Skip to main content

Simple library for loading checkpoints of language models.

Project description

lm-checkpoints

Simple library for dealing with language model checkpoints.

Install using pip install lm-checkpoints.

Example

Say you want to compute some metrics for all model checkpoints of Pythia 160m, but only seed 0.

from lm_checkpoints import PythiaCheckpoints

for ckpt in PythiaCheckpoints(size=160,seed=[0]):
    # Do something with ckpt.model or ckpt.tokenizer

Or if you only want to load steps 0, 1, 2, 4, 8, 16 for all available seeds:

from lm_checkpoints import PythiaCheckpoints

for ckpt in PythiaCheckpoints(size=160,step=[0, 1, 2, 4, 8, 16]):
    # Do something with ckpt.model or ckpt.tokenizer

Alternatively, you may want to load all final checkpoints of MultiBERTs:

from lm_checkpoints import MultiBERTCheckpoints

for ckpt in MultiBERTCheckpoints.final_checkpoints():
    # Do something with ckpt.model or ckpt.tokenizer

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

lm_checkpoints-0.1.8.tar.gz (5.6 kB view details)

Uploaded Source

Built Distribution

lm_checkpoints-0.1.8-py3-none-any.whl (7.5 kB view details)

Uploaded Python 3

File details

Details for the file lm_checkpoints-0.1.8.tar.gz.

File metadata

  • Download URL: lm_checkpoints-0.1.8.tar.gz
  • Upload date:
  • Size: 5.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.2.2 CPython/3.10.13 Darwin/23.1.0

File hashes

Hashes for lm_checkpoints-0.1.8.tar.gz
Algorithm Hash digest
SHA256 08d55bed6446f4f3529dae00372ab687157faab7334af102aed10b30ec44f3da
MD5 366d9b84097d3e76f483624032dc4bf8
BLAKE2b-256 a2d0e5ac6d31d17da9d9b2652133d2afef3448d9cc9cabcbfdae813f76f19094

See more details on using hashes here.

File details

Details for the file lm_checkpoints-0.1.8-py3-none-any.whl.

File metadata

  • Download URL: lm_checkpoints-0.1.8-py3-none-any.whl
  • Upload date:
  • Size: 7.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.2.2 CPython/3.10.13 Darwin/23.1.0

File hashes

Hashes for lm_checkpoints-0.1.8-py3-none-any.whl
Algorithm Hash digest
SHA256 e3d3e722c010a263f0f159aa7438646c618ab765a67a98db13bf4dba9f77fff7
MD5 3b7ed533158a54122692ae15ac5df010
BLAKE2b-256 9e65ecb2b58d2cb07936f0435a386b1ddeb5f0ec4156a8f475f42a0c28a25a89

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page