Upload pipeline-ready SDFT-paper distil build

README.md

@@ -0,0 +1,36 @@
+---
+dataset_info:
+  features:
+  - name: prompt
+    list:
+    - name: content
+      dtype: string
+    - name: role
+      dtype: string
+  - name: teacher_prompt
+    list:
+    - name: content
+      dtype: string
+    - name: role
+      dtype: string
+  - name: chat_template_kwargs
+    struct:
+    - name: enable_thinking
+      dtype: bool
+  splits:
+  - name: train
+    num_bytes: 7814546
+    num_examples: 2674
+  - name: eval
+    num_bytes: 731812
+    num_examples: 507
+  download_size: 2754492
+  dataset_size: 8546358
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: eval
+    path: data/eval-*
+---

data/eval-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:844b403b650b69fe6b31671e8b1f92147b6ca6c09f7c886ac47a3799facac2be
+size 124957

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c80ece36d2c3d30fc66c31d67f7fc0794f8749a164f51a6d1092a2420c4e7baf
+size 2629535