Low-code framework for building custom LLMs, neural networks, and other AI models
APACHE-2.0 License
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.10.2...v0.10.3
Published by alexsherstinsky 7 months ago
true
in config (can be used in conjunction): adapter:
type: lora
use_rslora: false
use_dora: false
trainer:
eval_batch_size: auto
from_checkpoint=True
to LudwigModel.load()
:LudwigModel.load(model_dir, from_checkpoint=True)
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.10.1...v0.10.2
Published by alexsherstinsky 8 months ago
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.10.0...v0.10.1
Published by alexsherstinsky 8 months ago
get_quantization
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3926
use_reentrant
to True
to fix Mixtral-7b
bug by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3928
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.9.3...v0.10.0
microsoft/phi-2
by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3880
LLMEncoder
output to torch.float32
, freeze final layer at init. by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3900
LLMEncoder
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3902
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.9.2...v0.9.3
Published by alexsherstinsky 9 months ago
name
and description
classmethods to IA3Config
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3844
prepare_for_trianing
logic to ECD model for LLM encoder adapter initialization by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3874
libsox-dev
and update API by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3879
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.9.1...v0.9.2
Published by geoffreyangus 10 months ago
name
and description
classmethods to IA3Config
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3844
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.9...v0.9.1
Published by justinxzhao 10 months ago
combiner_registry
to combiner_config_registry
, update decorator name by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3516
response
column for text output features during postprocessing by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3521
datetime.date
date features by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3534
effective_batch_size
to auto-adjust gradient accumulation by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3533
LambdaLR
, and add cosine annealing LR scheduler as a decay method. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3555
prompt
validation check by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3564
check_module_parameters_updated
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3567
4.33.2
by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3637
model.predict
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3666
max_new_tokens
based on output feature length, GMSL and model window size by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3713
_initialize_llm
before moving to cuda. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3762
huggingface-hub>=0.19.0
by @brightsparc in https://github.com/ludwig-ai/ludwig/pull/3834
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8...v0.9
Published by justinxzhao about 1 year ago
max_new_tokens
based on output feature length, GMSL and model window size by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3713
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.5...v0.8.6
Published by justinxzhao about 1 year ago
model.predict
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3666
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.4...v0.8.5
Published by justinxzhao about 1 year ago
4.33.2
by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3637
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.3...v0.8.4
Published by justinxzhao about 1 year ago
prompt
validation check by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3564
check_module_parameters_updated
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3567
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.2...v0.8.3
Published by justinxzhao about 1 year ago
combiner_registry
to combiner_config_registry
, update decorator name by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3516
response
column for text output features during postprocessing by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3521
datetime.date
date features by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3534
effective_batch_size
to auto-adjust gradient accumulation by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3533
LambdaLR
, and add cosine annealing LR scheduler as a decay method. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3555
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.1...v0.8.2
Published by justinxzhao about 1 year ago
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8.1...v0.8.1.post1
Published by arnavgarg1 about 1 year ago
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.8...v0.8.1
Published by tgaddair about 1 year ago
Full Release Blog Post Here: https://predibase.com/blog/ludwig-v0-8-open-source-toolkit-to-build-and-fine-tune-custom-llms-on-your-data
Sarcos
dataset by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3169
model.type()
instead of instanceof(model)
. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3184
sample_ratio
, or while balancing data) by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3191
max_sequence_length
default to None
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3205
transformer
encoder and disable embed
encoder from SequenceCombiner
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3154
non_zero
to common_fields.NumFCLayersField
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3215
sequence_length
param by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3221
padding_value
by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3238
onehot
encoder outputs to float32 tensor by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3242
torchaudio
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3244
torchvision
and torchtext
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3248
pretrained_model_name_or_path
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3257
torchvision
, torchtext
, and torchaudio
. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3255
LUDWIG_SCHEMA_VALIDATION_POLICY
to change marshmallow validation strictness by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3226
model.compile
in PyTorch 2.0 by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3246
afn
parameter default to what's listed in HF docs. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3261
LudwigFeatureDict
to permit module keys that are rejected by torch ModuleDict. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3270
np.bool
to JSON NumpyEncoder
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3280
.get
to access LudwigFeatureDict
contents by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3279
kwargs
option to all file readers and feed nrows
where possible by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3266
fill_with_const
has fill_value
for binary features by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3278
type
param if it exists in the defaults config by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3296
kwargs
to read_parquet
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3293
transformers
by @geoffreyangus in https://github.com/ludwig-ai/ludwig/pull/3290
read_parquet
for remote filesystems by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3294
deberta
parameters by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3299
frac
arg in df.sample
instead of n
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3307
prompt_template
preprocessing param for text features by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3298
StringOptions
has duplicates and change validation errors to assertion errors for this field as well. by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3302
PROBABILITIES
and PREDICTIONS
to the prediction set for the sequence tagger decoder. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3300
trainer.gradient_accumulation_steps
for increasing effective batch size by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3305
iq
and all number feature preprocessing normalizations work on dask backends. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3297
ludwig.schema.hyperopt
by @jeffkinnison in https://github.com/ludwig-ai/ludwig/pull/3190
dist.barrier
instead of horovod legacy code by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3358
pip install expecttest
for torch nightly tests by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3366
expecttest
temporary install by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3371
eval
mode when using batch_size: 1
by @abidwael in https://github.com/ludwig-ai/ludwig/pull/3378
zscore
metadata by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3388
required
for input and output features to the Ludwig JSON schema by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3442
getdaft
to 0.1.6 by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3443
model_name
with required base_model
, add preset LLM registry, update internal adapter modules by @tgaddair in https://github.com/ludwig-ai/ludwig/pull/3423
tables
requirement as it causes issues installing ludwig in linux env. by @justinxzhao in https://github.com/ludwig-ai/ludwig/pull/3473
ludwig upload
to push artifacts to HuggingFace Hub by @arnavgarg1 in https://github.com/ludwig-ai/ludwig/pull/3480
sqlalchemy
to 1.x.x
versions by @ksbrar in https://github.com/ludwig-ai/ludwig/pull/3496
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.7...v0.8
Published by justinxzhao about 1 year ago
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.7.4...v0.7.5
Published by justinxzhao over 1 year ago
Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.7.3...v0.7.4
Published by tgaddair over 1 year ago
trainer.compile: true
(#3246)LUDWIG_SCHEMA_VALIDATION_POLICY
to change marshmallow validation strictness (#3226)sequence_length
capability (#3259)padding_value
(#3238)Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.7.2...v0.7.3
Published by tgaddair over 1 year ago
sample_ratio
, or while balancing data) (#3191)Full Changelog: https://github.com/ludwig-ai/ludwig/compare/v0.7.1...v0.7.2