Commits

Breadcrumbs

History for

onmain

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 8, 2023

• Integrate Fabric.init_module(empty_weights=True) into scripts (Lightning-AI#360)

  awaelchli

  authoredJun 8, 2023
  4c1fcd3

Commits on Jun 2, 2023

• Adapter annotated (Lightning-AI#352)

  Andrei-Aksionov

  authoredJun 2, 2023
  0819c7f

Commits on Jun 1, 2023

• Fix quantization attribute assignment

  carmocca

  authoredJun 1, 2023
  cc6795d

• Fix the multigate adapter fix for old and new checkpoint files (Lightning-AI#348)

  

  rasbt

  and

  awaelchli

  authoredJun 1, 2023
  2bc146c

Commits on May 31, 2023

• incremental_save fixes (Lightning-AI#343)

  carmocca

  authoredMay 31, 2023
  7278fc6

Commits on May 30, 2023

• Add incremental_save to utils, use in convert_hf_checkpoint (Lightning-AI#326)

  t-vi

  authoredMay 30, 2023
  2f2305a

Commits on May 29, 2023

• Multihead gating for adapter (Lightning-AI#297)

  

  3 peopleauthoredMay 29, 2023
  67c1e95

• Fix typos (Lightning-AI#336)

  Andrei-Aksionov

  authoredMay 29, 2023
  1c9977f

Commits on May 25, 2023

• Fix attention comment in model.py (Lightning-AI#325)

  gregor-soniox

  authoredMay 25, 2023
  33185d4

Commits on May 23, 2023

• Reset all caches on xla devices (Lightning-AI#317)

  gkroiz

  authoredMay 23, 2023
  26f33a3

Commits on May 22, 2023

• Adding cached KVs (Lightning-AI#266)

  

  gkroiz

  and

  carmocca

  authoredMay 22, 2023
  a24fc5e

• Typo

  carmocca

  committedMay 22, 2023
  206f316

• Adapter v2 (Lightning-AI#303)

  

  rasbt

  and

  carmocca

  authoredMay 22, 2023
  028946f

Commits on May 21, 2023

• Fix typo

  carmocca

  authoredMay 21, 2023
  cdfbd91

• LoRA annotated (Lightning-AI#277)

  

  Andrei-Aksionov

  and

  awaelchli

  authoredMay 21, 2023
  0c6520f

Commits on May 16, 2023

• Fix reset mmaps and buffers (Lightning-AI#275)

  lantiga

  authoredMay 16, 2023
  98f29db

Commits on May 11, 2023

• refine llama model lookup for the benefit of quantized weights (Lightning-AI#259)

  t-vi

  authoredMay 11, 2023
  59a7516

Commits on May 3, 2023

• Automatic vocabulary size padding (Lightning-AI#223)

  carmocca

  authoredMay 3, 2023
  8875f4c

• Implement data-sharding in PackedDataset (Lightning-AI#218)

  

  awaelchli

  and

  lantiga

  authoredMay 3, 2023
  087de88

• Add option to wrap dataset around (Lightning-AI#219)

  lantiga

  authoredMay 3, 2023
  7e95724

Commits on May 2, 2023

• Pre-training script (Lightning-AI#209)

  

  lantiga

  and

  awaelchli

  authoredMay 2, 2023
  366bca6

Commits on May 1, 2023

• support more general tensors, change lazy_load into context manager to not leak open file (Lightning-AI#196)

  t-vi

  authoredMay 1, 2023
  01e7bde

Commits on Apr 27, 2023

• Make Red Pajama prepare scripts use PackedDataset (Lightning-AI#199)

  lantiga

  authoredApr 27, 2023
  58b6b36

Commits on Apr 26, 2023

• Add packed dataset (Lightning-AI#192)

  lantiga

  authoredApr 26, 2023
  d3c0075

• Update default values for max sequence length (Lightning-AI#194)

  awaelchli

  authoredApr 26, 2023
  5f484dd

Commits on Apr 24, 2023

• Automatically infer model from state_dict (Lightning-AI#160)

  

  3 peopleauthoredApr 24, 2023
  2d3f560

Commits on Apr 21, 2023

• Fix .eval() merging of LoRA weights (Lightning-AI#185)

  awaelchli

  authoredApr 21, 2023
  a1ff82d

Commits on Apr 20, 2023

• Real ROPE (Lightning-AI#174)

  t-vi

  authoredApr 20, 2023
  14a40ae

• Adding a quantized kernel through triton (Lightning-AI#173)

  t-vi

  authoredApr 20, 2023
  5ee9b87

Commits on Apr 19, 2023

• Make adapter.LLaMA.from_name available (Lightning-AI#170)

  awaelchli

  authoredApr 19, 2023
  863985c

• Add lazyload from torchhacks (Lightning-AI#164)

  t-vi

  authoredApr 19, 2023
  88ba008

• Add Red Pajama conversion script for sample and full dataset (Lightning-AI#157)

  lantiga

  authoredApr 19, 2023
  c232f7a

Commits on Apr 17, 2023

• Save and load only adapter weights (Lightning-AI#151)

  awaelchli

  authoredApr 17, 2023
  c8502d0

• Make generate_adapter.py work with --quantize argument (Lightning-AI#149)

  awaelchli

  authoredApr 17, 2023
  f8cf484

• Add generation script for LoRA (Lightning-AI#145)

  

  lantiga

  and

  awaelchli

  authoredApr 17, 2023
  89f285e

Pagination

PreviousNext