ludwig

ludwig copied to clipboard

Add Ludwig schema definition to JSON Schema Store

2

We can enable auto-complete for Ludwig configs by uploading the definition to a common repo like https://www.schemastore.org/json/, which will make it available for folks using IntelliJ, VSCode, etc. This should...

tgaddair

Update text_encoders.py

1

This replaces https://huggingface.co/sileod/deberta-v3-base-tasksource-nli with https://huggingface.co/tasksource/deberta-base-long-nli The previous model is deprecated, and the newer one has doubled context window and more training data.

sileod

# Code Pull Requests Please provide the following: - a clear explanation of what your code does - if applicable, a reference to an issue - a reproducible test for...

arnavgarg1

Added LLaMa text encoder

1

Default is trained on alpaca-7b. Blocked by #3290.

tgaddair

Add Buildkite analytics integration

1

TODO: Condition to only run on merge to master, otherwise it'll get too noisy.

arnavgarg1

Fix NaN and INF in json

1

eval stats can be invalid json for np.nan and np.inf, so `test_statistics.json` is invalid

hungcs

Test Pandas 2.0 with Arrow backend

1

tgaddair

Re-enable GPU tests

1

geoffreyangus

Raise `ValueError` if no paths are successfully read for a given image feature column

2

geoffreyangus

test: tv encoders with large memory footprint

2

# Code Pull Requests Add unit test for TorchVision pretrained models that have a large memory footprint. These large memory models are excluded from the existing unit tests. This new...

jimthompson5802