Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on May 23, 2024

• Add recipe test for llama3 (pytorch#929)

  SLR722

  authored
  0a7d488

  Copy full SHA for 0a7d488

Commits on May 22, 2024

• Fix TypeError: tuple indices must be integers or slices, not str issue with collate_fn (pytorch#1011)

  tambulkar

  authored
  174bb15

  Copy full SHA for 174bb15

Commits on May 21, 2024

• Datasets tutorial improvements (pytorch#994)

  RdoubleA

  authored
  dc74267

  Copy full SHA for dc74267

• Save adapter config and remapped adapter weights for loading into PEFT (pytorch#933)

  ebsmothers

  authored
  dc9c697

  Copy full SHA for dc9c697

• Support for unstructured text corpus datasets for CPT (pytorch#868)

  RdoubleA

  authored
  bbf0010

  Copy full SHA for bbf0010

Commits on May 20, 2024

• TorchTune --> torchtune (pytorch#1007)

  joecummings

  authored
  29ae975

  Copy full SHA for 29ae975

• Filter out special tokens and placeholder tokens for Phi-3 (pytorch#983)

  joecummings

  authored
  27f6872

  Copy full SHA for 27f6872

Commits on May 19, 2024

• Add support for free generation tasks in evals (pytorch#975)

  joecummings

  authored
  4203311

  Copy full SHA for 4203311

Commits on May 17, 2024

• fix: different rope base between phi3 and lora_phi3 (pytorch#997)

  Optimox

  authored
  eb66b62

  Copy full SHA for eb66b62

• fix: lora dropout applied to all models (pytorch#995)

  Optimox

  authored
  9f31a88

  Copy full SHA for 9f31a88

• Fix Gemma 2B model forward call (pytorch#998)

  joecummings

  authored
  926c6d8

  Copy full SHA for 926c6d8

• Llama3-8b memory efficient full finetune (pytorch#990)

  rohan-varma

  authored
  46d7c83

  Copy full SHA for 46d7c83

• Fix position embeddings for Phi3 when packing + nits (pytorch#992)

  RdoubleA

  authored
  f35e5d6

  Copy full SHA for f35e5d6

Commits on May 16, 2024

• updating help docs for hf-token arg in download.py (pytorch#991)

  SalmanMohammadi

  authored
  3883081

  Copy full SHA for 3883081

• Add better formatting for Eleuther eval results (pytorch#986)

  joecummings

  authored
  a6f4cc5

  Copy full SHA for a6f4cc5

Commits on May 15, 2024

• Utilize compile on model rather than via torch API (pytorch#953)

  joecummings

  authored
  618e80a

  Copy full SHA for 618e80a

• Sample packing for map datasets with correct RoPE encoding and no cross-contamination (pytorch#875)

  RdoubleA

  authored
  1e8d081

  Copy full SHA for 1e8d081

Commits on May 13, 2024

• Delete init_weights_with_constant test util (pytorch#974)

  ebsmothers

  authored
  f3611e5

  Copy full SHA for f3611e5

Commits on May 11, 2024

• Update tokens_per_sec to tokens_per_sec_per_gpu (pytorch#956)

  kartikayk

  authored
  dc2b991

  Copy full SHA for dc2b991

• Update documentation tab to point to main instead of stable (pytorch#960)

  kartikayk

  authored
  cb8e65a

  Copy full SHA for cb8e65a

Commits on May 9, 2024

• Add batched inference (pytorch#947)

  joecummings

  authored
  30c75d4

  Copy full SHA for 30c75d4

• Update readme table (pytorch#955)

  kartikayk

  authored
  0217d44

  Copy full SHA for 0217d44

Commits on May 8, 2024

• Add LoRA and QLoRA Components, Models, and Configs for Phi3 Mini (pytorch#932)

  pbontrager

  authored
  1e56b80

  Copy full SHA for 1e56b80

• Update eval wrapper to match lm_eval's API change (pytorch#950)

  water-vapor

  authored
  9f140c9

  Copy full SHA for 9f140c9

Commits on May 7, 2024

• FSDP Llama3 wrapping improvements for full finetune (pytorch#865)

  rohan-varma

  authored
  fa1392b

  Copy full SHA for fa1392b

Commits on May 6, 2024

• Remove comment about limited QLoRA + compile support (pytorch#941)

  rohan-varma

  authored
  ded5764

  Copy full SHA for ded5764

Commits on May 4, 2024

• Loading mistral reward model checkpoints (pytorch#911)

  SalmanMohammadi

  authored
  057709e

  Copy full SHA for 057709e

Commits on May 3, 2024

• Feat: Add support of multiple datasets in config (pytorch#889)

  Show description for d36e818
  

  3 peopleauthored
  d36e818

  Copy full SHA for d36e818

• Fix decode for Phi-3 Mini (pytorch#923)

  joecummings

  authored
  9274c89

  Copy full SHA for 9274c89

Commits on May 2, 2024

• Update Phi 3 Config in README.md (pytorch#927)

  kunal-mansukhani

  authored
  7d05579

  Copy full SHA for 7d05579

• Update README.md to include configs for Phi3 (pytorch#926)

  Show description for ca18fff
  

  kunal-mansukhani

  and

  kartikayk

  authored
  ca18fff

  Copy full SHA for ca18fff

• Fix import error from quantization before PyTorch 2.3 (pytorch#920)

  andrewor14

  authored
  3c1010c

  Copy full SHA for 3c1010c

• Finish additions needed for Phi-3 Mini 4K (pytorch#912)

  joecummings

  authored
  f5b8eaf

  Copy full SHA for f5b8eaf

Commits on May 1, 2024

• Mistral testing (pytorch#888)

  Show description for 06c5fcb
  

  SalmanMohammadi

  and

  ebsmothers

  authored
  06c5fcb

  Copy full SHA for 06c5fcb

Commits on Apr 30, 2024

• Add support for 8da4w quantization (pytorch#884)

  andrewor14

  authored
  fb59735

  Copy full SHA for fb59735

Pagination

Previous

Next