metadata
model-index:
- name: vsd-benchmark/vsd-fashion
results:
- task:
type: in_catalog_retrieval_zero_shot
dataset:
type: vsd-benchmark/vsd-fashion
name: vsd-benchmark/vsd-fashion
config: in_catalog_retrieval_zero_shot
metrics:
- type: roc_auc
value: 0
name: ROC_AUC
- type: posicube/mean_reciprocal_rank
value: 0
args:
k: 5
name: MRR@5
- task:
type: in_catalog_open_catalog
dataset:
type: vsd-benchmark/vsd-fashion
name: vsd-benchmark/vsd-fashion
config: in_catalog_open_catalog
metrics:
- type: roc_auc
value: 0
name: ROC_AUC
- type: posicube/mean_reciprocal_rank
value: 0
args:
k: 5
name: MRR@5
- task:
type: in_catalog_closed_catalog
dataset:
type: vsd-benchmark/vsd-fashion
name: vsd-benchmark/vsd-fashion
config: in_catalog_closed_catalog
metrics:
- type: roc_auc
value: 0
name: ROC_AUC
- type: posicube/mean_reciprocal_rank
value: 0
args:
k: 5
name: MRR@5
- task:
type: consumer-catalog_wild_zero_shot
dataset:
type: vsd-benchmark/vsd-fashion
name: vsd-benchmark/vsd-fashion
config: consumer-catalog_wild_zero_shot
metrics:
- type: roc_auc
value: 0
name: ROC_AUC
- type: posicube/mean_reciprocal_rank
value: 0
args:
k: 5
name: MRR@5
metrics:
- roc_auc
- posicube/mean_reciprocal_rank
tags:
- vsd
This is an example model card for reporting metrics for the VSD fashion dataset.
To report your model metrics, you need to do 2 things:
- Tag your model with 'vsd'.
- In your model card README.md add a section with:
---
model-index:
- name: vsd-benchmark/vsd-fashion
results:
- task:
type: in_catalog_retrieval_zero_shot
dataset:
type: vsd-benchmark/vsd-fashion
name: vsd-benchmark/vsd-fashion
config: in_catalog_retrieval_zero_shot
metrics:
- type: roc_auc
value: 0
name: ROC_AUC
- type: posicube/mean_reciprocal_rank
value: 0
args:
k: 5
name: MRR@5
metrics:
- roc_auc
- posicube/mean_reciprocal_rank
tags:
- vsd
---
This will report a metric on the vsd-fashion dataset, on the in_catalog_retrieval_zero_shot task.