victormiller commited on
Commit
497eac9
1 Parent(s): 64d8614

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -3
README.md CHANGED
@@ -52,15 +52,16 @@ Curated datasets are typically structured and consistently formatted. TxT360 fil
52
  | Europarl | 6.1 GB | 1.96B | - |
53
  | StackExchange | 45 GB | 8.37B | Q4 2023 |
54
 
 
55
 
56
  ## CommonCrawl Data Filtering
57
- Follow [this link](https://llm360-txt360-new.hf.space/webdata#section1) to view all steps taken to filter the web data.
58
 
59
  ## Curated Source Filtering
60
- Each data source was filtered individually with respect to the underlying data. Full details and discussion on how each source is filter is covered [here](https://llm360-txt360-new.hf.space/curated#section1).
61
 
62
  ## Global Deduplication
63
- After the web and curated sources were filtered, they were globally deduplicated to create TxT360. The deduplication process is available [here](https://llm360-txt360-new.hf.space/common#section2).
64
 
65
  # Citation
66
 
 
52
  | Europarl | 6.1 GB | 1.96B | - |
53
  | StackExchange | 45 GB | 8.37B | Q4 2023 |
54
 
55
+ The [TxT360](https://huggingface.co/spaces/LLM360/TxT360) blog post provides all the details behind how we approached and implemented the following features:
56
 
57
  ## CommonCrawl Data Filtering
58
+ Complete discussion on how 99 Common Crawl snapshots were filtered and comparison to previous filtering techinques (e.g. Dolma, DataTrove, RedPajamaV2).
59
 
60
  ## Curated Source Filtering
61
+ Each data source was filtered individually with respect to the underlying data. Full details and discussion on how each source was filter are covered.
62
 
63
  ## Global Deduplication
64
+ After the web and curated sources were filtered, all sources globally deduplicated to create TxT360. The tips and tricks behind the deduplication process are included.
65
 
66
  # Citation
67