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
descriptor: struct<execution: struct<accumulator_dtype: string, attention_backend: string, batch_policy: struct< (... 1598 chars omitted)
child 0, execution: struct<accumulator_dtype: string, attention_backend: string, batch_policy: struct<algorithm: string, (... 759 chars omitted)
child 0, accumulator_dtype: string
child 1, attention_backend: string
child 2, batch_policy: struct<algorithm: string, max_sequences: int64, max_tokens: int64, window_size: int64>
child 0, algorithm: string
child 1, max_sequences: int64
child 2, max_tokens: int64
child 3, window_size: int64
child 3, compute_dtype: string
child 4, device_type: string
child 5, engine_sha256: string
child 6, kernel_sha256: null
child 7, numerical_policy: struct<autocast_dtype: string, autocast_enabled: bool, cuda_matmul_allow_tf32: bool, cudnn_allow_tf3 (... 339 chars omitted)
child 0, autocast_dtype: string
child 1, autocast_enabled: bool
child 2, cuda_matmul_allow_tf32: bool
child 3, cudnn_allow_tf32: bool
child 4, deterministic_algorithms: bool
child 5, esmc_fp8: bool
child 6, esmc_precision: struct<converted_projections: int64, device: string, enabled: bool, reason: string, transformer_engi (... 17 chars omitted)
child 0, converted_projections: int64
child 1, device: string
child 2, enabled: bool
child
...
4, singleton_variance: int64
child 5, variance_correction: int64
child 6, version: int64
child 3, version: string
child 12, sae: null
child 13, schema: string
child 14, special_tokens: string
child 15, structural: null
child 16, tokenizer: struct<content_sha256: string, repository: string, revision: string>
child 0, content_sha256: string
child 1, repository: string
child 2, revision: string
dtype: string
format: string
key: string
layout: string
positions: bool
width: int64
volume_directory: string
model_state_sha256: string
special_tokens: string
sae_layer: int64
normalization: string
streams: struct<final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, (... 35 chars omitted)
child 0, final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, sparse_count: null, wi (... 11 chars omitted)
child 0, descriptor_sha256: string
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, sparse_count: null
child 5, width: int64
schema: string
volume: string
max_residues: int64
label: string
store_kind: string
base: string
row_layout: string
random_init: struct<architecture: string, dtype: string, initializer: string, kind: string, seed: int64>
child 0, architecture: string
child 1, dtype: string
child 2, initializer: string
child 3, kind: string
child 4, seed: int64
hub_repo: string
to
{'base': Value('string'), 'hub_repo': Value('string'), 'label': Value('string'), 'max_residues': Value('int64'), 'model_state_sha256': Value('string'), 'normalization': Value('string'), 'random_init': {'architecture': Value('string'), 'dtype': Value('string'), 'initializer': Value('string'), 'kind': Value('string'), 'seed': Value('int64')}, 'row_layout': Value('string'), 'sae_layer': Value('int64'), 'schema': Value('string'), 'special_tokens': Value('string'), 'store_kind': Value('string'), 'streams': {'final_mean_var': {'descriptor_sha256': Value('string'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'sparse_count': Value('null'), 'width': Value('int64')}}, 'volume': Value('string'), 'volume_directory': 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
descriptor: struct<execution: struct<accumulator_dtype: string, attention_backend: string, batch_policy: struct< (... 1598 chars omitted)
child 0, execution: struct<accumulator_dtype: string, attention_backend: string, batch_policy: struct<algorithm: string, (... 759 chars omitted)
child 0, accumulator_dtype: string
child 1, attention_backend: string
child 2, batch_policy: struct<algorithm: string, max_sequences: int64, max_tokens: int64, window_size: int64>
child 0, algorithm: string
child 1, max_sequences: int64
child 2, max_tokens: int64
child 3, window_size: int64
child 3, compute_dtype: string
child 4, device_type: string
child 5, engine_sha256: string
child 6, kernel_sha256: null
child 7, numerical_policy: struct<autocast_dtype: string, autocast_enabled: bool, cuda_matmul_allow_tf32: bool, cudnn_allow_tf3 (... 339 chars omitted)
child 0, autocast_dtype: string
child 1, autocast_enabled: bool
child 2, cuda_matmul_allow_tf32: bool
child 3, cudnn_allow_tf32: bool
child 4, deterministic_algorithms: bool
child 5, esmc_fp8: bool
child 6, esmc_precision: struct<converted_projections: int64, device: string, enabled: bool, reason: string, transformer_engi (... 17 chars omitted)
child 0, converted_projections: int64
child 1, device: string
child 2, enabled: bool
child
...
4, singleton_variance: int64
child 5, variance_correction: int64
child 6, version: int64
child 3, version: string
child 12, sae: null
child 13, schema: string
child 14, special_tokens: string
child 15, structural: null
child 16, tokenizer: struct<content_sha256: string, repository: string, revision: string>
child 0, content_sha256: string
child 1, repository: string
child 2, revision: string
dtype: string
format: string
key: string
layout: string
positions: bool
width: int64
volume_directory: string
model_state_sha256: string
special_tokens: string
sae_layer: int64
normalization: string
streams: struct<final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, (... 35 chars omitted)
child 0, final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, sparse_count: null, wi (... 11 chars omitted)
child 0, descriptor_sha256: string
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, sparse_count: null
child 5, width: int64
schema: string
volume: string
max_residues: int64
label: string
store_kind: string
base: string
row_layout: string
random_init: struct<architecture: string, dtype: string, initializer: string, kind: string, seed: int64>
child 0, architecture: string
child 1, dtype: string
child 2, initializer: string
child 3, kind: string
child 4, seed: int64
hub_repo: string
to
{'base': Value('string'), 'hub_repo': Value('string'), 'label': Value('string'), 'max_residues': Value('int64'), 'model_state_sha256': Value('string'), 'normalization': Value('string'), 'random_init': {'architecture': Value('string'), 'dtype': Value('string'), 'initializer': Value('string'), 'kind': Value('string'), 'seed': Value('int64')}, 'row_layout': Value('string'), 'sae_layer': Value('int64'), 'schema': Value('string'), 'special_tokens': Value('string'), 'store_kind': Value('string'), 'streams': {'final_mean_var': {'descriptor_sha256': Value('string'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'sparse_count': Value('null'), 'width': Value('int64')}}, 'volume': Value('string'), 'volume_directory': 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.
ESMC-300 random init pooled embeddings
One vector per protein from ESMC-300 random init, in the FastPLMs feature-store format, for every protein of
Synthyra/canonical_embedding_sequences_v1.
The per-token streams of interaction proteins are in Synthyra/esmc_300_random_token_embeddings.
The weights are random: seed 20261001, the ESMC-300 architecture, no training. This is a control, not a protein model.
Each vector covers all l + 2 token rows of a protein (cropped from the N-terminus at 2046): row 0 is CLS, rows 1..l
are the residues, row l + 1 is EOS, and padding is the only masked position. final_mean_var is the mean, then the population
variance, of the final normalized hidden state. sae_max is the maximum of the SAE codes over those rows.
A row is keyed by the SHA-256 of the uppercased, whitespace-stripped sequence. The store only appends.
| stream | layout | width | dtype |
|---|---|---|---|
final_mean_var |
dense | 1920 | float32 |
Read it with foundry.embedding.open_canonical_view(root, inventory) (a FeatureView), or with
fastplms.features.FeatureStore directly. canonical.json lists the streams and pins each descriptor. Model state
SHA-256: 5cd00a95dc9126242a88c566007b82e10bb30449fe3f8615efc77c6b54d8a6e5.
- Downloads last month
- -