| --- |
| license: cc-by-sa-4.0 |
| language: |
| - en |
| pretty_name: SN22 sandbox embed texts |
| size_categories: |
| - 10K<n<100K |
| tags: |
| - embeddings |
| - bittensor |
| - desearch |
| configs: |
| - config_name: default |
| data_files: data/*.parquet |
| --- |
| |
| # SN22 sandbox embed texts |
|
|
| Texts and reference embeddings for testing a [Desearch Subnet 22](https://github.com/Desearch-ai/subnet-22) |
| embed miner locally, before it takes tasks on mainnet. |
|
|
| Run `python -m sandbox embed` in the subnet-22 repository. It downloads these files, queues the texts |
| as embed tasks in a local task API, and runs a local validator that checks your miner's embeddings |
| against the reference vectors here, the way mainnet validators check embed tasks. Point your miner at the |
| local API and see each verdict, the texts that did not match and by how much, and how fast your GPU |
| gets through a task. The [guide](https://github.com/Desearch-ai/subnet-22/blob/develop/docs/test-your-miner.md) |
| has the steps. |
|
|
| ## Dataset details |
|
|
| | | | |
| |---|---| |
| | Texts | 14,226: 3,000 page texts and 11,226 passages | |
| | Pages | 3,000 English Wikipedia articles, sampled at random | |
| | Characters | 15.6 million | |
| | Reference model | `desearch-4b-v1`: [desearch/Desearch-Embedding-4B](https://huggingface.co/desearch/Desearch-Embedding-4B) at revision `12e5c6910d49dd5534e8a77a6fa7d8a057bd534a` | |
| | Reference vectors | 2,560 dimensions, float32, L2-normalized | |
| | Files | `data/texts-NNNNN.parquet`, rows grouped by page; `manifest.json` lists them | |
| | Source | [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia), snapshot `20231101.en` | |
| | License | [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/) | |
|
|
| ## Columns |
|
|
| | Column | Type | | |
| |---|---|---| |
| | `text_id` | string | the text's id within its page: `<page_key>#page` or `<page_key>#p<n>` | |
| | `page_key` | string | the page the text was taken from | |
| | `url` | string | the article's Wikipedia URL | |
| | `content_sha1` | string | the version of the page the text was taken from | |
| | `kind` | string | `page` for a page's own text, `passage` for a passage of a long article | |
| | `index` | int32 | 0 for the page text, 1, 2, … for its passages | |
| | `text` | large string | exactly what is embedded | |
| | `reference` | binary | the reference vector: 2,560 little-endian float32 values | |
|
|
| The first seven columns are the columns of a mainnet embed task's input, so the sandbox serves these rows to |
| miners unchanged. |
|
|
| ## How the texts were made |
|
|
| Each article becomes texts the same way a crawled page does on mainnet, with the same code: |
|
|
| - **Page text:** the title, then the article's text, cut at 10,000 characters. |
| - **Passages:** an article longer than 2,000 characters also gives up to 16 passages of 600 to 1,500 |
| characters of consecutive paragraphs, split at its section headings, each starting with the article's title |
| and the section's heading. |
|
|
| Random articles give the mix of short and long pages a real task has. |
|
|
| ## How the reference vectors were computed |
|
|
| The reference vectors come from the exact model in full precision: vLLM 0.31.0 with `--runner pooling |
| --dtype float32 --no-enable-chunked-prefill --max-num-batched-tokens 16384`, attention kernels in float32 |
| too (`TRITON_F32_DEFAULT=ieee`). Each text was sent as is, with no prompt, and cut at 2,048 tokens |
| (`truncate_prompt_tokens: 2048`). |
|
|
| The model served as miners run it, in float16 with the documented command, matches these vectors with a |
| cosine similarity of at least 0.99998 on a random sample of 500 texts. Mainnet's check asks for 0.999. |
|
|
| ## Loading |
|
|
| ```python |
| import numpy as np |
| from datasets import load_dataset |
| |
| rows = load_dataset("desearch/sn22-sandbox-embed", split="train") |
| row = rows[0] |
| reference = np.frombuffer(row["reference"], dtype="<f4") # shape (2560,) |
| print(row["text_id"], row["kind"], len(row["text"]), reference.shape) |
| ``` |
|
|
| ## Source and license |
|
|
| The texts are adapted from Wikipedia articles by Wikipedia contributors, through the |
| [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) dataset (snapshot `20231101.en`). |
| Like Wikipedia, this dataset is licensed under |
| [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). |
|
|
