Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
validation_index: int64
domain: string
source: string
shard_index: int64
sample_index: int64
global_sample_index: int64
document_index: int64
document_chunk_index: int64
validation_metadata_sha256: string
dtype: string
validation: struct<path: string, tokens: int64, sequences: int64, sha256: string, usable_tokens: int64, discarde (... 21 chars omitted)
child 0, path: string
child 1, tokens: int64
child 2, sequences: int64
child 3, sha256: string
child 4, usable_tokens: int64
child 5, discarded_tail_tokens: int64
header_bytes: int64
magic: int64
validation_metadata_path: string
version: int64
format: string
endianness: string
sequence_length: int64
total_validation_usable_tokens: int64
total_sequences: int64
domain_totals: struct<thestackv1_concat_by_repo-65536: struct<sequences: int64>, book-65536: struct<sequences: int6 (... 40 chars omitted)
child 0, thestackv1_concat_by_repo-65536: struct<sequences: int64>
child 0, sequences: int64
child 1, book-65536: struct<sequences: int64>
child 0, sequences: int64
child 2, textbooks: struct<sequences: int64>
child 0, sequences: int64
tokenizer: struct<name: string, revision: string, vocab_size: int64, eos_token_id: int64>
child 0, name: string
child 1, revision: string
child 2, vocab_size: int64
child 3, eos_token_id: int64
validation_source: struct<dataset: string, revision: string, source_tokenizer: struct<name: string, revision: string>, (... 1071 chars omitted)
child 0, dataset: strin
...
ars omitted)
child 0, thestackv1_concat_by_repo-65536: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 1, book-65536: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 2, textbooks: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 10, converter_versions: struct<mosaicml-streaming: string, tokenizers: string, transformers: string>
child 0, mosaicml-streaming: string
child 1, tokenizers: string
child 2, transformers: string
child 11, script_sha256: string
total_validation_tokens: int64
to
{'format': Value('string'), 'tokenizer': {'name': Value('string'), 'revision': Value('string'), 'vocab_size': Value('int64'), 'eos_token_id': Value('int64')}, 'dtype': Value('string'), 'endianness': Value('string'), 'header_bytes': Value('int64'), 'magic': Value('int64'), 'version': Value('int64'), 'sequence_length': Value('int64'), 'validation': {'path': Value('string'), 'tokens': Value('int64'), 'sequences': Value('int64'), 'sha256': Value('string'), 'usable_tokens': Value('int64'), 'discarded_tail_tokens': Value('int64')}, 'validation_metadata_path': Value('string'), 'validation_metadata_sha256': Value('string'), 'total_validation_tokens': Value('int64'), 'total_validation_usable_tokens': Value('int64'), 'total_sequences': Value('int64'), 'domain_totals': {'thestackv1_concat_by_repo-65536': {'sequences': Value('int64')}, 'book-65536': {'sequences': Value('int64')}, 'textbooks': {'sequences': Value('int64')}}, 'validation_source': {'dataset': Value('string'), 'revision': Value('string'), 'source_tokenizer': {'name': Value('string'), 'revision': Value('string')}, 'target_tokenizer': {'name': Value('string'), 'revision': Value('string')}, 'split_special_tokens': Value('bool'), 'selection_seed': Value('int64'), 'selection': Value('string'), 'excluded_training': {'manifest_sha256': Value('string'), 'unique_source_samples': {'thestackv1_concat_by_repo-65536': Value('int64'), 'book-65536': Value('int64'), 'textbooks': Value('int64')}}, 'domain_sequence_counts': {'thestackv1_concat_by_repo-65536': Value('int64'), 'book-65536': Value('int64'), 'textbooks': Value('int64')}, 'stats': {'thestackv1_concat_by_repo-65536': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}, 'book-65536': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}, 'textbooks': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}}, 'converter_versions': {'mosaicml-streaming': Value('string'), 'tokenizers': Value('string'), 'transformers': Value('string')}, 'script_sha256': Value('string')}}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
validation_index: int64
domain: string
source: string
shard_index: int64
sample_index: int64
global_sample_index: int64
document_index: int64
document_chunk_index: int64
validation_metadata_sha256: string
dtype: string
validation: struct<path: string, tokens: int64, sequences: int64, sha256: string, usable_tokens: int64, discarde (... 21 chars omitted)
child 0, path: string
child 1, tokens: int64
child 2, sequences: int64
child 3, sha256: string
child 4, usable_tokens: int64
child 5, discarded_tail_tokens: int64
header_bytes: int64
magic: int64
validation_metadata_path: string
version: int64
format: string
endianness: string
sequence_length: int64
total_validation_usable_tokens: int64
total_sequences: int64
domain_totals: struct<thestackv1_concat_by_repo-65536: struct<sequences: int64>, book-65536: struct<sequences: int6 (... 40 chars omitted)
child 0, thestackv1_concat_by_repo-65536: struct<sequences: int64>
child 0, sequences: int64
child 1, book-65536: struct<sequences: int64>
child 0, sequences: int64
child 2, textbooks: struct<sequences: int64>
child 0, sequences: int64
tokenizer: struct<name: string, revision: string, vocab_size: int64, eos_token_id: int64>
child 0, name: string
child 1, revision: string
child 2, vocab_size: int64
child 3, eos_token_id: int64
validation_source: struct<dataset: string, revision: string, source_tokenizer: struct<name: string, revision: string>, (... 1071 chars omitted)
child 0, dataset: strin
...
ars omitted)
child 0, thestackv1_concat_by_repo-65536: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 1, book-65536: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 2, textbooks: struct<source_samples: int64, source_documents: int64, target_roundtrip_mismatches: int64, target_to (... 59 chars omitted)
child 0, source_samples: int64
child 1, source_documents: int64
child 2, target_roundtrip_mismatches: int64
child 3, target_tokens_scanned: int64
child 4, discarded_document_tail_tokens: int64
child 10, converter_versions: struct<mosaicml-streaming: string, tokenizers: string, transformers: string>
child 0, mosaicml-streaming: string
child 1, tokenizers: string
child 2, transformers: string
child 11, script_sha256: string
total_validation_tokens: int64
to
{'format': Value('string'), 'tokenizer': {'name': Value('string'), 'revision': Value('string'), 'vocab_size': Value('int64'), 'eos_token_id': Value('int64')}, 'dtype': Value('string'), 'endianness': Value('string'), 'header_bytes': Value('int64'), 'magic': Value('int64'), 'version': Value('int64'), 'sequence_length': Value('int64'), 'validation': {'path': Value('string'), 'tokens': Value('int64'), 'sequences': Value('int64'), 'sha256': Value('string'), 'usable_tokens': Value('int64'), 'discarded_tail_tokens': Value('int64')}, 'validation_metadata_path': Value('string'), 'validation_metadata_sha256': Value('string'), 'total_validation_tokens': Value('int64'), 'total_validation_usable_tokens': Value('int64'), 'total_sequences': Value('int64'), 'domain_totals': {'thestackv1_concat_by_repo-65536': {'sequences': Value('int64')}, 'book-65536': {'sequences': Value('int64')}, 'textbooks': {'sequences': Value('int64')}}, 'validation_source': {'dataset': Value('string'), 'revision': Value('string'), 'source_tokenizer': {'name': Value('string'), 'revision': Value('string')}, 'target_tokenizer': {'name': Value('string'), 'revision': Value('string')}, 'split_special_tokens': Value('bool'), 'selection_seed': Value('int64'), 'selection': Value('string'), 'excluded_training': {'manifest_sha256': Value('string'), 'unique_source_samples': {'thestackv1_concat_by_repo-65536': Value('int64'), 'book-65536': Value('int64'), 'textbooks': Value('int64')}}, 'domain_sequence_counts': {'thestackv1_concat_by_repo-65536': Value('int64'), 'book-65536': Value('int64'), 'textbooks': Value('int64')}, 'stats': {'thestackv1_concat_by_repo-65536': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}, 'book-65536': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}, 'textbooks': {'source_samples': Value('int64'), 'source_documents': Value('int64'), 'target_roundtrip_mismatches': Value('int64'), 'target_tokens_scanned': Value('int64'), 'discarded_document_tail_tokens': Value('int64')}}, 'converter_versions': {'mosaicml-streaming': Value('string'), 'tokenizers': Value('string'), 'transformers': Value('string')}, 'script_sha256': Value('string')}}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
ProLong SmolLM2 validation
Two validation sets for evaluating DCLM-trained language models, retokenized from the code, books, and textbooks subsets of princeton-nlp/prolong-data-64K.
| Folder | Context length | Sequences | Usable tokens | Stored tokens | Trailing EOS filler (unused) |
|---|---|---|---|---|---|
4k/ |
4,096 | 24,414 | 99,999,744 | 100,000,000 | 256 |
32k/ |
32,768 | 3,051 | 99,975,168 | 100,000,000 | 24,832 |
Each folder contains prolong_val_100m.bin, per-sequence source metadata in prolong_val_100m.json, and manifest-prolong-val.json with checksums and conversion settings.
Conversion
- Target tokenizer: HuggingFaceTB/SmolLM2-135M, revision
93efa2f097d58c2a74874c7e644dbc9b0cee75a2; vocabulary 49,152, EOS ID 0. - Source tokenizer:
princeton-nlp/Llama-3-8B-ProLong-64k-Base, revisionee4b1380059c70aaeab703934ca669f368cc7905. - Decode each original document segment independently. Remove source BOS/end markers before decoding; append target EOS only when the source segment had an end marker. Literal special-token spellings are encoded as ordinary text (
split_special_tokens=True), matching DCLM preparation. - Keep complete context-length chunks within each document segment; discard its remainder. Do not join separate document segments into one sequence.
- Code/books/textbooks sequence quotas use weights
204884:196544:7084, allocated by largest remainder. Source rows follow NumPy seeded permutations with seeds 42/43/44 in that domain order. - No source rows are excluded. These are newly selected validation subsets of the source corpus, not an upstream official validation split. No deduplication against DCLM was performed.
- The two lengths use the same selection rule but are not identical token streams: changing chunk length changes discarded remainders and how many source rows fill the budget.
- Each manifest records one text round-trip mismatch, in the same code source row (
global_sample_index=23002): the pinned native SmolLM2 tokenizer drops oneU+001Dcontrol character. All saved chunks of that row match native tokenization exactly. Books and textbooks had no recorded round-trip mismatches.
Download and read
hf download sycmucmu/prolong-smollm2-validation --repo-type dataset --local-dir prolong-smollm2-validation
Each binary has one 1,024-byte header (256 little-endian int32 values) followed by little-endian uint16 token IDs. Header fields 0–2 are 20240520, 1, and the stored token count. Headers are not tokens. The trainer must ignore the trailing filler outside complete sequences.
import json
from pathlib import Path
import numpy as np
root = Path("prolong-smollm2-validation/32k")
manifest = json.loads((root / "manifest-prolong-val.json").read_text())
val = manifest["validation"]
tokens = np.memmap(root / val["path"], dtype="<u2", mode="r",
offset=manifest["header_bytes"], shape=(val["tokens"],))
sequences = tokens[:val["usable_tokens"]].reshape(val["sequences"], manifest["sequence_length"])
Future training exclusions
Preserve the per-sequence JSON. Each record contains domain, source (original MDS shard), shard_index, sample_index, global_sample_index, document_index, and document_chunk_index.
For later 32k ProLong training, exclude every (domain, global_sample_index) appearing in 32k/prolong_val_100m.json, so all chunks from a validation source row remain held out. This is 1,496 source rows; all 1,326 source rows used by the 4k set are included, so this exclusion protects both sets. The source field also permits excluding their 685 whole original MDS shards if desired. Source-row exclusion does not establish document-level deduplication across different source rows.
- Downloads last month
- 332