Instructions to use Synthyra/ESMFold2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Synthyra/ESMFold2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="Synthyra/ESMFold2", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("Synthyra/ESMFold2", trust_remote_code=True) model = AutoModel.from_pretrained("Synthyra/ESMFold2", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 1,917 Bytes
80c2fed | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | """The FastPLMs feature store: one storage format for every embedding this workspace keeps.
A feature is one value per sequence, addressed by the SHA-256 of the sequence and by a key that
names the model, its revision, the sparse autoencoder, the layer, the pooling, the dtype, and the
residue limit. `store` holds the directory format and `layouts` the row layouts: dense vectors,
compressed-sparse pooled rows, ragged hidden states, and ragged top-k residue codes. `reader`
serves rows by random access to loops that read a batch at a time, `writing` commits a window of
rows as one segment, and `conversion` moves a cache another format holds into a store and proves the
rows survived.
"""
from .async_writer import AsyncFeatureWriter, PackedBatch
from .conversion import (
ConversionMismatch,
ConversionReceipt,
conversion_fingerprint,
convert_rows,
describe_file,
)
from .layouts import (
CSR,
DENSE,
LAYOUT_NAMES,
RAGGED,
RAGGED_TOPK,
SparseRow,
TopKRow,
dtype_name,
value_dtype,
)
from .reader import CsrRows, FeatureReader
from .store import (
FORMAT,
FeatureStore,
RowAddress,
SegmentReceipt,
SegmentWriter,
StoredFeature,
features_in,
open_feature,
partition_sequences,
sequence_digest,
)
from .writing import write_rows
__all__ = [
"CSR",
"DENSE",
"FORMAT",
"LAYOUT_NAMES",
"RAGGED",
"RAGGED_TOPK",
"AsyncFeatureWriter",
"ConversionMismatch",
"ConversionReceipt",
"CsrRows",
"FeatureReader",
"FeatureStore",
"PackedBatch",
"RowAddress",
"SegmentReceipt",
"SegmentWriter",
"SparseRow",
"StoredFeature",
"TopKRow",
"conversion_fingerprint",
"convert_rows",
"describe_file",
"dtype_name",
"features_in",
"open_feature",
"partition_sequences",
"sequence_digest",
"value_dtype",
"write_rows",
]
|