Upload folder using huggingface_hub

.gitattributes

@@ -57,3 +57,4 @@ 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
+train.parquet.bak filter=lfs diff=lfs merge=lfs -text

CHANGELOG

@@ -0,0 +1,5 @@
+v.01 - Inital upload
+v.02 - Further deduping
+v.03 - ToS filtered. Added filters script repo
+v.04 - Fixed end tags and emoticons having missing leading spaces
+v.05 - Added dataset pipeline

README.md

@@ -0,0 +1,232 @@
+---
+tags:
+    - discord
+    - chatml
+    - conversation
+    - dialogue
+    - multi-turn
+    - single-turn
+    - fine-tuning
+    - reward-model
+    - llm-training
+    - chat-dataset
+    - open-source
+    - anonymized-data
+    - casual-dialogue
+license: apache-2.0
+language:
+    - en
+pretty_name: Discord-Dialogues
+size_categories:
+    - 1M<n<10M
+---
+
+<span style="font-size:smaller;">This is a clone of mookiezi/Discord-Dialogues.</span>
+
+<p align="center">
+  <img src="assests/Discord-Dialogues.png" alt="Discord-Dialogues">
+</p>
+
+> **Discord-Dialogues** is a large-scale dataset of anonymized Discord conversations from late spring to early fall 2025 for training and evaluating realistic conversational AI models in a ChatML-friendly format.
+
+  This dataset contains 7.5 million exchanges spread out over 17 million turns, with more than 145 million words. 
+
+---
+
+<p align="center">
+  <a href="https://atlas.nomic.ai/data/mookiezi/discord-alpha/map">
+    <img src="assests/discord-alpha.png" alt="discord-alpha">
+  </a>
+</p>
+
+<p align="center">
+  <a href="https://atlas.nomic.ai/data/mookiezi/discord-alpha/map"><strong>Nomic Atlas Map</strong></a>
+</p>
+
+---
+
+## Features
+
+-   Mixed single and multi-turn exchanges
+-   Human-only dialogues (no bots)
+-   Filtered for ToS and harmful content
+-   Links, embeds, and commands removed
+-   Trading posts, code blocks, and LFG removed
+-   Two-author chains only
+-   Merged self-replies from the same author into a single message
+-   Cleaned and deduplicated for relevance
+-   Primarily English, with some other languages present
+
+---
+
+## Use
+
+-   Fine-tuning conversational models
+-   Training relevance/reward models
+-   Dialogue generation research
+
+Use case examples:  
+- [mookiezi/Discord-Micae-8B-Preview](https://huggingface.co/mookiezi/Discord-Micae-8B-Preview) — experimental larger model  
+- [mookiezi/Discord-Micae-Hermes-3-3B](https://huggingface.co/mookiezi/Discord-Micae-Hermes-3-3B) — stable smaller model
+
+---
+
+## Filtering Pipeline
+
+This dataset was constructed with a custom multi-stage filtering toolkit:
+
+1. **SQL filters** (`filter.sql`)  
+  Postgres regex/text filters for PII, bot/command patterns, links, embeds, and automation noise.
+
+2. **Smart cleaner** (`smartclean.py`)  
+  Multi-stage process: normalize text, slang replacement, resample by length, and enforce structural validation.
+  Filters out structural noise such as code blocks, trading posts, and LFG.
+
+3. **Dedupe** (`dedupe.py`)
+  Deduplicates conversations by hashing message chains
+  Keeps only unique rows preferring the longest final assistant message when duplicates occur.
+
+4. **Fix End** (`fixend.py`)
+  Strips any prefix of spaces, commas, or non-emoticon colons before `<|im_end|>` to the plain token.
+
+5. **ToS risk filter** (`tos.py`)
+  Drops or redacts unsafe categories (sexual violence, CSA, slurs, harassment, doxxing, self-harm, extremism) and PII.
+  Uses fuzzy/leet/diacritic-aware regex.
+
+The full filtering scripts are open source at the [filters GitHub repository](https://github.com/mookiezi/filters).
+
+---
+
+## Dataset Pipeline
+
+The full end-to-end pipeline is documented in the [dataset-pipeline GitHub repository](https://github.com/mookiezi/dataset-pipeline).
+
+---
+
+## Collection Policy
+
+-   All data was collected adhering to Discord's [Terms of Service](https://discord.com/terms) and [Community Guidelines](https://discord.com/guidelines).
+
+---
+
+## Dataset Statistics <span style="font-weight:normal;">(using the [NousResearch/Hermes-3-Llama-3.1-8B tokenizer](https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B))</span>
+
+<div style="display:flex; gap:20px; align-items:flex-start;">
+
+<div>
+
+| Metric                   |           Value  |
+| ------------------------ | --------------:  |
+| Samples (count)          |        7,546,294 |
+| Min length (tokens)      |                7 |
+| Max length (tokens)      |            5,979 |
+| Mean length (tokens)     |            33.02 |
+| Median length (tokens)   |               29 |
+| Std dev (tokens)         |            17.39 |
+| Skew                     |            26.46 |
+| Kurtosis                 |         7,487.55 |
+| Total tokens             |      249,193,745 |
+| Total characters         |    1,291,480,299 |
+| Total words              |      145,887,976 |
+| Avg chars per sample     |           171.14 |
+| Avg words per sample     |            19.33 |
+| Avg chars per word       |             8.85 |
+| Tokens per char          |             0.19 |
+| Total assistant blocks   |        9,341,891 |
+
+</div>
+
+<div>
+
+
+| Tokens    |     Count  |
+| --------- | --------:  |
+| 0–8       |         1  |
+| 8–16      |    110,310 |
+| 16–32     |  4,382,094 |
+| 32–64     |  2,674,780 |
+| 64–128    |    360,401 |
+| 128–256   |     18,083 |
+| 256–384   |        417 |
+| 384–512   |         75 |
+| 512–768   |         78 |
+| 768–1024  |         30 |
+| 1024–2048 |         18 |
+| 2048–4096 |          3 |
+
+</div>
+
+<div>
+
+
+| Turns |   Count    |
+| ----- | --------:  |
+| 2     |  5,969,540 |
+| 3     |  1,080,526 |
+| 4     |    319,794 |
+| 5     |    102,553 |
+| 6     |     41,246 |
+| 7     |     16,904 |
+| 8     |      7,715 |
+| 9     |      3,691 |
+| 10    |      1,867 |
+| 11    |      1,007 |
+| 12    |        575 |
+| 13    |        334 |
+| 14    |        189 |
+| 15    |        129 |
+| 16    |         67 |
+| 17    |         62 |
+| 18    |         32 |
+| 19    |         21 |
+| 20    |          8 |
+| 21    |         11 |
+| 22    |         11 |
+| 23    |          2 |
+| 24    |          1 |
+| 25    |          3 |
+| 27    |          2 |
+| 29    |          1 |
+| 32    |          1 |
+| 33    |          2 |
+
+</div>
+
+</div>
+
+---
+
+## Disclaimer
+
+Although filtering reduced the exchanges by about 75% (leaving roughly 7.5% of the full data dump), this dataset is still intended as a large-scale dump. For best training results, further curation to target high-signal data relevant to your goals is recommended.
+
+---
+
+## License
+
+This project is licensed under the Apache License 2.0.
+
+---
+
+## How to cite:
+
+```bibtex
+@misc{discord-dialogues-2025,
+  title = {Discord-Dialogues},
+  author = {mookiezi},
+  year = {2025},
+  url={https://huggingface.co/datasets/mookiezi/Discord-Dialogues}
+}
+```
+
+---
+
+## Related
+
+-   [mookiezi/Discord-Micae-Hermes-3-3B](https://huggingface.co/mookiezi/Discord-Micae-Hermes-3-3B)
+-   [mookiezi/Discord-OpenMicae](https://huggingface.co/datasets/mookiezi/Discord-OpenMicae)
+-   [NousResearch/Hermes-3-Llama-3.2-8B](https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B)
+
+[​](https://20000.online/micae)
+[​](https://20000.online/openmicae)
+[​](https://20000.online/discord-dialogues)

dataset_infos.json

@@ -0,0 +1,29 @@
+{
+  "default": {
+    "description": "Discord-Dialogues is a large-scale dataset of anonymized Discord conversations formatted for ChatML. It includes mixed single- and multi-turn exchanges between two human participants, cleaned of bots, links, embeds, commands, ToS breaking content, and duplicate messages—primarily in English, suitable for fine-tuning conversational AI models.",
+    "citation": "@misc{discord-dialogues-2025,\n  title = {Discord-Dialogues},\n  author = {mookiezi},\n  year = {2025},\n  url = {https://huggingface.co/datasets/mookiezi/Discord-Dialogues}\n}",
+    "homepage": "https://huggingface.co/datasets/mookiezi/Discord-Dialogues",
+    "license": "Apache License 2.0",
+    "features": {
+      "text": { "dtype": "string", "_type": "Value" },
+      "tokens": { "dtype": "int64", "_type": "Value" },
+      "turns": { "dtype": "int64", "_type": "Value" },
+      "characters": { "dtype": "int64", "_type": "Value" },
+      "words": { "dtype": "int64", "_type": "Value" }
+    },
+    "splits": {
+      "train": {
+        "name": "train",
+        "num_bytes": 362022690,
+        "num_examples": 7546294,
+        "dataset_name": "default"
+      }
+    },
+    "download_size": 362022690,
+    "dataset_size": 362022690,
+    "size_in_bytes": 362022690,
+    "data_files": {
+      "train": [{ "filename": "train.parquet" }]
+    }
+  }
+}

tokens.log

@@ -0,0 +1,135 @@
+Stats for text:
+  min: 7
+  max: 5979
+  mean: 33.02200325086725
+  median: 29.0
+  std: 17.390580671916503
+  skew: 26.456841814125784
+  kurt: 7487.549682758939
+  count: 7546294
+  sum: 249193745
+  99.9%: 152.0
+  1%: 15.0
+  2%: 16.0
+  3%: 16.0
+  4%: 17.0
+  5%: 17.0
+  6%: 18.0
+  7%: 18.0
+  8%: 18.0
+  9%: 19.0
+  10%: 19.0
+  11%: 19.0
+  12%: 19.0
+  13%: 20.0
+  14%: 20.0
+  15%: 20.0
+  16%: 20.0
+  17%: 21.0
+  18%: 21.0
+  19%: 21.0
+  20%: 21.0
+  21%: 22.0
+  22%: 22.0
+  23%: 22.0
+  24%: 22.0
+  25%: 22.0
+  26%: 23.0
+  27%: 23.0
+  28%: 23.0
+  29%: 23.0
+  30%: 24.0
+  31%: 24.0
+  32%: 24.0
+  33%: 24.0
+  34%: 24.0
+  35%: 25.0
+  36%: 25.0
+  37%: 25.0
+  38%: 25.0
+  39%: 26.0
+  40%: 26.0
+  41%: 26.0
+  42%: 26.0
+  43%: 27.0
+  44%: 27.0
+  45%: 27.0
+  46%: 27.0
+  47%: 28.0
+  48%: 28.0
+  49%: 28.0
+  50%: 29.0
+  51%: 29.0
+  52%: 29.0
+  53%: 29.0
+  54%: 30.0
+  55%: 30.0
+  56%: 30.0
+  57%: 31.0
+  58%: 31.0
+  59%: 31.0
+  60%: 32.0
+  61%: 32.0
+  62%: 32.0
+  63%: 33.0
+  64%: 33.0
+  65%: 34.0
+  66%: 34.0
+  67%: 34.0
+  68%: 35.0
+  69%: 35.0
+  70%: 36.0
+  71%: 36.0
+  72%: 37.0
+  73%: 37.0
+  74%: 38.0
+  75%: 38.0
+  76%: 39.0
+  77%: 39.0
+  78%: 40.0
+  79%: 41.0
+  80%: 42.0
+  81%: 42.0
+  82%: 43.0
+  83%: 44.0
+  84%: 45.0
+  85%: 46.0
+  86%: 47.0
+  87%: 48.0
+  88%: 49.0
+  89%: 51.0
+  90%: 52.0
+  91%: 54.0
+  92%: 56.0
+  93%: 58.0
+  94%: 60.0
+  95%: 64.0
+  96%: 68.0
+  97%: 73.0
+  98%: 80.0
+  99%: 95.0
+  100%: 5979.0
+  total_chars: 1291480299
+  total_words: 145887976
+  avg_chars: 171.14099967480726
+  avg_words: 19.332400248386826
+  avg_chars_per_word: 8.852547923483426
+  avg_chars_per_sample: 171.14099967480726
+  avg_words_per_sample: 19.332400248386826
+  tokens_per_char: 0.19295202969255673
+  bin_0-8: 1
+  bin_8-16: 110310
+  bin_16-32: 4382094
+  bin_32-64: 2674780
+  bin_64-128: 360401
+  bin_128-256: 18083
+  bin_256-384: 417
+  bin_384-512: 75
+  bin_512-768: 78
+  bin_768-1024: 30
+  bin_1024-2048: 18
+  bin_2048-4096: 3
+  assistant_blocks: 9341891
+
+Total tokens across all columns: 249193745
+Total assistant blocks: 9341891

tokenstats.txt

@@ -0,0 +1,135 @@
+Stats for text:
+  min: 8
+  max: 5979
+  mean: 33.024094354044
+  median: 29.0
+  std: 17.402215788976577
+  skew: 26.64223385581348
+  kurt: 7506.831274998159
+  count: 7543842
+  sum: 249128550
+  99.9%: 153.0
+  1%: 15.0
+  2%: 16.0
+  3%: 16.0
+  4%: 17.0
+  5%: 17.0
+  6%: 18.0
+  7%: 18.0
+  8%: 18.0
+  9%: 19.0
+  10%: 19.0
+  11%: 19.0
+  12%: 19.0
+  13%: 20.0
+  14%: 20.0
+  15%: 20.0
+  16%: 20.0
+  17%: 21.0
+  18%: 21.0
+  19%: 21.0
+  20%: 21.0
+  21%: 22.0
+  22%: 22.0
+  23%: 22.0
+  24%: 22.0
+  25%: 22.0
+  26%: 23.0
+  27%: 23.0
+  28%: 23.0
+  29%: 23.0
+  30%: 24.0
+  31%: 24.0
+  32%: 24.0
+  33%: 24.0
+  34%: 24.0
+  35%: 25.0
+  36%: 25.0
+  37%: 25.0
+  38%: 25.0
+  39%: 26.0
+  40%: 26.0
+  41%: 26.0
+  42%: 26.0
+  43%: 27.0
+  44%: 27.0
+  45%: 27.0
+  46%: 27.0
+  47%: 28.0
+  48%: 28.0
+  49%: 28.0
+  50%: 29.0
+  51%: 29.0
+  52%: 29.0
+  53%: 29.0
+  54%: 30.0
+  55%: 30.0
+  56%: 30.0
+  57%: 31.0
+  58%: 31.0
+  59%: 31.0
+  60%: 32.0
+  61%: 32.0
+  62%: 32.0
+  63%: 33.0
+  64%: 33.0
+  65%: 34.0
+  66%: 34.0
+  67%: 34.0
+  68%: 35.0
+  69%: 35.0
+  70%: 36.0
+  71%: 36.0
+  72%: 37.0
+  73%: 37.0
+  74%: 38.0
+  75%: 38.0
+  76%: 39.0
+  77%: 39.0
+  78%: 40.0
+  79%: 41.0
+  80%: 42.0
+  81%: 42.0
+  82%: 43.0
+  83%: 44.0
+  84%: 45.0
+  85%: 46.0
+  86%: 47.0
+  87%: 48.0
+  88%: 49.0
+  89%: 51.0
+  90%: 52.0
+  91%: 54.0
+  92%: 56.0
+  93%: 58.0
+  94%: 60.0
+  95%: 64.0
+  96%: 68.0
+  97%: 73.0
+  98%: 80.0
+  99%: 95.0
+  100%: 5979.0
+  total_chars: 1290998934
+  total_words: 145717457
+  avg_chars: 171.13281720375375
+  avg_words: 19.316080188317837
+  avg_chars_per_word: 8.85960378789756
+  avg_chars_per_sample: 171.13281720375375
+  avg_words_per_sample: 19.316080188317837
+  tokens_per_char: 0.19297347460087058
+  bin_0-8: 0
+  bin_8-16: 109538
+  bin_16-32: 4381031
+  bin_32-64: 2674243
+  bin_64-128: 360330
+  bin_128-256: 18072
+  bin_256-384: 418
+  bin_384-512: 78
+  bin_512-768: 77
+  bin_768-1024: 30
+  bin_1024-2048: 17
+  bin_2048-4096: 4
+  assistant_blocks: 9339690
+
+Total tokens across all columns: 249128550
+Total assistant blocks: 9339690

train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe52210778f5d71724664d9c365a71599913f93c30cc60df8e674dd3c45c08ca
+size 362018517