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
base: string
hardware_class: struct<compute_capability: list<item: int64>, device_type: string, id: string, multiprocessors: int6 (... 2 chars omitted)
child 0, compute_capability: list<item: int64>
child 0, item: int64
child 1, device_type: string
child 2, id: string
child 3, multiprocessors: int64
hub_repo: string
label: string
max_residues: int64
model_state_sha256: string
normalization: string
path_in_repo: string
profile: struct<id: string, sha256: string>
child 0, id: string
child 1, sha256: string
random_init: null
row_layout: string
sae_layer: int64
schema: string
special_tokens: string
store_kind: string
store_model: string
streams: struct<final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, (... 157 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
child 1, sae_max: 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
visibility: string
volume: string
...
string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 3, sae_codes: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 4, sae_max: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 5, structural: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 13, visibility: string
updated_utc: timestamp[s]
repository: string
to
{'repository': Value('string'), 'schema': Value('string'), 'stores': {'profile_v1/esmc_300/pooled': {'base': Value('string'), 'bytes': Value('int64'), 'files': Value('int64'), 'hardware_class': {'compute_capability': List(Value('int64')), 'device_type': Value('string'), 'id': Value('string'), 'multiprocessors': Value('int64')}, 'index': {'parts': Value('string'), 'rows': Value('string')}, 'kind': Value('string'), 'label': Value('string'), 'model': Value('string'), 'model_state_sha256': Value('string'), 'path': Value('string'), 'profile': {'id': Value('string'), 'sha256': Value('string')}, 'proteins': Value('int64'), 'streams': {'final_mean_var': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_max': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}}, 'visibility': Value('string')}, 'profile_v1/esmc_300/token': {'base': Value('string'), 'bytes': Value('int64'), 'files': Value('int64'), 'hardware_class': {'compute_capability': List(Value('int64')), 'device_type': Value('string'), 'id': Value('string'), 'multiprocessors': Value('int64')}, 'index': {'parts': Value('string'), 'rows': Value('string')}, 'kind': Value('string'), 'label': Value('string'), 'model': Value('string'), 'model_state_sha256': Value('string'), 'path': Value('string'), 'profile': {'id': Value('string'), 'sha256': Value('string')}, 'proteins': Value('int64'), 'streams': {'final_hidden': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'final_mean_var': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'layer_hidden': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_codes': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_max': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'structural': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}}, 'visibility': Value('string')}}, 'updated_utc': Value('timestamp[s]')}
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 2951, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
base: string
hardware_class: struct<compute_capability: list<item: int64>, device_type: string, id: string, multiprocessors: int6 (... 2 chars omitted)
child 0, compute_capability: list<item: int64>
child 0, item: int64
child 1, device_type: string
child 2, id: string
child 3, multiprocessors: int64
hub_repo: string
label: string
max_residues: int64
model_state_sha256: string
normalization: string
path_in_repo: string
profile: struct<id: string, sha256: string>
child 0, id: string
child 1, sha256: string
random_init: null
row_layout: string
sae_layer: int64
schema: string
special_tokens: string
store_kind: string
store_model: string
streams: struct<final_mean_var: struct<descriptor_sha256: string, dtype: string, key: string, layout: string, (... 157 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
child 1, sae_max: 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
visibility: string
volume: string
...
string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 3, sae_codes: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 4, sae_max: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 5, structural: struct<bytes: int64, dtype: string, key: string, layout: string, parts: int64, proteins: int64, widt (... 9 chars omitted)
child 0, bytes: int64
child 1, dtype: string
child 2, key: string
child 3, layout: string
child 4, parts: int64
child 5, proteins: int64
child 6, width: int64
child 13, visibility: string
updated_utc: timestamp[s]
repository: string
to
{'repository': Value('string'), 'schema': Value('string'), 'stores': {'profile_v1/esmc_300/pooled': {'base': Value('string'), 'bytes': Value('int64'), 'files': Value('int64'), 'hardware_class': {'compute_capability': List(Value('int64')), 'device_type': Value('string'), 'id': Value('string'), 'multiprocessors': Value('int64')}, 'index': {'parts': Value('string'), 'rows': Value('string')}, 'kind': Value('string'), 'label': Value('string'), 'model': Value('string'), 'model_state_sha256': Value('string'), 'path': Value('string'), 'profile': {'id': Value('string'), 'sha256': Value('string')}, 'proteins': Value('int64'), 'streams': {'final_mean_var': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_max': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}}, 'visibility': Value('string')}, 'profile_v1/esmc_300/token': {'base': Value('string'), 'bytes': Value('int64'), 'files': Value('int64'), 'hardware_class': {'compute_capability': List(Value('int64')), 'device_type': Value('string'), 'id': Value('string'), 'multiprocessors': Value('int64')}, 'index': {'parts': Value('string'), 'rows': Value('string')}, 'kind': Value('string'), 'label': Value('string'), 'model': Value('string'), 'model_state_sha256': Value('string'), 'path': Value('string'), 'profile': {'id': Value('string'), 'sha256': Value('string')}, 'proteins': Value('int64'), 'streams': {'final_hidden': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'final_mean_var': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'layer_hidden': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_codes': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'sae_max': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}, 'structural': {'bytes': Value('int64'), 'dtype': Value('string'), 'key': Value('string'), 'layout': Value('string'), 'parts': Value('int64'), 'proteins': Value('int64'), 'width': Value('int64')}}, 'visibility': Value('string')}}, 'updated_utc': Value('timestamp[s]')}
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.
Synthyra/esmc_embeddings
Embeddings of the public ESMC models,
computed under an embedding profile: a pinned container, pinned numerics, one fixed batch shape per protein and a continuous
SAE readout, so every row is a function of its protein alone and bitwise reproducible on the hardware class that made it.
Each store is a FastPLMs feature store at <profile>/<model>/<kind>/; catalog.json lists them all.
| store | model | kind | proteins | GiB | streams | profile |
|---|---|---|---|---|---|---|
profile_v1/esmc_300/pooled |
ESMC-300 | pooled | 2,392,727 | 55.6 | final_mean_var, sae_max |
embedding_profile_v1 |
profile_v1/esmc_300/token |
ESMC-300 | token | 173,754 | 301.3 | final_hidden, final_mean_var, layer_hidden, sae_codes, sae_max, structural |
embedding_profile_v1 |
A row is keyed by the SHA-256 of the uppercased, whitespace-stripped sequence. Per-token streams hold l + 2 rows per protein (row 0 CLS, rows 1..l residues, row l + 1 EOS, cropped at 2046 residues); pooled streams hold one vector.
Each store carries index/rows.parquet (per protein: the segment, part and row of each stream, sorted by key) and
index/parts.parquet (per file: path, rows, bytes, SHA-256), so one protein or one shard is fetched without the rest:
from foundry.embedding.hub import fetch_rows
from foundry.embedding import open_canonical_view, sequence_inventory
root = fetch_rows("Synthyra/esmc_embeddings", "profile_v1/esmc_300/pooled", ["MKTAYIAKQR"], revision="<commit>", local_dir="embeddings")
with open_canonical_view(root, sequence_inventory(["MKTAYIAKQR"])) as view:
vector = view.pooled(["MKTAYIAKQR"], ["final_mean_var"])
Without the workspace, read index/rows.parquet with any parquet reader filtered on sha256, then download the listed part
with huggingface_hub.hf_hub_download; parts are safetensors files.
- Downloads last month
- -