Merge dev onto main for v1.3 release#1585

Merged

RyanMullins merged 88 commits intomainfrom dev

Oct 22, 2024

+2,896-9,866

Commits on Jun 28, 2024

Internal Change

cpka145
authored and
LIT team
committed

Commits on Jul 3, 2024

Internal Change

cpka145
authored and
LIT team
committed
Fix broken links to lit demos and documentation.

cpka145
authored and
LIT team
committed

Commits on Jul 9, 2024

Internal change.

bdu91
authored and
LIT team
committed

Commits on Jul 10, 2024

Internal Change

cpka145
authored and
LIT team
committed

Commits on Jul 25, 2024

No public description
Googler
committed

Commits on Aug 6, 2024

Make sequence length a flag when launching the prompt debugging demo.

bdu91
authored and
LIT team
committed

Commits on Aug 7, 2024

feat: Add initial version of LIT demos for generative models.

llcourage
authored and
LIT team
committed

Commits on Aug 8, 2024

Upgrade LIT to MobX v6.

RyanMullins
authored and
LIT team
committed
Fix Classification Results module coloring for correct class

RyanMullins
authored and
LIT team
committed
Add Gemma model to VertexModelGardenModel class.

llcourage
authored and
LIT team
committed
Fix range access issue in Classification Module

RyanMullins
authored and
LIT team
committed
Make aip_endpoint_name optional for SelfHostedGenerativeModel, so that user can input the endpoint_name in LIT UI.

llcourage
authored and
LIT team
committed

Commits on Aug 9, 2024

Set default temperature to 0.7 for GCP models.

llcourage
authored and
LIT team
committed
Enable GCP demo use Gemma model.

llcourage
authored and
LIT team
committed

Commits on Aug 13, 2024

Remove a TODO that is infeasible. The generate_content function converts a list of input to a single content instead of a batch of inputs.(https://source.corp.google.com/piper///depot/google3/third…

llcourage
authored and
LIT team
committed

Commits on Aug 16, 2024

Internal change.

bdu91
authored and
LIT team
committed

Commits on Aug 20, 2024

Documentation update.

bdu91
authored and
LIT team
committed

Commits on Sep 4, 2024

Factor LLM model name generation into a utility function.

RyanMullins
authored and
LIT team
committed

Commits on Sep 5, 2024

Fix indexing issue in Sequence Salience module.

RyanMullins
authored and
LIT team
committed

Commits on Sep 6, 2024

Refactor LIT Vertex AI demos to the lit.nlp.examples.vertexAI folder.

llcourage
authored and
LIT team
committed
Fix TS2869 compilation issues with TypeScript 5.6.
Googler
committed

Commits on Sep 10, 2024

LIT: Refactor model field names and specs into a constants module.

RyanMullins
authored and
LIT team
committed
Add Dockerfile for LIT Vertex AI demo

llcourage
authored and
LIT team
committed

Commits on Sep 12, 2024

LIT: Use ENV VAR=value form in Dockerfiles

RyanMullins
authored and
LIT team
committed
LIT: Change the demo_mode flag to be False in Vertex AI demo.

llcourage
authored and
LIT team
committed
LIT: Update numpy version to be smaller than 2.0.0.

llcourage
authored and
LIT team
committed

Commits on Sep 13, 2024

Internal Change

cpka145
authored and
LIT team
committed
Fix dependency issue at OSS.

bdu91
authored and
LIT team
committed

Commits on Sep 16, 2024

LIT: Load multiple model wrappers with shared model via the /create_model API

RyanMullins
authored and
LIT team
committed

Commits on Sep 18, 2024

Move the file caching logic to within the model __init__ function.

bdu91
authored and
LIT team
committed
LIT: Update layout for LIT vertexai demo.

llcourage
authored and
LIT team
committed
Make transformer lms HFGenerativeModel output generated texts.

bdu91
authored and
LIT team
committed
Add the custom model and dataset loaders to prompt debugging notebook.

bdu91
authored and
LIT team
committed

Commits on Sep 20, 2024

Add instructions on how to debug LIT UI in Colab.

bdu91
authored and
LIT team
committed

Commits on Oct 2, 2024

LIT: Fix assertion error when generation stops short of max_length

RyanMullins
authored and
LIT team
committed
LIT: Isolate DL runtime-specific generation logic for Transformers

RyanMullins
authored and
LIT team
committed

Commits on Oct 3, 2024

LIT:Fix type hints for model initializers.

llcourage
authored and
LIT team
committed
LIT:Update default DL runtime to torch.

llcourage
authored and
LIT team
committed

Commits on Oct 10, 2024

Internal change.

bdu91
authored and
LIT team
committed

Commits on Oct 11, 2024

Converting hosted demos images to multi-stage builds
RyanMullins
committed
Adding initial dockerfile for GCP use case
RyanMullins
committed
Update lit-on-gcp requirements.

llcourage
authored and
RyanMullins
committed
Updating date on license declaration
RyanMullins
committed
Updating dockerfile for LIT on GCP
RyanMullins
committed
Add model_server to lit-on-gcp,

llcourage
authored and
RyanMullins
committed
Updating copyright year in Dockerfile
RyanMullins
committed
Add unit test for model_server.

llcourage
authored and
RyanMullins
committed
Add webtest package to lit_on_gcp requirement file.

llcourage
authored and
RyanMullins
committed
Adding gunicorn config for model_server.py
RyanMullins
committed
Add unit test for model_server of lit on GCP.

llcourage
authored and
RyanMullins
committed
Update pyproject.toml and requirment_test files.

llcourage
authored and
RyanMullins
committed
Fetch model_name from model_config for lit-on-gcp model server.

llcourage
authored and
RyanMullins
committed
Fixing KeyError in tests related to default name change
RyanMullins
committed
Fixing requirements and udpating README
RyanMullins
committed
Adding testing instructions to README
RyanMullins
committed
Correct the model name of model_server.

llcourage
authored and
RyanMullins
committed
Fix wsgi handler name.

llcourage
authored and
RyanMullins
committed
Modify the output to be a list to ensure compatibility with JSON serialization.

llcourage
authored and
RyanMullins
committed
Adopting the generate_model_group_names() API from PD Utils
RyanMullins
committed
Using generate_model_group_names() API in model server test
RyanMullins
committed
Code health update on model server tests
RyanMullins
committed
Linting cleanup in the model server
RyanMullins
committed
Adds LIT app server code
RyanMullins
committed
Adding LIT App server for LIT LLMs on GCP
RyanMullins
committed
Fixing enum accessor in UTs
RyanMullins
committed
Migrate to updated pd_models.get_models kw args
RyanMullins
committed
Fix errors in gcp model_server and models.

llcourage
authored and
RyanMullins
committed
Combine vertexai docker image and gcp_server image.

llcourage
authored and
RyanMullins
committed
Add identity_token as LLM_ON_GCP init spec.

llcourage
authored and
RyanMullins
committed
Rename the model loader name.

llcourage
authored and
RyanMullins
committed
Updating requirements for TF 2.16 compatibility
RyanMullins
committed

Commits on Oct 15, 2024

Commits on Oct 16, 2024

Merge pull request #1546 from PAIR-code:lit-on-gcp
LIT team
committed

Commits on Oct 18, 2024

Adding pandas optional deps for TFDS dataset compatibility

RyanMullins
authored and
LIT team
committed

Commits on Oct 22, 2024

README updates to accommodate LIT on GCP changes.

RyanMullins
authored and
LIT team
committed
LIT 1.3 release notes.

RyanMullins
authored and
LIT team
committed
Updating website for v1.3 launch
RyanMullins
committed
Removing errant package-lock file
RyanMullins
committed
Merge pull request #1586 from PAIR-code:website-1_3_0
LIT team
committed
Fixing prod pip installs in Dockerfiles
RyanMullins
committed