Add evaluation metrics to model card

Adds notebook-derived evaluation protocol and metrics from train/ml-32m-train.ipynb.

README.md

@@ -9,6 +9,25 @@ tags:
 - movielens
 - implicit-feedback
 - implicit-als
+metrics:
+- recall
+- ndcg
+model-index:
+- name: ml32m-als128-v1
+  results:
+  - task:
+      type: top-n-recommendation
+      name: Top-N recommendation
+    dataset:
+      name: MovieLens ml-32m
+      type: MovieLens
+    metrics:
+    - type: recall_at_10
+      value: 0.08418674438152764
+      name: Recall@10
+    - type: ndcg_at_10
+      value: 0.06680269709374106
+      name: NDCG@10
 ---
 
 # ml32m-als128-v1
@@ -29,6 +48,25 @@ This repository contains an implicit-feedback ALS recommender artifact trained f
 - Saved user factors: false
 - Saved at UTC: 2026-07-02T15:18:10.162757+00:00
 
+## Evaluation
+
+Evaluation metrics were taken from `train/ml-32m-train.ipynb`.
+
+Protocol:
+
+- Held-out relevance: test-set ratings `>= 4.0`.
+- Seen-item filtering: movies present in each user's training history were excluded from the ranked list.
+- Ranking sample: 10,000 users sampled with seed `0` from 187,278 users with at least one relevant test item.
+- Cutoff: `K=10`.
+
+| Model | Recall@10 | NDCG@10 | Users |
+| --- | ---: | ---: | ---: |
+| ALS implicit | 0.0842 | 0.0668 | 10,000 |
+| Popularity baseline | 0.0489 | 0.0430 | 10,000 |
+| Explicit MF baseline | 0.0005 | 0.0005 | 10,000 |
+
+The ALS model is optimized for implicit-feedback ranking, so rating-prediction RMSE/MAE is not reported for this artifact.
+
 ## Repository Files
 
 - `als_model.npz`: ALS artifact containing saved item factors and training parameters.