Upload folder using huggingface_hub

README.md

@@ -0,0 +1,154 @@
+---
+license: odc-by
+task_categories:
+  - text-generation
+language:
+  - code
+size_categories:
+  - 1B<n<10B
+tags:
+  - curated
+  - starcoderdata
+  - code
+  - structured-data
+  - multi-task-filter
+---
+
+# StarCoderData Curated
+
+A curated subset of [StarCoderData](https://huggingface.co/datasets/bigcode/starcoderdata)
+optimised for training a 500M parameter model focused on structured data output
+(JSON generation, function calling, schema compliance).
+
+## Dataset Summary
+
+- **Total code files:** 5,203,508
+- **Total tokens:** 3.9B (target: 3.5B)
+- **Source:** bigcode/starcoderdata
+- **Classifier:** [mdonigian/code-curator-v1](https://huggingface.co/mdonigian/code-curator-v1) (UniXcoder-base, multi-task)
+- **Curation:** Per-language-slice filtering + compression ratio pre-filter + MinHash deduplication
+
+## Filtering Strategy
+
+Each language slice has its own filtering strategy and token budget:
+
+1. **Schema languages** (JSON/YAML/SQL/protobuf/GraphQL): ~800M tokens, light filter (quality floor only)
+2. **TypeScript** (relevance ≥ 2): ~600M tokens, relevance classifier
+3. **Python** (relevance ≥ 2): ~600M tokens, relevance classifier
+4. **Rust/Go/Java** (relevance ≥ 2): ~600M tokens, relevance classifier
+5. **Jupyter notebooks**: ~400M tokens, passthrough (already structured)
+6. **GitHub Issues** (technical): ~500M tokens, keyword filter
+
+Pre-filtering applied to all slices:
+- zlib compression ratio < 0.10 catches extreme repetition
+- Quality floor: drop files with predicted quality ≤ 1 (broken/gibberish)
+- MinHash LSH deduplication with line-level 5-gram shingles
+
+## Language Slice Distribution
+
+| Slice | Strategy | Target | Actual | % |
+|-------|----------|--------|--------|---|
+| schema_languages | light_filter | 800M | 799M | 99.9% |
+| typescript | relevance_filter | 600M | 598M | 99.7% |
+| python | relevance_filter | 600M | 594M | 99.1% |
+| rust_go_java | relevance_filter | 600M | 485M | 80.8% |
+| jupyter | relevance_filter | 400M | 0M | 0.0% |
+| github_issues | keyword_filter | 500M | 426M | 85.2% |
+| general_code | light_filter | 1000M | 999M | 99.9% |
+
+
+## Content Group Distribution
+
+| Group | % of Tokens | Tokens | Files |
+|-------|-------------|--------|-------|
+| Library/Package | 27.7% | 1,079,381,502 | 1,075,672 |
+| Application | 1.4% | 56,360,750 | 118,200 |
+| Script/CLI | 0.5% | 17,853,662 | 24,871 |
+| Test Code | 2.3% | 91,370,763 | 48,003 |
+| Config/Data/Generated/Other | 68.1% | 2,656,726,472 | 3,936,762 |
+
+
+## Structured Data Relevance Distribution
+
+The strongest classifier signal (Spearman 0.81). SD2+ files contain significant
+structured data patterns (API endpoints, JSON parsing, schema definitions, etc.).
+
+| Level | Range | Target % | Actual % | Files |
+|-------|-------|----------|----------|-------|
+| SD0 | [0.0, 0.5) | 10.0% | 70.1% | 3,649,912 |
+| SD1 | [0.5, 1.5) | 20.0% | 0.0% | 0 |
+| SD2 | [1.5, 2.5) | 35.0% | 0.9% | 49,213 |
+| SD3 | [2.5, 3.5) | 35.0% | 28.9% | 1,504,383 |
+
+
+## Quality Distribution
+
+Quality mean: 1.13, Median: 0.00.
+
+| Level | Description | Files |
+|-------|-------------|-------|
+| 1 | Broken/gibberish | 3,649,912 |
+| 2 | Functional but poor | 42,668 |
+| 3 | Decent | 129,945 |
+| 4 | Good | 1,380,674 |
+| 5 | Excellent | 309 |
+
+
+## Programming Languages
+
+| Language | % Tokens | Files |
+|----------|----------|-------|
+| typescript | 15.3% | 841,426 |
+| python | 15.2% | 567,721 |
+| github-issues-filtered-structured | 10.9% | 485,384 |
+| markdown | 8.9% | 351,728 |
+| json | 8.7% | 1,124,326 |
+| go | 8.5% | 73,899 |
+| sql | 5.9% | 121,035 |
+| javascript | 5.8% | 281,216 |
+| yaml | 5.8% | 957,872 |
+| java | 3.2% | 57,787 |
+| c-sharp | 3.0% | 114,063 |
+| html | 2.9% | 53,527 |
+| c | 2.8% | 75,899 |
+| haskell | 2.2% | 84,862 |
+| rust | 0.7% | 12,752 |
+
+
+## Token Count Distribution
+
+| Percentile | Tokens |
+|------------|--------|
+| P10 | 55 |
+| P25 | 111 |
+| P50 (median) | 255 |
+| P75 | 631 |
+| P90 | 1,416 |
+| Mean | 749 |
+
+## Schema
+
+Each row contains:
+
+| Field | Type | Description |
+|-------|------|-------------|
+| `content` | string | Source code text |
+| `lang` | string | Programming language |
+| `size` | int | File size in bytes |
+| `token_count` | int | Estimated token count (size // 4) |
+| `quality` | float | Code quality score (1-5) |
+| `structured_data` | float | Structured data relevance (0-3) |
+| `content_type` | string | Content type (9 classes) |
+| `language_slice` | string | Language slice (schema_languages, typescript, python, etc.) |
+| `relevance_score` | float | Composite relevance score |
+
+## Methodology
+
+1. **Classification:** Single multi-task UniXcoder-base model with three heads runs
+   on each code file. Stores raw float scores for maximum flexibility.
+2. **Pre-filtering:** zlib compression ratio filter removes repetitive boilerplate
+   before GPU inference.
+3. **Filtering:** Quality floor + priority-based sampling targeting high structured
+   data relevance and balanced content type distribution.
+4. **Deduplication:** MinHash LSH (128 perms, 5-line shingles, 0.7 Jaccard threshold).
+   Highest-relevance code file kept from each cluster.

data_0000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e80d9ff747cb6b6b4fb41f320cd20b6a949769673a907ed9fbfe2bc6e459e76
+size 1368339706

data_0001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ee4da50457384e1be54ac87f7737c6d30ca10ce425db6bd5fb5fdc0d7ec1492
+size 1483550079

data_0002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c966e282de204164c8870c5af8287e647e5b6a511f236eeba4ec19ea711afd27
+size 1347029171

data_0003.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:564d1f393566a599f9734206526285ba145a3ccc826e7fa21dc5569a1a0b297d
+size 900649083

data_0004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:504d5ccc627cc0c5091f8441317ebd0abdfe33d6d3dbef575caac8f2278ae295
+size 676566278

data_0005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc160ca6f322af12adb23b2b8b899f8ad5b063c858ab4ea06feca71198ff5452
+size 125706433

dataset_stats.json

@@ -0,0 +1,376 @@
+{
+  "total_documents": 5203508,
+  "total_tokens": 3901693149,
+  "num_output_files": 6,
+  "group_distribution": {
+    "library": {
+      "display_name": "Library/Package",
+      "docs": 1075672,
+      "tokens": 1079381502,
+      "pct_tokens": 27.66,
+      "pct_of_total": 27.66
+    },
+    "application": {
+      "display_name": "Application",
+      "docs": 118200,
+      "tokens": 56360750,
+      "pct_tokens": 1.44,
+      "pct_of_total": 1.44
+    },
+    "script": {
+      "display_name": "Script/CLI",
+      "docs": 24871,
+      "tokens": 17853662,
+      "pct_tokens": 0.46,
+      "pct_of_total": 0.46
+    },
+    "test": {
+      "display_name": "Test Code",
+      "docs": 48003,
+      "tokens": 91370763,
+      "pct_tokens": 2.34,
+      "pct_of_total": 2.34
+    },
+    "low_value": {
+      "display_name": "Config/Data/Generated/Other",
+      "docs": 3936762,
+      "tokens": 2656726472,
+      "pct_tokens": 68.09,
+      "pct_of_total": 68.09
+    }
+  },
+  "content_type_distribution": {
+    "library": 1075672,
+    "application": 118200,
+    "test": 48003,
+    "config": 3133,
+    "tutorial": 2843,
+    "data": 5585,
+    "generated": 275289,
+    "script": 24871,
+    "other": 0
+  },
+  "quality_distribution": {
+    "1": 3649912,
+    "2": 42668,
+    "3": 129945,
+    "4": 1380674,
+    "5": 309
+  },
+  "quality_stats": {
+    "mean": 1.131,
+    "std": 1.75,
+    "median": 0.0
+  },
+  "sd_distribution": {
+    "SD0": {
+      "range": "[0.0, 0.5)",
+      "docs": 3649912,
+      "tokens": 2224333326,
+      "pct": 70.14,
+      "target_pct": 10.0
+    },
+    "SD1": {
+      "range": "[0.5, 1.5)",
+      "docs": 0,
+      "tokens": 0,
+      "pct": 0.0,
+      "target_pct": 20.0
+    },
+    "SD2": {
+      "range": "[1.5, 2.5)",
+      "docs": 49213,
+      "tokens": 55140562,
+      "pct": 0.95,
+      "target_pct": 35.0
+    },
+    "SD3": {
+      "range": "[2.5, 3.5)",
+      "docs": 1504383,
+      "tokens": 1622219261,
+      "pct": 28.91,
+      "target_pct": 35.0
+    }
+  },
+  "sd_stats": {
+    "mean": 0.849,
+    "std": 1.306,
+    "median": 0.0
+  },
+  "top_languages": [
+    {
+      "language": "typescript",
+      "tokens": 598255831,
+      "docs": 841426,
+      "pct_tokens": 15.33
+    },
+    {
+      "language": "python",
+      "tokens": 594491555,
+      "docs": 567721,
+      "pct_tokens": 15.24
+    },
+    {
+      "language": "github-issues-filtered-structured",
+      "tokens": 425892784,
+      "docs": 485384,
+      "pct_tokens": 10.92
+    },
+    {
+      "language": "markdown",
+      "tokens": 347112952,
+      "docs": 351728,
+      "pct_tokens": 8.9
+    },
+    {
+      "language": "json",
+      "tokens": 339833981,
+      "docs": 1124326,
+      "pct_tokens": 8.71
+    },
+    {
+      "language": "go",
+      "tokens": 332569389,
+      "docs": 73899,
+      "pct_tokens": 8.52
+    },
+    {
+      "language": "sql",
+      "tokens": 231050279,
+      "docs": 121035,
+      "pct_tokens": 5.92
+    },
+    {
+      "language": "javascript",
+      "tokens": 228398196,
+      "docs": 281216,
+      "pct_tokens": 5.85
+    },
+    {
+      "language": "yaml",
+      "tokens": 228221967,
+      "docs": 957872,
+      "pct_tokens": 5.85
+    },
+    {
+      "language": "java",
+      "tokens": 125354344,
+      "docs": 57787,
+      "pct_tokens": 3.21
+    },
+    {
+      "language": "c-sharp",
+      "tokens": 115640806,
+      "docs": 114063,
+      "pct_tokens": 2.96
+    },
+    {
+      "language": "html",
+      "tokens": 111255466,
+      "docs": 53527,
+      "pct_tokens": 2.85
+    },
+    {
+      "language": "c",
+      "tokens": 109359872,
+      "docs": 75899,
+      "pct_tokens": 2.8
+    },
+    {
+      "language": "haskell",
+      "tokens": 87567023,
+      "docs": 84862,
+      "pct_tokens": 2.24
+    },
+    {
+      "language": "rust",
+      "tokens": 26660301,
+      "docs": 12752,
+      "pct_tokens": 0.68
+    },
+    {
+      "language": "jupyter-scripts-dedup-filtered",
+      "tokens": 28403,
+      "docs": 11,
+      "pct_tokens": 0.0
+    }
+  ],
+  "slice_distribution": {
+    "schema_languages": {
+      "description": "Schema/data/query languages \u2014 inherently structured, light filter only",
+      "strategy": "light_filter",
+      "languages": [
+        "json",
+        "yaml",
+        "sql",
+        "protocol-buffer",
+        "thrift",
+        "xslt"
+      ],
+      "target_tokens": 800000000,
+      "actual_tokens": 799106227,
+      "docs": 2203233,
+      "pct_of_target": 99.9,
+      "pct_of_total": 20.5
+    },
+    "typescript": {
+      "description": "TypeScript \u2014 strong type system, filter by structured data relevance \u2265 2",
+      "strategy": "relevance_filter",
+      "languages": [
+        "typescript"
+      ],
+      "target_tokens": 600000000,
+      "actual_tokens": 598255831,
+      "docs": 841426,
+      "pct_of_target": 99.7,
+      "pct_of_total": 15.3
+    },
+    "python": {
+      "description": "Python \u2014 filter by structured data relevance \u2265 2",
+      "strategy": "relevance_filter",
+      "languages": [
+        "python"
+      ],
+      "target_tokens": 600000000,
+      "actual_tokens": 594491555,
+      "docs": 567721,
+      "pct_of_target": 99.1,
+      "pct_of_total": 15.2
+    },
+    "rust_go_java": {
+      "description": "Rust/Go/Java \u2014 strongly typed, filter by structured data relevance \u2265 2",
+      "strategy": "relevance_filter",
+      "languages": [
+        "rust",
+        "go",
+        "java"
+      ],
+      "target_tokens": 600000000,
+      "actual_tokens": 484584034,
+      "docs": 144438,
+      "pct_of_target": 80.8,
+      "pct_of_total": 12.4
+    },
+    "jupyter": {
+      "description": "Jupyter notebooks \u2014 filter by structured data relevance \u2265 2",
+      "strategy": "relevance_filter",
+      "languages": [
+        "jupyter-scripts-dedup-filtered"
+      ],
+      "target_tokens": 400000000,
+      "actual_tokens": 28403,
+      "docs": 11,
+      "pct_of_target": 0.0,
+      "pct_of_total": 0.0
+    },
+    "github_issues": {
+      "description": "GitHub issues (technical) \u2014 keyword filter for structured data topics",
+      "strategy": "keyword_filter",
+      "languages": [
+        "github-issues-filtered-structured"
+      ],
+      "target_tokens": 500000000,
+      "actual_tokens": 425892784,
+      "docs": 485384,
+      "pct_of_target": 85.2,
+      "pct_of_total": 10.9
+    },
+    "general_code": {
+      "description": "General code \u2014 random sample from all other languages for diversity",
+      "strategy": "light_filter",
+      "languages": [
+        "ada",
+        "agda",
+        "alloy",
+        "antlr",
+        "applescript",
+        "assembly",
+        "augeas",
+        "awk",
+        "batchfile",
+        "bluespec",
+        "c",
+        "c-sharp",
+        "clojure",
+        "cmake",
+        "coffeescript",
+        "common-lisp",
+        "cpp",
+        "css",
+        "cuda",
+        "dart",
+        "dockerfile",
+        "elixir",
+        "elm",
+        "emacs-lisp",
+        "erlang",
+        "f-sharp",
+        "fortran",
+        "glsl",
+        "groovy",
+        "haskell",
+        "html",
+        "idris",
+        "isabelle",
+        "java-server-pages",
+        "javascript",
+        "julia",
+        "jupyter-structured-clean-dedup",
+        "kotlin",
+        "lean",
+        "literate-agda",
+        "literate-coffeescript",
+        "literate-haskell",
+        "lua",
+        "makefile",
+        "maple",
+        "markdown",
+        "mathematica",
+        "matlab",
+        "ocaml",
+        "pascal",
+        "perl",
+        "php",
+        "powershell",
+        "prolog",
+        "r",
+        "racket",
+        "restructuredtext",
+        "rmarkdown",
+        "ruby",
+        "sas",
+        "scala",
+        "scheme",
+        "shell",
+        "smalltalk",
+        "solidity",
+        "sparql",
+        "stan",
+        "standard-ml",
+        "stata",
+        "systemverilog",
+        "tcl",
+        "tcsh",
+        "tex",
+        "verilog",
+        "vhdl",
+        "visual-basic",
+        "yacc",
+        "zig"
+      ],
+      "target_tokens": 1000000000,
+      "actual_tokens": 999334315,
+      "docs": 961295,
+      "pct_of_target": 99.9,
+      "pct_of_total": 25.6
+    }
+  },
+  "token_count_distribution": {
+    "p10": 55,
+    "p25": 111,
+    "p50": 255,
+    "p75": 631,
+    "p90": 1416,
+    "mean": 749
+  }
+}