Fix split to 35k train / 500 val / 500 test

README.md

@@ -20,13 +20,11 @@ The data is extracted from `bigcode/the-stack-v2-dedup`, Python subset. The extr
 
 ## Splits
 
-- `train.jsonl`: 26,000
-- `val.jsonl`: 5,000
-- `test.jsonl`: 5,000
+- `train.jsonl`: 35,000
+- `val.jsonl`: 500
+- `test.jsonl`: 500
 - `all.jsonl`: 36,000
 
-Note: the requested split `35k/5k/5k` sums to 45k, not 36k. This 36k snapshot keeps validation and test at 5k each and assigns the remaining 26k to train.
-
 ## Record Format
 
 Each JSONL row is one Python class sample. Important fields include:

all.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:06fc3f22d8fea4edfa0b2230f24110c75d1098184f7f603dbb52a9ffefcf72cc
-size 444140865
+oid sha256:8a2be28fdd4e89ae72ae5f1bbfb7e1fc5e535f3c8186a44d7ec96aa989aa26ea
+size 444179967

metadata.json

@@ -6,11 +6,10 @@
   "snapshot_size": 36000,
   "snapshot_seed": 42,
   "bad_json_lines_skipped": 0,
-  "requested_split_note": "User requested a 36k dataset and also 35k/5k/5k splits, which sums to 45k. This 36k snapshot uses 26k/5k/5k to keep validation and test at 5k each.",
   "splits": {
-    "train": 26000,
-    "val": 5000,
-    "test": 5000
+    "train": 35000,
+    "val": 500,
+    "test": 500
   },
   "format": "class-level Python code generation with prompt, skeleton, method bodies, and reconstructed solution",
   "filters": {
@@ -26,48 +25,48 @@
     "pyflakes_hard_filter": false
   },
   "method_count_distribution": {
-    "2": 21570,
-    "3": 8008,
-    "4": 3586,
-    "5": 1856,
-    "6": 980
+    "2": 21551,
+    "3": 8022,
+    "4": 3599,
+    "5": 1850,
+    "6": 978
   },
   "top_repos": [
     [
       "cohesity/management-sdk-python",
-      456
+      449
     ],
     [
       "microsoftgraph/msgraph-sdk-python",
-      401
+      394
     ],
     [
       "home-assistant/core",
-      291
+      286
     ],
     [
-      "jansel/pytorch-jit-paritybench",
-      124
+      "sungminoh/algorithms",
+      122
     ],
     [
-      "sungminoh/algorithms",
-      124
+      "jansel/pytorch-jit-paritybench",
+      122
     ],
     [
       "953250587/leetcode-python",
-      121
+      119
     ],
     [
       "Shiv2157k/leet_code",
-      118
+      117
     ],
     [
       "Jimmy-INL/google-research",
-      101
+      100
     ],
     [
       "metoppv/improver",
-      91
+      90
     ],
     [
       "bssrdf/pyleet",
@@ -75,45 +74,45 @@
     ],
     [
       "dealflowteam/Idfy",
-      78
-    ],
-    [
-      "vincent-lg/tsunami",
       76
     ],
-    [
-      "oxhead/CodingYourWay",
-      75
-    ],
     [
       "vsdrun/lc_public",
       75
     ],
     [
       "JavaRod/SP_Python220B_2019",
-      75
+      74
     ],
     [
-      "mindspore-ai/models",
-      72
+      "oxhead/CodingYourWay",
+      73
     ],
     [
-      "log2timeline/plaso",
-      68
+      "vincent-lg/tsunami",
+      73
+    ],
+    [
+      "mindspore-ai/models",
+      71
     ],
     [
       "inventree/InvenTree",
       67
     ],
     [
-      "yiming1012/MyLeetCode",
-      64
+      "log2timeline/plaso",
+      66
     ],
     [
       "luque/better-ways-of-thinking-about-software",
       63
+    ],
+    [
+      "yiming1012/MyLeetCode",
+      63
     ]
   ],
-  "avg_body_lines": 9.144970414201184,
+  "avg_body_lines": 9.1440909372996,
   "task": "Given class/method signatures and docstrings, generate class skeleton and method bodies for Diffusion + AR hybrid code generation experiments."
 }

test.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:45bb3d8da6ea0f12a501f16316d4edfa56f3caaba3656c0e3b3a6844139806c1
-size 61490756
+oid sha256:45428812dea1255651ef87f387dcdc9fda1dc948c77f439fb93814cc0b1a363f
+size 6119695

train.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfe4cc8c0d8451b01dbfdb0591e4bfad53f55e5b19f9784d0276a4847a09938d
-size 320962541
+oid sha256:2085f61548444e77c5d824a18111f593779d70060833f336435a93f82ae232b5
+size 431871109

val.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bff1b6b656ddd819af37d18d505169ee8432ab0aed08156963202af330156a6f
-size 61687568
+oid sha256:6ed2725fd07e42e8d4557667a7edbd5eeab56399c024afb5e559de11b95a9a77
+size 6189163