Repository navigation

Rebase hf - #7

Merged

zachares merged 936 commits into

mainfrom

rebase-hf

Aug 11, 2023

Merged

Rebase hf#7
zachares merged 936 commits into
mainfrom
rebase-hf

This pull request is big! We're only showing the most recent 250 commits

Commits on Jul 18, 2023
- Enable `ZeroShotAudioClassificationPipelineTests::test_small_model_pt` (#24882)
  ydshieh
  authored
  57da42a
- Add DINOv2 (#24016)
  NielsRogge
  authored
  3ec10e6
- [`InstructBlip`] Fix int8/fp4 issues (#24888)
  younesbelkada
  authored
  a9e067a
- [`Blip`] Fix blip output name (#24889)
  younesbelkada
  authored
  5c5cb4e
- check if eval dataset is dict (#24877)
  hwijeen
  authored
  dd49404
- Separate CircleCI cache between `main` and `pull` (or other branches) (#24886)
  ydshieh
  authored
  30c172f
- [`Llama2`] Add support for Llama 2 (#24891)
  
  ArthurZucker
  and
  sgugger
  authored
  07360b6
- Disable ipex env var if false (#24885)
  muellerzr
  authored
  a982c02
- Check for accelerate env var when doing CPU only (#24890)
  muellerzr
  authored
  476be08
Commits on Jul 19, 2023
Commits on Jul 20, 2023
Commits on Jul 21, 2023
- fix fsdp checkpointing issues (#24926)
  pacman100
  authored
  1c7e5e2
- fix: cast input pixels to appropriate dtype for image_to_text pipelines (#24947)
  JimAllanson
  authored
  83f9314
- 🌐 [i18n-KO] Fixed Korean and English `quicktour.md` (#24664)
  
  authored
  ec3dfe5
- fsdp fixes and enhancements (#24980)
  pacman100
  authored
  f4eb459
- Fix missing spaces in system prompt of Llama2 tokenizer (#24930)
  
  chenjoya
  and
  ArthurZucker
  authored
  f74560d
- [`LlamaConfig`] Nit: pad token should be None by default (#24958)
  ArthurZucker
  authored
  0511369
- Remove tokenizers from the doc table (#24963)
  sgugger
  authored
  640e1b6
- Avoid importing all models when instantiating a pipeline (#24960)
  sgugger
  authored
  5b7ffd5
- Fix type annotation for deepspeed training arg (#24988)
  sgugger
  authored
  a6484c8
- Use main_input_name for include_inputs_for_metrics (#24993)
  sgugger
  authored
  a7d2131
- Fix `llama` tokenization doctest (#24990)
  ydshieh
  authored
  f1a1eb4
- [`bnb`] Add simple check for bnb import (#24995)
  younesbelkada
  authored
  d3ce048
- [`Llama`] remove persistent `inv_freq` tensor (#24998)
  ArthurZucker
  authored
  95f96b4
- improve from_pretrained for zero3 multi gpus mode (#24964)
  
  1ytic
  and
  stas00
  authored
  ea41e18
- Move template doc file to md (#25004)
  sgugger
  authored
  87fba94
- 🌐 [i18n-KO] Updated Korean `serialization.md` (#24686)
  wonhyeongseo
  authored
  b257c46
Commits on Jul 24, 2023
- [check_config_docstrings.py] improve diagnostics (#25012)
  stas00
  authored
  c9a82be
- [`logging.py`] set default `stderr` path if `None` (#25033)
  ArthurZucker
  authored
  0906d21
- fix(integrations): store serialized `TrainingArgs` to `wandb.config` without sanitization. (#25035)
  
  parambharat
  and
  Bharat Ramanathan
  authored
  54ba860
- [docs] Performance docs tidy up, part 1 (#23963)
  
  MKhalusova
  and
  stevhliu
  authored
  75317ae
- Support GatedRepoError + use raise from (#25034)
  
  Wauplin
  and
  sgugger
  authored
  6704923
- Better handling missing SYS in llama conversation tokenizer (#24997)
  ichernev
  authored
  efb2ba6
- 🌐[i18n-KO] Translated performance.md to Korean (#24883)
  
  augustinLib
  and
  kihoon71
  authored
  383be1b
- 🌐 [i18n-KO] Translated `testing.md` to Korean (#24900)
  Sunmin0520
  authored
  9d2b983
- Add dispatch_batches to training arguments (#25038)
  muellerzr
  authored
  3b734f5
- Fix typo in LlamaTokenizerFast docstring example (#25018)
  sbrunk
  authored
  8f1f0bf
- Make more test models smaller (#25005)
  sgugger
  authored
  42571f6
- Comment again print statement
  sgugger
  committed
  afe8bfc
- Pvt model (#24720)
  Xrenya
  authored
  a03d13c
- compute_loss in trainer failing to label shift for PEFT model when label smoothing enabled. (#25044)
  
  njbrake
  and
  Nathan Brake
  authored
  3611fc9
- [`8bit`] Fix 8bit corner case with Blip2 8bit (#25047)
  younesbelkada
  authored
  b08f41e
- 🌐 [i18n-KO] Translated `perf_train_cpu.md` to Korean (#24911)
  
  seank021
  and
  54data
  authored
  c0d1c33
- Better error message when signal is not supported on OS (#25049)
  sgugger
  authored
  d229570
Commits on Jul 25, 2023
Commits on Jul 26, 2023
- documentation for llama2 models (#25102)
  shauray8
  authored
  277d3ae
- 🌐[i18n-KO] Translated pipeline_webserver.md to Korean (#24828)
  
  authored
  ee63520
- Fix `PvtModelIntegrationTest::test_inference_fp16` (#25106)
  ydshieh
  authored
  31acba5
- Add descriptive docstring to TemperatureLogitsWarper (#24892)
  
  nablabits
  and
  gante
  authored
  04a5c85
- fix "UserWarning: Creating a tensor from a list of numpy.ndarrays is … (#24772)
  
  liucw2012
  and
  刘长伟
  authored
  c53c8e4
- update `use_auth_token` -> `token` (#25083)
  ydshieh
  authored
  224da5d
- Fix past CI after #24334 (#25113)
  ydshieh
  authored
  d30cf3d
- Move common image processing methods to BaseImageProcessor (#25089)
  asartran
  authored
  1486d2a
- Fix ViT docstring regarding default dropout values. (#25118)
  ebezzam
  authored
  b914ec9
- MaskFormer - enable return_dict in order to compile (#25052)
  asartran
  authored
  659829b
- Move center_crop to BaseImageProcessor (#25122)
  asartran
  authored
  1689aea
Commits on Jul 27, 2023
- fix deepspeed load best model at end when the model gets sharded (#25057)
  pacman100
  authored
  a004237
- fix delete all checkpoints when save_total_limit is set to 1 (#25136)
  Pbihao
  authored
  de9e3b5
- [`T5/LlamaTokenizer`] default legacy to `None` to not always warn (#25131)
  ArthurZucker
  authored
  9429642
- Clarify 4/8 bit loading log message (#25134)
  BramVanroy
  authored
  9a220ce
- 🚨🚨🚨Change default from `adamw_hf` to `adamw_torch` 🚨🚨🚨 (#25109)
  muellerzr
  authored
  a1c4954
- [`MptConfig`] support from pretrained args (#25116)
  
  ArthurZucker
  and
  younesbelkada
  authored
  9cea3e7
- Add offload support to Bark (#25037)
  
  authored
  0b92ae3
- More `token` things (#25146)
  ydshieh
  authored
  0c790dd
- Add bloom flax (#25094)
  
  authored
  e931036
- Add new model in doc table of content (#25148)
  sgugger
  authored
  400e76e
Commits on Jul 28, 2023
- Fix `.push_to_hub` and cleanup `get_full_repo_name` usage (#25120)
  Wauplin
  authored
  6232c38
- Add test when downloading from gated repo (#25039)
  Wauplin
  authored
  c1dba11
- override .cuda() to check if model is already quantized (#25166)
  ranchlai
  authored
  2a78720
- Represent query_length in a different way to solve jit issue (#25164)
  jiqing-feng
  authored
  d23d2c2
- make run_generation more generic for other devices (#25133)
  
  ji-huazhong
  and
  statelesshz
  authored
  afa96ff
- added compiled model support for inference (#25124)
  
  Alexander Markov
  and
  amarkov
  authored
  3cbc560
- Update `use_auth_token` -> `token` in example scripts (#25167)
  ydshieh
  authored
  d53b8ad
- [`Mpt`] Fix mpt slow test (#25170)
  younesbelkada
  authored
  add0895
- [`InstructBlip`] Fix instructblip slow test (#25171)
  younesbelkada
  authored
  dd9d45b
- 🌐 [i18n-KO] Translated `transformers_agents.md` to Korean (#24881)
  
  authored
  31f137c
- Fix beam search to sample at least 1 non eos token (#25103) (#25115)
  yonigottesman
  authored
  c90e14f
- [MusicGen] Fix integration tests (#25169)
  sanchit-gandhi
  authored
  03f98f9
- 🚨🚨🚨 Fix rescale ViVit Efficientnet (#25174)
  asartran
  authored
  05cda5d
Commits on Jul 31, 2023
- Musicgen: CFG is manually added (#25173)
  gante
  authored
  4a56449
- Better error message in `_prepare_output_docstrings` (#25202)
  ydshieh
  authored
  67b85f2
- [`PreTrainedModel`] Wrap `cuda` and `to` method correctly (#25206)
  younesbelkada
  authored
  59dcea3
- Fix `all_model_classes` in `FlaxBloomGenerationTest` (#25211)
  ydshieh
  authored
  9ca3aa0
- [quantization.md] fix (#25190)
  stas00
  authored
  5220606
- [`pipeline`] revisit device check for pipeline (#25207)
  younesbelkada
  authored
  e0c50b2
- Update tiny model info. and pipeline testing (#25213)
  ydshieh
  authored
  1b4f619
- Fix docker image build failure (#25214)
  ydshieh
  authored
  0fd8d2a
Commits on Aug 1, 2023
Commits on Aug 2, 2023
Commits on Aug 3, 2023
- Docs: Update list of `report_to` logging integrations in docstring (#25281)
  tomaarsen
  authored
  15082a9
- Update InstructBLIP & Align values after rescale update (#25209)
  asartran
  authored
  30409af
- Docs: separate generate section (#25235)
  gante
  authored
  a881737
- Update bark doc (#25234)
  
  ylacombe
  and
  sanchit-gandhi
  authored
  8455346
- add generate method to SpeechT5ForTextToSpeech (#25233)
  
  ylacombe
  and
  sgugger
  authored
  6d3f9c1
- Add timeout parameter to load_image function (#25184)
  
  rolisz
  and
  asartran
  authored
  d114a6b
- [JAX] Bump min version (#25286)
  sanchit-gandhi
  authored
  66c240f
- [small] llama2.md typo (#25295)
  H-Huang
  authored
  33da2db
- Fix typo: Roberta -> RoBERTa (#25302)
  MrGeislinger
  authored
  641adca
Commits on Aug 4, 2023
- Move usage of deprecated logging.warn to logging.warning (#25310)
  PeterJCLaw
  authored
  6768309
- Give more memory in test_disk_offload (#25315)
  sgugger
  authored
  fab1a0a
- Generate: get generation mode as an enum (#25292)
  gante
  authored
  bff4313
- Add offline mode for agents (#25226)
  sgugger
  authored
  aeb5a08
- Deal with nested configs better in base class (#25237)
  
  sgugger
  and
  younesbelkada
  authored
  29f0400
- Document check copies (#25291)
  sgugger
  authored
  f0fd73a
- Make `bark` could have tiny model (#25290)
  ydshieh
  authored
  ce6d153
- Document toc check and doctest check scripts (#25319)
  sgugger
  authored
  fdaef33
- [Whisper] Better error message for outdated generation config (#25298)
  sanchit-gandhi
  authored
  fdd81ae
- Remove jnp.DeviceArray since it is deprecated. (#24875)
  
  mariecwhite
  and
  sanchit-gandhi
  authored
  a6e6b1c
Commits on Aug 6, 2023
- add CFG for .generate() (#24654)
  Vermeille
  authored
  d533465
Commits on Aug 7, 2023
Commits on Aug 8, 2023
- [DOCS] Add descriptive docstring to MinNewTokensLength (#25196)
  
  nablabits
  and
  asartran
  authored
  a23ac36
- Register ModelOutput subclasses as supported torch.utils._pytree nodes (#25358)
  Matthew Hoffman
  authored
  d4bd33c
- Fix `test_model_parallelism` (#25359)
  ydshieh
  authored
  6ea3ee3
- Add warning for missing attention mask when pad tokens are detected (#25345)
  hackyon
  authored
  5ea2595
- [ASR Pipeline] Clarify return timestamps (#25344)
  sanchit-gandhi
  authored
  dedd111
- MaskFormer, Mask2Former - replace einsum for tracing (#25297)
  asartran
  authored
  36d5b8b
- Load state in else (#25318)
  muellerzr
  authored
  01ab39b
- Fix `token` in example template (#25351)
  ydshieh
  authored
  5744482
- Enable tests to run on third-party devcies (#25327)
  
  ji-huazhong
  and
  statelesshz
  authored
  26ce4dd
- 🌐 [i18n-KO] Translated `add_tensorflow_model.md` to Korean (#25017)
  keonju2
  authored
  6247d1b
- Fix `torch_job` worker(s) crashing (#25374)
  ydshieh
  authored
  9e57e0c
- Generate: add config-level validation (#25381)
  gante
  authored
  5bd8c01
- Fix missing usage of `token` (#25382)
  ydshieh
  authored
  9c7b744
- Use small config for `OneFormerModelTest.test_model_with_labels` (#25383)
  ydshieh
  authored
  5b517e1
- Add copied from for image processor methods (#25121)
  asartran
  authored
  e349010
- change version (#25387)
  SunMarc
  authored
  3a05e01
- [DOCS] Add example for `TopPLogitsWarper` (#25361)
  chiral-carbon
  authored
  41c5f45
Commits on Aug 9, 2023
- 🌐 [i18n-KO] Translated `perf_train_cpu_many.md` to Korean (#24923)
  
  nuatmochoi
  and
  wo0jun
  authored
  1367142
- 16059 - Add missing type hints for ASTModel (#25364)
  
  nablabits
  and
  Rocketknight1
  authored
  1564a81
- rm useless condition since the previous condition contains it. (#25403)
  jiqing-feng
  authored
  85447bb
- Fix path for dynamic module creation (#25402)
  sgugger
  authored
  5993771
- YOLOS - Revert default return_pixel_mask value (#25404)
  asartran
  authored
  ea5dda2
- Docs: introduction to generation with LLMs (#25240)
  
  authored
  d59b872
- Generate: length validation (#25384)
  gante
  authored
  3deed1f
- Improve training args (#25401)
  ji-huazhong
  authored
  00b93cd
- Generate: generation config validation fixes in docs (#25405)
  gante
  authored
  f456b4d
- 16059 - Add extra type hints for AltCLIPModel (#25399)
  nablabits
  authored
  ef74da6
- Generate: lower severity of parameterization checks (#25407)
  gante
  authored
  eb3ded1
- VQA task guide (#25244)
  
  authored
  f2a43c7
- 🌐 [i18n-KO] Translated `add_new_model.md` to Korean (#24957)
  
  authored
  133aac0
- 🌐 [i18n-KO] Translated `model_summary.md` to Korean (#24625)
  
  authored
  cf84738
- Update Bark generation configs and tests (#25409)
  ylacombe
  authored
  704bf59
- aligned sample_beam output selection with beam_search (#25375)
  hukuda222
  authored
  cb3c821
- Enable passing number of channels when inferring data format (#25412)
  asartran
  authored
  944ddce
- Bark: flexible generation config overload (#25414)
  gante
  authored
  d0c1aeb
Commits on Aug 10, 2023
- [DINOv2] Update pooler output (#25392)
  NielsRogge
  authored
  b175fc3
- 🌐 [i18n-KO] Translated `philosophy.md` to Korean (#25010)
  TaeYupNoh
  authored
  b14d464
- Doc checks (#25408)
  sgugger
  authored
  16edf4d
- Generation: strict generation config validation at save time (#25411)
  gante
  authored
  123ad53
- [WavLM] Fix Arxiv link and authors (#25415)
  sanchit-gandhi
  authored
  d0839f1
- Generate: Load generation config when `device_map` is passed (#25413)
  gante
  authored
  3e41cf1
- Fix rendering for `torch.compile()` docs (#25432)
  merveenoyan
  authored
  e7b001d
- Add `examples` to tests to run when `setup.py` is modified (#25437)
  ydshieh
  authored
  2d6839e
- Fix issue with ratio evaluation steps and auto find batch size (#25436)
  muellerzr
  authored
  a7da299
- docs: add LLaMA-Efficient-Tuning to awesome-transformers (#25441)
  
  ji-huazhong
  and
  statelesshz
  authored
  3470012
- GPTQ integration (#25062)
  
  authored
  55db70c
Commits on Aug 11, 2023
- Fix for #25437 (#25454)
  ydshieh
  authored
  454957c
- not debugged code
  
  zachares
  authored and
  vahanhov
  committed
  8db3720
- reference code so nothing is lost
  
  zachares
  authored and
  vahanhov
  committed
  7fcb1ab
- novelty
  
  zachares
  authored and
  vahanhov
  committed
  fe48c3f
- added docstrings
  
  zachares
  authored and
  vahanhov
  committed
  e47b6e8
- fixed some relative import errors
  
  zachares
  authored and
  vahanhov
  committed
  96d2814
- fixed small bugs
  
  zachares
  authored and
  vahanhov
  committed
  1090256
- added linear layers to bloom
  
  zachares
  authored and
  vahanhov
  committed
  5beaab7
- removed impossible embedding method
  
  zachares
  authored and
  vahanhov
  committed
  c7166f1
- Update src/transformers/models/bloom/desequence_graph_ids.py
  
  zachares
  and
  vahanhov
  committed
  d7052dd
- Update src/transformers/models/bloom/desequence_graph_ids.py
  
  zachares
  and
  vahanhov
  committed
  591cf9c
- feat(causal message passing) (#2)
  
  zachares
  authored and
  vahanhov
  committed
  9f156f0
- Clearer code and simpler method for within LLM message passing (#3)
  
  zachares
  and
  vahanhov
  committed
  3c1ca82
- memory efficient message passing (#4)
  
  zachares
  authored and
  vahanhov
  committed
  13affe7
- rebase from HF
  vahanhov
  committed
  4eb3009
- Merge branch 'main' into rebase-hf
  zachares
  authored
  c5e232d

Commits

Commits on Jul 18, 2023

Commits on Jul 19, 2023

Commits on Jul 20, 2023

Commits on Jul 21, 2023

Commits on Jul 24, 2023

Commits on Jul 25, 2023

Commits on Jul 26, 2023

Commits on Jul 27, 2023

Commits on Jul 28, 2023

Commits on Jul 31, 2023