Downloads · 30 days
0
qmddn/seal-water-bottle-cap-stage1-resnet18-reward-v1
seal-water-bottle-cap-stage1-resnet18-reward-v1 is a machine learning model from qmddn. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
- RGB three-view vertical stack in Wan world-model format. - View order: camhigh + camleftwrist + camrightwrist. - Tensor shape: [3, 544, 320], CHW. - Accepted value range: uint8 [0, 255] or float [0, 1]; the model pr…
Downloads · 30 days
0
Access
Public
Updated Jun 12, 2026
Repo size
45.3 MB
Likes
0
Public
Click a slice to open those files.
.pt45.3 MB · 100%
From the Hugging Face model README
cam_high + cam_left_wrist + cam_right_wrist.[3, 544, 320], CHW.[0, 255] or float [0, 1]; the model preprocessing converts to float [0, 1] and applies ImageNet normalization.full_weights.pt is a PyTorch state_dict for the RLinf-compatible ResNet18 binary reward model.stage1 has been completed.sigmoid(logit) is the stage1 completion probability.sigmoid(logit) >= 0.499094.| File | Description |
|---|---|
full_weights.pt | ResNet18 stage1 reward checkpoint. |
model_metadata.json | Input/output schema, dataset counts, and model metadata. |
eval_summary.json | Validation metrics and F1 threshold sweep result. |
README.md | This model card. |
| score | threshold | F1 | precision | recall | accuracy | TP | FP | TN | FN |
|---|---|---|---|---|---|---|---|---|---|
| probability | 0.499094 | 0.904943 | 0.832168 | 0.991667 | 0.895833 | 119 | 24 | 96 | 1 |
| logit | -0.003625 | 0.904943 | 0.832168 | 0.991667 | 0.895833 | 119 | 24 | 96 | 1 |
| split | samples | positives | negatives | loss | AUC | [email protected] | [email protected] | positive prob mean | negative prob mean |
|---|---|---|---|---|---|---|---|---|---|
| val | 240 | 120 | 120 | 0.601936 | 0.914722 | 0.891667 | 0.900763 | 0.971991 | 0.202400 |