Finalize native Hub loading documentation
Browse files
README.md
CHANGED
|
@@ -142,7 +142,7 @@ The same direct form applies to `samples/dedup_bucket=<bucket>/*.parquet` and
|
|
| 142 |
|
| 143 |
`revision=` is optional for ordinary use because `main` is the current release.
|
| 144 |
For reproducible training, pass the immutable data commit shown below as
|
| 145 |
-
`revision="
|
| 146 |
`@<revision>` to an `hf://` URL. The manifest run ID remains inside
|
| 147 |
`manifest.json` as provenance; it is not part of the Hub path.
|
| 148 |
|
|
@@ -156,7 +156,8 @@ dataset membership.
|
|
| 156 |
|
| 157 |
## Publication verification
|
| 158 |
|
| 159 |
-
Data commit `
|
|
|
|
| 160 |
The published tree was checked as 3,293 data objects totaling
|
| 161 |
41,557,605,070 bytes: 3,292 Parquet shards plus `manifest.json`. Every local
|
| 162 |
object size and SHA-256 matched its remote Hub object. Pinned streaming loads
|
|
|
|
| 142 |
|
| 143 |
`revision=` is optional for ordinary use because `main` is the current release.
|
| 144 |
For reproducible training, pass the immutable data commit shown below as
|
| 145 |
+
`revision="4b9c79918adcb426fa7d0b27056fcde9366357af"` to `load_dataset`, or add
|
| 146 |
`@<revision>` to an `hf://` URL. The manifest run ID remains inside
|
| 147 |
`manifest.json` as provenance; it is not part of the Hub path.
|
| 148 |
|
|
|
|
| 156 |
|
| 157 |
## Publication verification
|
| 158 |
|
| 159 |
+
Data commit `4b9c79918adcb426fa7d0b27056fcde9366357af` is the immutable revision for this
|
| 160 |
+
release.
|
| 161 |
The published tree was checked as 3,293 data objects totaling
|
| 162 |
41,557,605,070 bytes: 3,292 Parquet shards plus `manifest.json`. Every local
|
| 163 |
object size and SHA-256 matched its remote Hub object. Pinned streaming loads
|