Add files using upload-large-folder tool

.gitattributes

@@ -58,3 +58,15 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+title/test.jsonl filter=lfs diff=lfs merge=lfs -text
+title/val_meta.jsonl filter=lfs diff=lfs merge=lfs -text
+title/test_meta.jsonl filter=lfs diff=lfs merge=lfs -text
+author/val_meta.jsonl filter=lfs diff=lfs merge=lfs -text
+author/val.jsonl filter=lfs diff=lfs merge=lfs -text
+author/test.jsonl filter=lfs diff=lfs merge=lfs -text
+author/test_meta.jsonl filter=lfs diff=lfs merge=lfs -text
+title/val.jsonl filter=lfs diff=lfs merge=lfs -text
+title/train.jsonl filter=lfs diff=lfs merge=lfs -text
+author/train.jsonl filter=lfs diff=lfs merge=lfs -text
+author/train_meta.jsonl filter=lfs diff=lfs merge=lfs -text
+title/train_meta.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,49 @@
+---
+license: cc-by-4.0
+task_categories:
+  - text-generation
+language:
+  - bo
+tags:
+  - tibetan
+  - metadata
+  - title
+  - author
+  - llm-sft
+  - tilamb
+size_categories:
+  - 100K<n<1M
+---
+
+# Tibetan metadata LLM SFT dataset
+
+Supervised fine-tuning data for **title** and **author** span extraction from BDRC outliner segments, built for [TiLamb-7B](https://huggingface.co/YoLo2000/TiLamb-7B).
+
+## Layout
+
+```
+title/{train,val,test}.jsonl       # Alpaca format for LLaMA-Factory
+title/{train,val,test}_meta.jsonl
+author/{train,val,test}.jsonl
+author/{train,val,test}_meta.jsonl
+dataset_info.json                  # LLaMA-Factory registry snippet
+reports/crop_stats.json
+```
+
+Each training row:
+
+- `instruction` — fixed task prompt (title or author)
+- `input` — cropped segment text (token-budget ≤3584 via TiLamb tokenizer)
+- `output` — JSON `{"spans":[{"text","start","end"}]}` (crop-relative offsets)
+
+## Cropping
+
+| Kind | When |
+|------|------|
+| `full` | Whole segment fits token budget |
+| `positive` | Random window containing gold span (anti position-bias) |
+| `negative` | Random window, empty spans for that task |
+
+Source: [ganga4364/tibetan-metadata-extracted](https://huggingface.co/datasets/ganga4364/tibetan-metadata-extracted) (3794 docs).
+
+Built with [OpenPecha/tibetan-text-meta-detection](https://github.com/OpenPecha/tibetan-text-meta-detection) `llm_sft` package.

author/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:918e346aa9398aa6c6cd54cc974e77c152ef4d4f9da9567d3f3204a0836e9659
+size 151104227

author/test_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db207ffdd2075a3ef970d05166902a8fa531ddfa7d4833b7a044a17def0bbfa3
+size 153496285

author/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:329873390fe98b2fa02e0589988a237ecf739352741d5fefbba1da24e29cac98
+size 1295318184

author/train_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a5ec3be22db69b2e99de328dfab3e2545cf5165e6ea548c97af1fa7bf0786c9
+size 1315581432

author/val.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a89147285e8764701ef8c6a3726a751da9371e86071719294bd3852db81beb22
+size 15351707

author/val_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6bf857492a4564601e534b64eb6f0ddc3e88e6f09c74fd0965299c0f414b564
+size 15569501

dataset_info.json

@@ -0,0 +1,18 @@
+{
+  "tibetan_title_sft": {
+    "file_name": "title/train.jsonl",
+    "columns": {
+      "prompt": "instruction",
+      "query": "input",
+      "response": "output"
+    }
+  },
+  "tibetan_author_sft": {
+    "file_name": "author/train.jsonl",
+    "columns": {
+      "prompt": "instruction",
+      "query": "input",
+      "response": "output"
+    }
+  }
+}

reports/crop_stats.json

@@ -0,0 +1,42 @@
+{
+  "rows": 134835,
+  "segments": 57091,
+  "errors": 29037,
+  "crop_kind": {
+    "negative": 14635,
+    "positive": 74535,
+    "full": 45665
+  },
+  "task": {
+    "title": 69058,
+    "author": 65777
+  },
+  "split": {
+    "train:title": 60359,
+    "train:author": 58268,
+    "val:title": 1006,
+    "val:author": 626,
+    "test:title": 7693,
+    "test:author": 6883
+  },
+  "span_position_ratio_histogram": {
+    "0.0-0.1": 55956,
+    "0.1-0.2": 1651,
+    "0.2-0.3": 826,
+    "0.3-0.4": 463,
+    "0.4-0.5": 342,
+    "0.5-0.6": 336,
+    "0.6-0.7": 526,
+    "0.7-0.8": 1173,
+    "0.8-0.9": 4224,
+    "0.9-1.0": 54703
+  },
+  "span_position_ratio_median": 0.7,
+  "config": {
+    "tokenizer": "YoLo2000/TiLamb-7B",
+    "max_context_tokens": 3584,
+    "crops_per_positive": 3,
+    "crops_per_negative": 1,
+    "seed": 42
+  }
+}

title/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdd0f1c9034bd73e5992b54c0f95220fda5a2ed980e3be806a21dc8efbaff439
+size 154468940

title/test_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54e0250326460f792c096a2df65830fc0224ead1414fd33df499756e9950c379
+size 157120121

title/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5419aa2b58c3c1762dd4120e07600214c7d6f028d0be9224865f3290e431d6e2
+size 1233059814

title/train_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f392475cb061d5ec07628347438240c48ddefb40deb4537a8568787e0a71dda4
+size 1253854094

title/val.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:004c2e02e79d4a285ab0df9cd0795943ef959d822f5d2406683ab30dbe914f93
+size 18447356

title/val_meta.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8179529ba0ff754f33557d237ebf37e6d5320037177d1fd635e8723d2cb72232
+size 18794088