A fully sharded data parallel trainer for jax
Project description
jax-trainer
An efficient distributed FSDP in jax
Running Examples
Single host trainer
Run imdb movie classification from the long range arena dataset with batch parallelisation given by jit function:
pdm run bash ./tax/examples/bin/run_lra.sh $PWD imdb.yaml lra_imdb
For Library users see example usages in:
tax/tax/examples/lra/exp.py
FSDP Trainer
Train a gemma language model with the FSDP trainer
pdm run bash ./tax/examples/bin/run_lm.sh $PWD lm.yaml test_trainer
For Library users see example usages in:
tax/tax/examples/gemma/exp.py
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file tax_dp-0.1.1.tar.gz.
File metadata
- Download URL: tax_dp-0.1.1.tar.gz
- Upload date:
- Size: 40.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: pdm/2.19.1 CPython/3.10.6 Linux/6.8.0-45-generic
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
e19db2db107460d56c8b8c349a45a928ec30811e57c81b2020494b9cfa696623
|
|
| MD5 |
babb9aeed85b47e3bb041712931073f2
|
|
| BLAKE2b-256 |
0384f000f90e0951ed7e2a0d259b2542392c94e6b314b72719c5ad6b6613d3ad
|
File details
Details for the file tax_dp-0.1.1-py3-none-any.whl.
File metadata
- Download URL: tax_dp-0.1.1-py3-none-any.whl
- Upload date:
- Size: 51.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: pdm/2.19.1 CPython/3.10.6 Linux/6.8.0-45-generic
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
6b2cd0eeb198ab00ab3e53af03e2e0191868526b9b7c38c96960e082856f45a0
|
|
| MD5 |
fed388db2985412abfa722636dbd60bd
|
|
| BLAKE2b-256 |
f99edb6eb11e2afda415c7e6c809379e48c05df4636218095bbd415bf9476880
|