[WIP] Upload folder using huggingface_hub (multi-commit ce81c354cbbc5d09d95e593286ea94557727334e788e6dd25f37cf8520dc1a4f)

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/dclm_crossdeduped/part.0.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:249a684b535dd5f7d425559f576a07fa5fa663fe06d0c4e5b65bea953477a319
-size 173518457

sample/100BT/dclm_crossdeduped/part.1.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f676c3c815ac1a61292e4bdefe0792d9ac92e0ec3253d05c65056dfd4eacaba6
-size 175420877

sample/100BT/dclm_crossdeduped/part.10.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a3a1dbaf577fbbc7dd6aee9df834e30f979932cf9ec25a5ed10cc7bd2f9d513b
-size 174990498

sample/100BT/dclm_crossdeduped/part.100.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d9031342990f69c9519ed11139e8007c6fe59fae77e59ab2a81c48c53865ae63
-size 172879381

sample/100BT/dclm_crossdeduped/part.101.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:350afda70dd1f8eb3c48e3da705f9ee54075e9e38243259ffa27b404628eda53
-size 174167679

sample/100BT/dclm_crossdeduped/part.102.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:edd9f23539111197565bb5e331768d5a687332fe8cdf18f45177f623e5017b1e
-size 177140924

sample/100BT/dclm_crossdeduped/part.103.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f9e316ca66e30b3510f9f6128b96c171e3ac1f3f74b1ee122a9a936c818d50ff
-size 181552253

sample/100BT/dclm_crossdeduped/part.104.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:71490982aac848516770a86fdbbe473c81b02871fd87a5ad216745c0e4e661f8
-size 180841997

sample/100BT/dclm_crossdeduped/part.105.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:515cf00e4cb893c0b173093226f7d302e2256279bcf4ffc5a74648da0aee807e
-size 179611946

sample/100BT/dclm_crossdeduped/part.106.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:91d2abae383b1dfd3a61078584811eb1c414a24d17a0aa56c23be1e302ebafeb
-size 173168670

sample/100BT/dclm_crossdeduped/part.107.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a0e684cd78c2ea3ac3aa9b4f6345c870ff0a8ce4a7d2b4adc757a6147187d890
-size 173241712

sample/100BT/dclm_crossdeduped/part.108.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0ecdcb6923f8d78d4748732c1162218aa570750c61d061828fec4ae46be7093d
-size 173623933

sample/100BT/dclm_crossdeduped/part.109.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:71432e50cb44bc29c8736b220d730697c75d665eed515c71ab4869dcac381ca4
-size 178155314

sample/100BT/dclm_crossdeduped/part.11.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:821ee538bc1884def47a15fa3157a2288859006ce71db7394b3ccd320da0c414
-size 182831725

sample/100BT/dclm_crossdeduped/part.110.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fa83f4907af3343300a34888b4240e921355faa00caa1f2bb67c1ff6959cef0d
-size 181103011

sample/100BT/dclm_crossdeduped/part.111.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:019cc6cc0e2729763e4867db6b1ec411ad897c14b50e6a5ee67c621175f956ab
-size 181377641

sample/100BT/dclm_crossdeduped/part.112.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bc71921b636792014a8bcab9d4fd23d6a4e7c81fbdc1d958e87da3dede630c06
-size 179165917

sample/100BT/dclm_crossdeduped/part.113.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0157fa64a4aeefc4a06724aa7cea21ef3f5af82e88e2641f247d8493e94d9cbf
-size 172870636

sample/100BT/dclm_crossdeduped/part.114.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3970c4de61b074dca26d80fc84130d9e0be797feb3f0f5f421b070cd94865e2a
-size 174861613

sample/100BT/dclm_crossdeduped/part.115.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:480e80aaeb187b415c8f8e947825b93eac4fed91c1e8409ec16bcb1ef47abd2c
-size 172485371

sample/100BT/dclm_crossdeduped/part.116.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ae4855e8a8cf7c514866bf37e64826516cd439ce7ce1aa54edaccc2046dd1fae
-size 174929250

sample/100BT/dclm_crossdeduped/part.117.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f2d8d9e6abaf5c45b161580d890d5f072b2d83443b1924fc7cbc1352a01d6d7f
-size 172018492

sample/100BT/dclm_crossdeduped/part.118.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b8bf99f1f6b66ea932685a5858a64605b9fe83cc33d125579e00bba01cda3135
-size 177621583

sample/100BT/dclm_crossdeduped/part.119.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:dbeddba6ade6d9d1b881719063c5ff10a81991f47f0b1b097aa4bf01073f1146
-size 178848599

sample/100BT/dclm_crossdeduped/part.12.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:78c3b60c9edf91528551e16a5778e56466c32f3c25663bc027164dabaa9f51ce
-size 177865734

sample/100BT/dclm_crossdeduped/part.120.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1273f75a1b00c35e7dc90ce54331585a269be7b01c47f60c02e513e5f6b8de20
-size 183002733

sample/100BT/dclm_crossdeduped/part.121.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:57f4f18ba75a88d900bf1ad16e15bc1578284605960c27a96bb00bc81d491134
-size 173078938

sample/100BT/dclm_crossdeduped/part.122.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:20a9cb1a680e3a6cc7326b352ceaf164ae70d49d3ee9d23887c26205bba86468
-size 169601210

sample/100BT/dclm_crossdeduped/part.123.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6d016babd1997b7333b8a82e46a7619a832a222ea03c425d363bc27829c15d0b
-size 172930128

sample/100BT/dclm_crossdeduped/part.124.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:84443d2b1baec6f82a69d7bdc11661b19ec965e41ed07cb2945249238d192237
-size 180649215

sample/100BT/dclm_crossdeduped/part.125.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:944e38d479a00512558055ecad7481ebd14b9886c3911376cd99bc5c5f0680b5
-size 184330169

sample/100BT/dclm_crossdeduped/part.126.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:dc2e0876d3da47cb036ed526564c8cb313645082c56d12d16acc13480935ef12
-size 177088039

sample/100BT/dclm_crossdeduped/part.127.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2ce846693392b5ce2bf794f6fc1141f0b73ea469396fd5e541409d9dbe1eb80a
-size 174284344

sample/100BT/dclm_crossdeduped/part.128.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:62cf0522acb5fa207880d2c2715601a5136bfe9ad9e1b2dd023190103c1877f2
-size 173235029

sample/100BT/dclm_crossdeduped/part.129.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6a20103aa1ed68aef2452a3488d6858dacd8e174aeac9fd3edc2ffb4f1ee14c8
-size 175019811

sample/100BT/dclm_crossdeduped/part.13.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e41e38c4babbbb6f477b9a493ebe146d8e3489ad0ac44666c50b0483bbfe366f
-size 183468211

sample/100BT/dclm_crossdeduped/part.130.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:76e741cfae21b6614b33f0a55bb24e79bfdada37312c5b51a775ec6642103ba2
-size 172423384

sample/100BT/dclm_crossdeduped/part.131.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d71110ece1164653996e9a1705529ebfb8ba624968f53a1611f39943d301fc3d
-size 180335334

sample/100BT/dclm_crossdeduped/part.132.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f9247834f606d2b7c0c173c6b280ac9d9f5f9dcb7cfe886089cb45890f1bdb90
-size 179793684

sample/100BT/dclm_crossdeduped/part.133.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:457a3071ee6a5c14a47ffa81a531f33c347d3c9fcbfc26aacb152243525910e7
-size 180347272

sample/100BT/dclm_crossdeduped/part.134.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bf91d12152b171e1db631a84a54907163ec8a183df1bd3e7831a5ad67dc858d6
-size 176039061

sample/100BT/dclm_crossdeduped/part.135.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:88573a6cb1b27336ab7642c3e7267cba4ffe9583a3495485f17b25ad238f0fd7
-size 172459663

sample/100BT/dclm_crossdeduped/part.136.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:704e452415ae8c4a335984c99df5a2d03981075d2b50dd2104b4ce2a1c3dbe51
-size 174386373

sample/100BT/dclm_crossdeduped/part.137.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b4b7119526891b4fa5cf700a28c34009c5d46778b74af4f772cbcc769732778f
-size 177922365

sample/100BT/dclm_crossdeduped/part.138.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bff4eecf77bd3367f1b64dbfecb387c94bea7eeb55df2c03c86f5f3f35b69538
-size 180340267

sample/100BT/dclm_crossdeduped/part.139.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2f3bff731dc5869b05314f90a73013f485a89715810fcba2dece9351e7e94edd
-size 182570883

sample/100BT/dclm_crossdeduped/part.14.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f015fd78c321e12a733bac60976c2c1f68ea5c6991a77b9dcbef842bb6a93bb7
-size 174698357

sample/100BT/dclm_crossdeduped/part.140.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f09b402307ba7a417649d37241e1252023cf73b710fc57d7155808cfeeb4446e
-size 178262870

sample/100BT/dclm_crossdeduped/part.141.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9737454edde0c81a6a62eaec3a07d791b4c04fe5fc6217a05d17d99feaba7c27
-size 173951502