Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    TypeError
Message:      int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1539, in _prepare_split_single
                  for key, record in generator:
                                     ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 131, in _generate_examples
                  for example_idx, example in enumerate(self._get_pipeline_from_tar(tar_path, tar_iterator)):
                                              ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 32, in _get_pipeline_from_tar
                  for filename, f in tar_iterator:
                                     ^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/track.py", line 49, in __iter__
                  for x in self.generator(*self.args):
                           ~~~~~~~~~~~~~~^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 1419, in _iter_from_urlpath
                  with xopen(urlpath, "rb", download_config=download_config, block_size=0) as f:
                       ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 976, in xopen
                  file_obj = fs.open(paths[0], mode)
                File "<string>", line 3, in open
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1176, in __call__
                  return self._mock_call(*args, **kwargs)
                         ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1180, in _mock_call
                  return self._execute_mock_call(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/unittest/mock.py", line 1247, in _execute_mock_call
                  result = effect(*args, **kwargs)
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 794, in wrapped
                  tracker.files[urlpath] = {"read": 0, "size": int(f.size)}
                                                               ~~~^^^^^^^^
              TypeError: int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1401, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1579, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

py
unknown
__key__
string
__url__
string
[ 102, 114, 111, 109, 32, 116, 114, 97, 110, 115, 102, 111, 114, 109, 101, 114, 115, 32, 105, 109, 112, 111, 114, 116, 32, 80, 114, 101, 116, 114, 97, 105, 110, 101, 100, 67, 111, 110, 102, 105, 103, 10, 10, 99, 108, 97, 115, ...
runtime/anytok_pool_20k/hf_hypernet/configuration_hypernet
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 80, 121, 84, 111, 114, 99, 104, 32, 105, 109, 112, 108, 101, 109, 101, 110, 116, 97, 116, 105, 111, 110, 32, 111, 102, 32, 90, 101, 84, 84, 39, 115, 32, 70, 108, 97, 120, 32, 72, 121, 112, 101, 114, 82, 111, ...
runtime/anytok_pool_20k/hf_hypernet/hyper_roberta
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 80, 121, 84, 111, 114, 99, 104, 32, 112, 111, 114, 116, 32, 111, 102, 32, 90, 101, 84, 84, 39, 115, 32, 70, 108, 97, 120, 72, 121, 112, 101, 114, 84, 53, 69, 110, 99, 111, 100, 101, 114, 77, 111, 100, 117, 10...
runtime/anytok_pool_20k/hf_hypernet/hyper_t5
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 90, 101, 84, 84, 39, 115, 32, 111, 114, 105, 103, 105, 110, 97, 108, 45, 116, 111, 107, 101, 110, 32, 101, 109, 98, 101, 100, 100, 105, 110, 103, 32, 119, 97, 114, 109, 117, 112, 44, 32, 114, 101, 115, 116, 114...
runtime/anytok_pool_20k/hf_hypernet/identity_warmup
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 68, 97, 116, 97, 32, 97, 110, 100, 32, 108, 111, 115, 115, 32, 104, 101, 108, 112, 101, 114, 115, 32, 102, 111, 114, 32, 105, 110, 112, 117, 116, 45, 111, 110, 108, 121, 32, 106, 111, 105, 110, 116, 32, 116, 11...
runtime/anytok_pool_20k/hf_hypernet/joint_training
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 102, 114, 111, 109, 32, 46, 99, 111, 110, 102, 105, 103, 117, 114, 97, 116, 105, 111, 110, 95, 104, 121, 112, 101, 114, 110, 101, 116, 32, 105, 109, 112, 111, 114, 116, 32, 90, 101, 116, 116, 72, 121, 112, 101, 114, 110, 101...
runtime/anytok_pool_20k/hf_hypernet/modeling_hypernet
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 85, 110, 105, 102, 111, 114, 109, 32, 115, 97, 109, 112, 108, 105, 110, 103, 32, 111, 102, 32, 105, 109, 109, 117, 116, 97, 98, 108, 101, 32, 115, 97, 118, 101, 100, 32, 116, 111, 107, 101, 110, 105, 122, 101, ...
runtime/anytok_pool_20k/hf_hypernet/pool_sampling
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 80, 114, 101, 112, 97, 114, 101, 32, 97, 32, 116, 105, 101, 100, 45, 119, 101, 105, 103, 104, 116, 32, 99, 97, 117, 115, 97, 108, 32, 76, 77, 32, 102, 111, 114, 32, 105, 110, 112, 117, 116, 45, 111, 110, 108, ...
runtime/anytok_pool_20k/hf_hypernet/smollm_setup
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 82, 101, 112, 114, 101, 115, 101, 110, 116, 32, 98, 121, 116, 101, 45, 108, 101, 118, 101, 108, 32, 116, 97, 114, 103, 101, 116, 32, 116, 111, 107, 101, 110, 115, 32, 102, 111, 114, 32, 97, 32, 115, 111, 117, 1...
runtime/anytok_pool_20k/hf_hypernet/tokenizer_inputs
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
null
runtime/anytok_pool_20k/rust_utils/target/release/librust_utils
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 74, 111, 105, 110, 116, 108, 121, 32, 116, 114, 97, 105, 110, 32, 83, 109, 111, 108, 76, 77, 50, 32, 97, 110, 100, 32, 97, 110, 32, 105, 110, 112, 117, 116, 45, 111, 110, 108, 121, 32, 90, 101, 84, 84, 32, 10...
runtime/anytok_pool_raw_bpb_rl_20k/scripts/train_joint_smollm_input
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 82, 101, 118, 101, 114, 115, 105, 98, 108, 101, 44, 32, 115, 104, 97, 114, 101, 100, 32, 116, 114, 97, 105, 110, 105, 110, 103, 32, 99, 111, 110, 116, 101, 120, 116, 115, 32, 97, 110, 100, 32, 97, 32, 112, 114,...
runtime/anytok_pool_regret_rl_20k/hf_hypernet/paired_reward
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 65, 32, 115, 104, 97, 114, 101, 100, 32, 115, 101, 116, 116, 105, 110, 103, 115, 45, 116, 111, 45, 108, 111, 103, 105, 116, 32, 110, 101, 116, 119, 111, 114, 107, 59, 32, 110, 111, 32, 112, 101, 114, 45, 116, 1...
runtime/anytok_pool_settings_rl_20k/hf_hypernet/settings_policy
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 73, 115, 111, 108, 97, 116, 101, 100, 32, 102, 117, 108, 108, 45, 118, 111, 99, 97, 98, 117, 108, 97, 114, 121, 32, 99, 104, 117, 110, 107, 105, 110, 103, 32, 101, 120, 112, 101, 114, 105, 109, 101, 110, 116, 4...
scripts/chunked_hypernetwork
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 79, 112, 116, 45, 105, 110, 32, 100, 105, 97, 103, 110, 111, 115, 116, 105, 99, 32, 112, 97, 116, 99, 104, 101, 115, 59, 32, 110, 101, 118, 101, 114, 32, 105, 109, 112, 111, 114, 116, 101, 100, 32, 98, 121, 32,...
scripts/local_memory_variants
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 67, 104, 117, 110, 107, 32, 76, 97, 121, 101, 114, 78, 111, 114, 109, 32, 102, 111, 114, 119, 97, 114, 100, 59, 32, 114, 101, 116, 97, 105, 110, 32, 116, 104, 101, 32, 111, 114, 105, 103, 105, 110, 97, 108, 32,...
scripts/lossless_layer_norm
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 78, 97, 116, 105, 118, 101, 32, 102, 117, 108, 108, 45, 115, 104, 97, 112, 101, 32, 77, 76, 80, 32, 109, 97, 116, 104, 32, 119, 105, 116, 104, 32, 101, 120, 112, 108, 105, 99, 105, 116, 32, 97, 99, 116, 105, ...
scripts/lossless_mlp
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
[ 34, 34, 34, 84, 104, 101, 32, 71, 80, 85, 45, 111, 110, 108, 121, 32, 101, 120, 101, 99, 117, 116, 105, 111, 110, 32, 118, 101, 114, 105, 102, 105, 101, 100, 32, 105, 110, 32, 116, 104, 101, 32, 99, 111, 109, 112, 108, 101...
scripts/pool_training_memory
hf://datasets/cminst/anytok-training-inputs@c2421b0c046967a7eb05fa628546c5058f69d035/training-code.tar.gz
End of preview.

AnyTok training inputs

Exact saved inputs for the AnyTok uniform-pool and adversarial tokenizer-selection experiments. This bundle supports the corrected raw-BPB versus regret comparison; the training code is in qingy1337/anytok, branch zett-modal-20k.

The archive contains all 1,000 native tokenizer JSON files and their generator settings, the exact 23-bucket training corpus (10 billion raw UTF-8 bytes), the corpus mixture recipe, original SmolLM2-1.7B model/canonical tokenizer, identity-warmed input hypernetwork, fixed sanity-evaluation data and complete checksums. These are initialization inputs; no partially trained RL checkpoint or access credentials are included.

The tokenizer pool contains 332 BPE, 328 WordPiece and 340 Unigram models. Vocabulary targets span 32k–128k, with regex, line-chunk and no-regex pretokenization. Each tokenizer was already trained on its saved 200 MB mixture. Do not regenerate them for a reproduction.

training-config.json records the intended fresh 20k-step raw-BPB experiment: entropy weight 0, corrected paired boundaries, batch 128, microbatch 32, maximum 128 total LM positions. The architecture and scoring protocol are unchanged. This file is documentation; use the repository launcher to train.

Download and verify

Download anytok-training-inputs-v1.tar.gz, its .sha256 sidecar and bundle-receipt.json from this repository. Verify the archive checksum, extract it into a cloud filesystem, then run the repository's scripts/verify_training_inputs.py --root /path/to/extracted to verify every file. The extracted layout is input/ plus BUNDLE_MANIFEST.json.

The agent-facing handoff README has installation, launch, resume and memory-fix details. Use the complete corpus and pool. All large files should move directly between cloud services.

Provenance

Original source: Modal profile queues, volume anytok-pool-rl-20k, prefix input/. input/recipe.json records upstream dataset names and pinned revisions; input/corpus/manifest.json records exact corpus hashes. BUNDLE_MANIFEST.json and input/rl-input-audit.json record all file hashes. The archive is reread and every member checksum verified before publication.

The corpus includes FineWeb/FineWeb-2, Wikipedia, GitHub code, peS2o, FineMath, Stack Exchange, Gutenberg, PubMed and legal text. Original dataset/model terms continue to apply; provenance is preserved in the bundle.

Downloads last month
53