800 rank 128** for the current multimodal segmentation |
| protocol: |
|
|
| ```text |
| experiments/multimodal/e6_sourcecap32_e125_v1/runs/mix704_step8800/final/pytorch_lora_weights.safetensors |
| ``` |
|
|
| | Field | Value | |
| | --- | --- | |
| | Size | 1,699,695,208 bytes | |
| | SHA-256 | `cb21fc1044e7f7721f0242b05581fa386e240eb35da4d9a6ef9cd27ddddfe830` | |
| | Training mixture | 128 PanNuke H&E plus 576 source-balanced non-H&E images | |
| | Task LoRA | rank 128, alpha 128, dropout 0 | |
| | Inference | BF16, 4 steps, seeds 42/314159/271828 | |
| | Decoder | `qwen_topology16_majority_v1` | |
|
|
| ## Results |
|
|
| On the previously viewed 822-image CellSAM non-H&E cohort: |
|
|
| | Method | Image-mean F1 | Pooled F1 | PQ | AJI+ | FG Dice | |
| | --- | ---: | ---: | ---: | ---: | ---: | |
| | Mix704 step 8,800 | 0.8428 | 0.8165 | 0.6914 | 0.7072 | 0.8793 | |
| | Mix448 E6 | 0.7538 | 0.7049 | 0.5943 | 0.6229 | 0.8099 | |
| | CellSAM v1.2 official | 0.8927 | 0.8926 | 0.7465 | 0.7692 | 0.9070 | |
|
|
| This is retrospective same-protocol evidence, not blinded SOTA or clinical |
| validation. Qwen remains below CellSAM overall. |
|
|
| ## Download |
|
|
| The source repository provides pinned HF-Mirror commands, SHA-256 verification, |
| local path placement, and preflight instructions: |
|
|
| - [Weight download guide](https://github.com/Biogod2020/spatial-banana/blob/main/docs/WEIGHTS.md) |
| - [Full model card](https://github.com/Biogod2020/spatial-banana/blob/main/docs/model_cards/MIX704_STEP8800_R128.md) |
| - [Current results](https://github.com/Biogod2020/spatial-banana/blob/main/docs/CURRENT_STATUS.md) |
|
|
| ## Archive contents |
|
|
| The other LoRAs cover nuclei few-shot experiments, epoch-aligned training, |
| Data-Frontier 256, MM144 adaptation, source-balance screening, all eight |
| Mix448 screen arms, and development-only biological applications. They are |
| development or recovery artifacts unless their evidence package explicitly |
| says otherwise. The presence of a weight does not mean it passed a scientific |
| gate. |
|
|
| `ARCHIVE_MANIFEST.tsv` records every archived weight path, byte size, and |
| SHA-256. Resolved configs, prompts, metrics, per-image rows, and completion |
| audits remain in the GitHub source repository. |
|
|
| ## Limitations and terms |
|
|
| The recommended adapter requires the Qwen base and Lightning LoRA and remains |
| subject to their licenses and terms. It is not intended for clinical diagnosis |
| or treatment decisions. This release grants no rights to upstream models or |
| training/evaluation datasets. |
|
|
