Upload folder using huggingface_hub

README.md

@@ -0,0 +1,39 @@
+---
+configs:
+- config_name: cluster_0
+  default: true
+  data_files:
+  - split: train
+    path: "cluster_0/*.parquet"
+- config_name: cluster_1
+  data_files:
+  - split: train
+    path: "cluster_1/*.parquet"
+- config_name: cluster_2
+  data_files:
+  - split: train
+    path: "cluster_2/*.parquet"
+- config_name: cluster_3
+  data_files:
+  - split: train
+    path: "cluster_3/*.parquet"
+---
+
+# 10model_rank_k4
+
+This repository contains k-means cluster subsets for training. Each subset is a config (`name=cluster_X`), backed by Parquet.
+
+## Load examples
+
+```python
+from datasets import load_dataset
+
+# Single cluster
+ds0 = load_dataset("{username}/10model_rank_k4", name="cluster_0")
+
+# Specific cluster
+ds3 = load_dataset("{username}/10model_rank_k4", name="cluster_3")
+
+# Or with data_files directly (optional alternative):
+ds = load_dataset("{username}/10model_rank_k4", data_files={"train": "cluster_3/*.parquet"}, split="train")
+```

cluster_0/data.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e708490764aa24ece6b0b3db8e1fa7c17c3a833c0123a1513b97e2a4f65bf41
+size 806415859

cluster_1/data.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:554452e7fcb51907cce0822f48d0c0f080250588a863c47ff13319520733db44
+size 55386562

cluster_2/data.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:372702e990dc20db42c2b4c0f71f5ef7ad287034f5195bc4d7ceae55360ab2e4
+size 634832272

cluster_3/data.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ee8d6f74f2040c0c856d4200ce2561c7d9c9e71338efffcc33db22fe430b15b
+size 24889946

manifest.json

@@ -0,0 +1,32 @@
+{
+  "totals_per_cluster": {
+    "0": 106120,
+    "1": 66029,
+    "2": 42061,
+    "3": 35790
+  },
+  "datasets_order": [
+    "acereason_f50k",
+    "openr1_f50k",
+    "openthoughts_f50k",
+    "openo1_f50k",
+    "nemotron_f50k"
+  ],
+  "dataset_sizes": [
+    50000,
+    50000,
+    50000,
+    50000,
+    50000
+  ],
+  "dataset_offsets": [
+    0,
+    50000,
+    100000,
+    150000,
+    200000
+  ],
+  "per_cluster_sample": null,
+  "seed": 42,
+  "source_npz": "/home/adanato/Research/data_val/packages/fykcluster/cluster_results/10model_rank_k4/sampled_data.npz"
+}