Upload folder using huggingface_hub

README.md

@@ -12,10 +12,6 @@ configs:
     data_files:
       - split: train
         path: data/*/*/*
-  - config_name: sample-100BT
-    data_files:
-      - split: train
-        path: sample/100BT/*/*
   - config_name: dclm_crossdeduped
     data_files:
       - split: train
@@ -55,45 +51,21 @@ According to our evaluations, Zyda-2 is the most performant per-token open datas
 For more information, please see our [technical blog](https://www.zyphra.com/post/building-zyda-2).
 
 ## How to download
-We preserved the schemas of original component datasets, meaning that every component has its own schema. For that reason attempting to download the whole dataset using `datasets.load_dataset()` will fail during the stage of generating a split. If you attempt to stream the default config, it will also fail.
+Since we preserved the schemas of original component datasets, attempting to download the whole dataset using `datasets.load_dataset()` might fail during the stage of generating a split.
 
 To download the whole dataset we recommend to either clone the repository, or, if you must use the `datasets.load_dataset()`, download individual components separately.
 
-Only `nemo_id` and `text` are common columns between the components. Select those for every component first, and only then interleave the datasets with optimal weights (see example at the bottom of this section).
-
 Example command to clone the repository using huggingface-cli: `huggingface-cli download Zyphra/Zyda-2 --repo-type dataset`
 
 Commands to download individual components:
- - DCLM:  `ds_dclm = datasets.load_dataset("Zyphra/Zyda-2", name="dclm_crossdeduped", split="train")`
- - Zyda:  `ds_zyda = datasets.load_dataset("Zyphra/Zyda-2", name="zyda_crossdeduped-filtered", split="train")`
- - Dolma-CC:  `ds_dolma = datasets.load_dataset("Zyphra/Zyda-2", name="dolma-cc_crossdeduped-filtered", split="train")`
- - Fineweb-Edu:  `ds_fwe = datasets.load_dataset("Zyphra/Zyda-2", name="fwe3", split="train")`
+ - DCLM:  `ds = datasets.load_dataset("Zyphra/Zyda-2", name="dclm_crossdeduped", split="train")`
+ - Zyda:  `ds = datasets.load_dataset("Zyphra/Zyda-2", name="zyda_crossdeduped-filtered", split="train")`
+ - Dolma-CC:  `ds = datasets.load_dataset("Zyphra/Zyda-2", name="dolma-cc_crossdeduped-filtered", split="train")`
+ - Fineweb-Edu:  `ds = datasets.load_dataset("Zyphra/Zyda-2", name="fwe3", split="train")`
 
 In this repository we provide raw results of cross deduplication and filtering. To achieve the best possible performance, one will need to use appropriate weights during training.
-We found the following optimal weights by number of tokens (in the sense of weights in the resultant dataset): DCLM - 4.0, FWE3 - 4.0, Zyda - 0.16, Dolma-CC - 0.24.
-
-Below you will find an example of how to get proper dataset object.
-It demonstrates how to select only `nemo_id` and `text` columns, and then interleave the datasets with probabilities computed from the weights above.
-One needs to be careful with weights normalization, as `interleave_datasets()` returns documents, while our weights are token-wise. We provide precomputed document-wise weights in the example below.
-To stream the dataset, add `streaming=True` to the `load_dataset()` commands.
-
-```
-common_columns = ["nemo_id", "text"]
-ds_dclm = ds_dclm.select_columns(common_columns)
-ds_zyda = ds_zyda.select_columns(common_columns)
-ds_dolma = ds_dolma.select_columns(common_columns)
-ds_fwe = ds_zyda.select_columns(common_columns)
-norm_weights = [0.4038, 0.0316, 0.0585, 0.5061]
-ds = datasets.interleave_datasets([ds_dclm, ds_zyda, ds_dolma, ds_fwe], probabilities=norm_weights, stopping_strategy="all_exhausted")
-```
-
-### (Smaller) sample version
-Along with the configs above, you can also download a smaller version of the dataset with the following config:
-- `sample-100BT`: a subset randomly sampled from the whole dataset of around 100B gpt-neox tokens (252GB, 91.2M documents).
-
-This sample only has common columns `nemo-id` and `text`. In addition, it was sampled according to optimal weights, so you can start using it directly.
+We found the following optimal weights (in the sense of weights in the resultant dataset): DCLM - 4.0, FWE3 - 4.0, Zyda - 0.16, Dolma-CC - 0.24.
 
-`ds_sample = datasets.load_dataset("Zyphra/Zyda-2", name="sample-100BT", split="train")`
 
 ## Breakdown by component
 

sample/100BT/dolma_cc-crossdeduped-filtered/part.0.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:789f3e073ad5c89a43e65da564ce80726ae4b718579e6cf76e7dd39836bc4e23
-size 164068717

sample/100BT/dolma_cc-crossdeduped-filtered/part.1.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9c2d27c54b6e7b18bf4bfbb7b3735f09105aac44678d304245a58f71207dcb2f
-size 164427826

sample/100BT/dolma_cc-crossdeduped-filtered/part.10.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:94760d58cecfa8aa965cf5df25a280b1e38d6e8288f2192b2995366d4f9cb2de
-size 165480354

sample/100BT/dolma_cc-crossdeduped-filtered/part.11.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1db18e89a45796119d1a4d8bf8992fad901458a0cb3e2f5bc754a87afc492d64
-size 164177699

sample/100BT/dolma_cc-crossdeduped-filtered/part.12.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f770bcea60ed2b730bc72b65de9b6324897cdf09da669a2aa7a905d14180e5a7
-size 166038057

sample/100BT/dolma_cc-crossdeduped-filtered/part.13.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c88ba6c14703102b0791cf5a75d27b9ac453ba45d6e72859dd26e07f99e29f4d
-size 166263781

sample/100BT/dolma_cc-crossdeduped-filtered/part.14.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:343b318501f8dcfc290632b5d0ccae63673f9d4727d6d3dd16746843f087f088
-size 166714793

sample/100BT/dolma_cc-crossdeduped-filtered/part.15.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:78e3151993b96780e745007b07e096f1fd4b4255be24588f0289de78511542d5
-size 163741403

sample/100BT/dolma_cc-crossdeduped-filtered/part.16.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7bb88c162e9121b9a0da79aae6c63172414ecd2b30a20ddd0a4ff75e49af7586
-size 161875767

sample/100BT/dolma_cc-crossdeduped-filtered/part.17.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:93e9a42d92da00086a3a563ef4d0c034833d79043612e0d47b6dad764ce4bc70
-size 163441204

sample/100BT/dolma_cc-crossdeduped-filtered/part.18.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c674b73cc76f3e8affb103a9a156a013393e1940ab4f369d2ff5677e088cdc9b
-size 162643378

sample/100BT/dolma_cc-crossdeduped-filtered/part.19.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4890bf18a12192c07ef14b7aad7d7d95315cbe299c50aac5049a0584a99396cb
-size 161097893

sample/100BT/dolma_cc-crossdeduped-filtered/part.2.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2cc9dcfba09b757988b12aa6b9e09f4f41162388ffba71a31db955901279e612
-size 166949993

sample/100BT/dolma_cc-crossdeduped-filtered/part.20.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:cac65dc0d15cbfada40400d7360926d78d3c22718bbb67cac87012977237c9a8
-size 174137790

sample/100BT/dolma_cc-crossdeduped-filtered/part.21.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:33d43cf6d8c585faa8309e6546cf43f0778195b97fc1fd364fbd89ea80289b4d
-size 178097901

sample/100BT/dolma_cc-crossdeduped-filtered/part.22.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:dbbe4a87df25a5bb083ca809a7554bb3a6621d07b19da684dc6b5ce4bcf522a0
-size 166533119

sample/100BT/dolma_cc-crossdeduped-filtered/part.23.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5308a186cf878966a57e700a2679fcf27555468779967afe536ff1dfe636f2b7
-size 166037002

sample/100BT/dolma_cc-crossdeduped-filtered/part.24.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4775db252f536b23c9c0f0f034ce1643bbd8cf7dc825b8c3aae4ca7ed8f2b1cc
-size 170309163

sample/100BT/dolma_cc-crossdeduped-filtered/part.25.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6c281a967377903c1375b0b856c53dc1dd0cf330abba9f4df2609d565494315e
-size 178279658

sample/100BT/dolma_cc-crossdeduped-filtered/part.26.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6d90e5ee9dcaab22d69df9f1a8ac0c947acff83104b359e2b1d96adb4f5b32d1
-size 177612667

sample/100BT/dolma_cc-crossdeduped-filtered/part.27.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:761ec1bfb18352cd87ab16d4aca16c52ea5cdd3f65f3890c4150e8a88202a3da
-size 176501635

sample/100BT/dolma_cc-crossdeduped-filtered/part.28.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bf5ab684faf0e50defc9d70821d27ebb2a42a92b7a5120ec3c42eeba449a67b2
-size 172251331

sample/100BT/dolma_cc-crossdeduped-filtered/part.29.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:46ca44effb6576251118005e3921581ea81a0cb2b4044e51676f8e7701d3317f
-size 169207830

sample/100BT/dolma_cc-crossdeduped-filtered/part.3.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:27a6d6834f81c1a7e0f6ac10b1a6b1ad4e93a2a87780d58ffbf897ae7f37559c
-size 165493788

sample/100BT/dolma_cc-crossdeduped-filtered/part.30.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:06030cbac206281b20c925b718e1088a74f489be6b450a82c5cfc3d9f2179d37
-size 167113421

sample/100BT/dolma_cc-crossdeduped-filtered/part.31.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:44b9806306e3f9f0ac05adb1f0313ecbc06811fd5d26191903d93cb7dfc4a061
-size 165132529

sample/100BT/dolma_cc-crossdeduped-filtered/part.32.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2194323b6ec6b68825f7083065a4a819cc4501fb9c88c619f595cbdba2e8db34
-size 166202797

sample/100BT/dolma_cc-crossdeduped-filtered/part.33.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9e88527e73f3d59500b33f5571707c08dce3fb1c1b8a87b4555ba55167808eb9
-size 165992349

sample/100BT/dolma_cc-crossdeduped-filtered/part.34.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:be3020ab05294ad31415211646719e0d02f9419248aa1f2f4d12045ca0344692
-size 165429234

sample/100BT/dolma_cc-crossdeduped-filtered/part.35.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:efaa4368422d4928d3f5fd45c3ca4017597d1c573e31dd881af245a980c1d8c1
-size 166092761

sample/100BT/dolma_cc-crossdeduped-filtered/part.36.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0db49302490e1a6a41ba8db805db0b2f84bab9398eed7e2d52f07e7b9b7fe64a
-size 166732896

sample/100BT/dolma_cc-crossdeduped-filtered/part.37.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:badc7ea755621e5c3d95fbcfd47b53b8ef7ac0b94ab42d4996699e839f4ca386
-size 165863882

sample/100BT/dolma_cc-crossdeduped-filtered/part.38.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d6910df5353c00f9abeb14d58cf98dd8ece3b76f5616d4a09071bc0c500bf74b
-size 164324078

sample/100BT/dolma_cc-crossdeduped-filtered/part.39.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b15cd059bb056632e596136ed0ea3ee6faa69448625f85cc4afeb164e803050a
-size 161409078

sample/100BT/dolma_cc-crossdeduped-filtered/part.4.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c4494626b5dd1e23d6022e9ca963bc8a30ed1a28b378b7678c29d15f28e1bbcb
-size 163812825

sample/100BT/dolma_cc-crossdeduped-filtered/part.40.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:76c34447cdb93fa81d6405c23efd3e2c888edcf1071c05c443581f93f7b94a72
-size 161107115

sample/100BT/dolma_cc-crossdeduped-filtered/part.41.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:67437095d61fbc26d17ada6636053e790b9a6a7560e3b5cc60f416ffbbb999fc
-size 163006291

sample/100BT/dolma_cc-crossdeduped-filtered/part.42.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d40c8098e722d80cffffa638d11a64494ac0bcc3fb9ce350e675f4707cfafc0b
-size 161669195

sample/100BT/dolma_cc-crossdeduped-filtered/part.43.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:afa3609f016f1931982e0103a70e481e26ca25a1f1277142811e8b3e2f3835b1
-size 159251808

sample/100BT/dolma_cc-crossdeduped-filtered/part.44.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:194d37a41f758f5034297de0b908bae1d296844e29058cabc20b8f35016b23a4
-size 162762875

sample/100BT/dolma_cc-crossdeduped-filtered/part.45.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fd53d3e95c10021a023f539e1d72664bede4c03120ce98146a6466c80c427377
-size 163459358

sample/100BT/dolma_cc-crossdeduped-filtered/part.46.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:93993f15543c99fe3bc652ef0558649589627e1db8c2b75edea415a3cb49c010
-size 163617527

sample/100BT/dolma_cc-crossdeduped-filtered/part.47.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b102d8befb6d306bbf6558039942a89a40c2c26f85a25943be4a0247371aaf70
-size 166077046

sample/100BT/dolma_cc-crossdeduped-filtered/part.48.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:98d88f521f9bdb2de0b441cb9a8c23897c8a7678d38e1969fb609ad97b8b21b4
-size 150111428

sample/100BT/dolma_cc-crossdeduped-filtered/part.5.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b7ed9b0c1c6173d21cffe6bbf9d0d0878463be68cc2dc1b64bc75ceb6101614d
-size 167144148

sample/100BT/dolma_cc-crossdeduped-filtered/part.6.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:02b24970bb4421b174500a4fc791bc776a3c9ed9cb9e782b558ea98ba922d36f
-size 166136261

sample/100BT/dolma_cc-crossdeduped-filtered/part.7.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3d0d63e915a7699e49e41e497374a70b93d297e5fb6830383e27a6eb74438306
-size 165133123

sample/100BT/dolma_cc-crossdeduped-filtered/part.8.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a59d6e380ced282fb800c03519497ed49a2857a9e7e04749772b490fe4f1a831
-size 166424467

sample/100BT/dolma_cc-crossdeduped-filtered/part.9.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7b31500086bac168e90b83254f53b362b0eef953440be83871c910acab72e0bd
-size 166602443