Upload fine-tuned YOLO11n-Seg model and metadata

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch0_labels.jpg filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch0_pred.jpg filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch1_labels.jpg filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch1_pred.jpg filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch2_labels.jpg filter=lfs diff=lfs merge=lfs -text
+model/evaluation/val_batch2_pred.jpg filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,78 @@
+---
+license: agpl-3.0
+library_name: ultralytics
+pipeline_tag: image-segmentation
+tags:
+  - yolo
+  - yolo11
+  - instance-segmentation
+  - robotics
+  - synthetic-data
+  - so-101
+---
+
+# SO-101 YOLO11n-Seg plastic cup model
+
+This repository contains a YOLO11n-Seg model fine-tuned to segment the `plastic_cup` class in the SO-101 simulation scene. It also includes the synthetic dataset used for training and evaluation.
+
+## Repository contents
+
+- `best.pt`: fine-tuned Ultralytics checkpoint.
+- `model/metrics.json`: metrics from the held-out synthetic test split.
+- `model/training-config.yaml`: portable training settings.
+- `model/evaluation/`: test plots and prediction previews.
+- `dataset/`: YOLO segmentation images, polygon labels, per-sample truth, and the dataset manifest.
+- `SHA256SUMS`: hashes for the uploaded files.
+
+## Dataset
+
+The dataset has 1,200 synthetic 640 x 480 images rendered from the SO-101 MuJoCo scene through `task_camera`. Object-ID segmentation supplied the label masks. The split is fixed by seed range:
+
+| Split | Images | Labels | Seeds |
+| --- | ---: | ---: | --- |
+| train | 800 | 800 | 100000-100799 |
+| validation | 200 | 200 | 200000-200199 |
+| test | 200 | 200 | 300000-300199 |
+
+The only class is `plastic_cup` (`class_id=0`). Samples include no-cup, one-cup, and multi-cup scenes. The manifest records the scenario, seed, visible instance count, image path, label path, and synthetic truth path for each sample.
+
+Dataset generator commit: `2be8df09302feabffc7f028b16c90d06867f8055` (`build(perception): pin training runtime and labeled overlays`).
+
+## Training
+
+The run started from `yolo11n-seg.pt` and used Ultralytics segmentation training with these main settings:
+
+| Setting | Value |
+| --- | --- |
+| image size | 640 |
+| epochs | 100 |
+| batch size | 32 |
+| seed | 20260831 |
+| deterministic | true |
+| device | CUDA |
+| AMP | false |
+| optimizer | auto |
+
+Checkpoint SHA-256: `f281d25258493e2c7c220dd1d84a7ca4f0501adf99ed4a921a065d74ace40781`.
+
+## Synthetic test results
+
+| Metric | Boxes | Masks |
+| --- | ---: | ---: |
+| precision | 0.999735 | 0.999735 |
+| recall | 1.000000 | 1.000000 |
+| mAP50 | 0.995000 | 0.995000 |
+| mAP50-95 | 0.995000 | 0.973662 |
+
+These numbers describe the fixed synthetic test split. They do not establish accuracy on real cameras, unfamiliar cup appearances, or physical grasp success.
+
+## Use with Ultralytics
+
+```python
+from ultralytics import YOLO
+
+model = YOLO("best.pt")
+results = model.predict("image.png")
+```
+
+The checkpoint inherits the licensing requirements of its Ultralytics YOLO11 base model. The repository is private because a separate license has not been declared for the synthetic dataset.

best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f281d25258493e2c7c220dd1d84a7ca4f0501adf99ed4a921a065d74ace40781
+size 6001316

model/metrics.json

@@ -0,0 +1,24 @@
+{
+  "names": {
+    "0": "plastic_cup"
+  },
+  "results_dict": {
+    "fitness": 1.9686621837012601,
+    "metrics/mAP50(B)": 0.995,
+    "metrics/mAP50(M)": 0.995,
+    "metrics/mAP50-95(B)": 0.9949999999999999,
+    "metrics/mAP50-95(M)": 0.9736621837012602,
+    "metrics/precision(B)": 0.9997351414137473,
+    "metrics/precision(M)": 0.9997351414137473,
+    "metrics/recall(B)": 1.0,
+    "metrics/recall(M)": 1.0
+  },
+  "speed_ms": {
+    "inference": 0.9762399850001202,
+    "loss": 0.00019861499822582118,
+    "postprocess": 0.394080710000253,
+    "preprocess": 1.2785653700029798
+  },
+  "split": "test"
+}
+

model/provenance.json

@@ -0,0 +1,25 @@
+{
+  "schema_version": 1,
+  "task": "segment",
+  "class_names": {
+    "0": "plastic_cup"
+  },
+  "base_model": "yolo11n-seg.pt",
+  "checkpoint_sha256": "f281d25258493e2c7c220dd1d84a7ca4f0501adf99ed4a921a065d74ace40781",
+  "dataset_generator_commit": "2be8df09302feabffc7f028b16c90d06867f8055",
+  "dataset_generator_commit_subject": "build(perception): pin training runtime and labeled overlays",
+  "training_run": "full-exp-012",
+  "training_seed": 20260831,
+  "dataset_sample_count": 1200,
+  "dataset_split_counts": {
+    "train": 800,
+    "val": 200,
+    "test": 200
+  },
+  "dataset_image_size": {
+    "width": 640,
+    "height": 480
+  },
+  "dataset_camera": "task_camera",
+  "dataset_manifest_schema_version": 1
+}

model/training-config.yaml

@@ -0,0 +1,19 @@
+task: segment
+model: yolo11n-seg.pt
+data: dataset/dataset.yaml
+classes:
+  - 0
+imgsz: 640
+epochs: 100
+batch: 32
+workers: 8
+seed: 20260831
+deterministic: true
+device: cuda
+amp: false
+optimizer: auto
+patience: 20
+save: true
+plots: true
+verbose: true
+