databricks/lilac

Issues

• UI progress bars lost count, ex/sec, description information.

  #1017 opened a year ago
  0

• Loading a dataset should not block the REST endpoint.

  #1012 opened a year ago
  0

• Allow datasets with spaces in the name.

  #990 opened a year ago
  0

• Markdown code block aware chunking.

  #989 opened a year ago
  0

• ShareGPT format: compare to menu should show the title instead of `conversation.0.value`

  #988 opened a year ago
  3

• Add ability to jump to a random integer offset in single-item view.

  #982 opened a year ago
  0

• Add keyboard shortcuts for fast labeling

  #979 opened a year ago
  1

• Allow user-provided embeddings when ingesting from a source.

  #978 opened a year ago
  0

• Default to combine_cols=True

  #973 opened a year ago
  0

• labels stay in ui after deletion.

  #972 opened a year ago
  0

• Rendering bug when we have `a[].b[]`

  #971 opened a year ago
  1

• Error Computing Embedding with Parquet File in Lilac v0.1.24 on Windows 11

  #969 opened a year ago
  1

• Add `dataset.map(embeddings=True)`

  #967 opened a year ago
  0

• Datasets w/ map bug: DataType has no attribute 'key_type'

  #963 opened a year ago
  0

• Confused by HDBscan in UI

  #960 opened a year ago
  2

• Add `dataset.count(filters=...)` public API

  #954 opened a year ago
  0

• Remove num_rows from manifest and add a count() endpoint

  #953 opened a year ago
  1

• duckdb syntax error

  #949 opened a year ago
  0

• Dockerfile to build application front and back from source

  #945 opened a year ago
  4

• website: cross-wired footer follow links

  #940 opened a year ago
  2

• Add documentation in the guide for limit & filter.

  #939 opened a year ago
  0

• I want to modify the Datasets directly

  #937 opened a year ago
  1

• Allow deletion of columns created from map from the UI.

  #936 opened a year ago
  1

• Multiprocess `map` got ~40% slower

  #928 opened a year ago
  1

• Multithreaded map only shows progress from shard 0

  #927 opened a year ago
  1

• `Map` spans the ipython output cell and freezes the notebook while running

  #926 opened a year ago
  1

• Break up the TaskManager.

  #922 opened a year ago
  0

• Use a single tqdm for task, instead of showing the sharded tqdms.

  #921 opened a year ago
  0

• Ergonomics and small bugs with `dataset.map`

  #915 opened a year ago
  2

• Write a getting started guide.

  #906 opened a year ago
  0

• Improve the CSS for markdown rendered content.

  #905 opened a year ago
  1

• Build versioning of source & map data.

  #903 opened a year ago
  0

• Add a simple signal python editor from the UI.

  #902 opened a year ago
  0

• Compute signal page is broken -- no fields available.

  #876 opened a year ago
  0

• Add defragmentation of signal output

  #872 opened a year ago
  0

• Allow subsampling datasets

  #871 opened a year ago
  1

• Add `input_cols` to `Dataset.map` to allow an input column to be processed.

  #868 opened a year ago
  2

• Add batch_size to Dataset.map()

  #867 opened a year ago
  1

• Improve bin legibility for floats.

  #857 opened a year ago
  0

• API error when attempting to group 'messages', '*', 'author_name' by anything other than the initial value it provides

  #851 opened a year ago
  3

• Data not showing up in Hugging Face UI after deploying local project using `ll.deploy_project()`

  #834 opened a year ago
  3

• stream of HTTP 500s after deleting a dataset

  #830 opened a year ago
  2

• Build error when deploying local project to Hugging Face Spaces

  #828 opened a year ago
  3

• Use presidio for PII detection.

  #825 opened a year ago
  0

• Add support for exporting directly to a HuggingFace dataset.

  #823 opened a year ago
  0

• Add support for reading TFRecords.

  #822 opened a year ago
  0

• Add column-level diffing.

  #821 opened a year ago
  0

• Add sampling to the jsonl reader.

  #820 opened a year ago
  1

• Add multi-processing support for Dataset.map()

  #819 opened a year ago
  1

• Allow customizing the OpenAI prompt for generating example positives when creating a concept.

  #818 opened a year ago
  0