Skip to content

GitLab

Switch branch/tag
  1. 02 May, 2024 1 commit
  3. 30 Apr, 2024 1 commit
  5. 07 Mar, 2024 1 commit
  7. 05 Mar, 2024 1 commit
  9. 16 Feb, 2024 1 commit
  11. 08 Feb, 2024 1 commit
  13. 15 Jan, 2024 1 commit
  15. 05 Jan, 2024 1 commit
  17. 08 Dec, 2023 1 commit
    • Tom Aarsen's avatar
      Generate: New `Cache` abstraction and Attention Sinks support (#26681) · 633215ba
      Tom Aarsen authored 
      * Draft version of new KV Caching

This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks)
/ StreamingLLM (https://arxiv.org/abs/2309.17453) to be easily implemented
in a third-party or in transformers directly

* Address numerous PR suggestions

1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic.
2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls.
3. Remove __bool__ and __getitem__ magic as they're confusing.
4. past_key_values.update(key, value, idx) now returns key, value.
5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR.
6. Separate key_cache and value_cache.

Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method.

* Implement the SinkCache through backward+forward rotations

* Integrate (Sink)Cache with Llama FA2

* Set use_legacy_cache=True as default, allows for test passes

* Move from/to_legacy_cache to ...Model class

* Undo unnecessary newline change

* Remove copy utility from deprecated OpenLlama

* Match import style

* manual rebase with main

* Cache class working with generate (#1)

* Draft version of new KV Caching

This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks)
/ StreamingLLM (https://arxiv.org/abs/2309.17453

) to be easily implemented
in a third-party or in transformers directly

* Address numerous PR suggestions

1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic.
2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls.
3. Remove __bool__ and __getitem__ magic as they're confusing.
4. past_key_values.update(key, value, idx) now returns key, value.
5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR.
6. Separate key_cache and value_cache.

Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method.

* Integrate (Sink)Cache with Llama FA2

* Move from/to_legacy_cache to ...Model class

* Undo unnecessary newline change

* Match import style

* working generate

* Add tests; Simplify code; Apply changes to Mistral and Persimmon

* fix rebase mess

* a few more manual fixes

* last manual fix

* propagate changes to phi

* upgrade test

* add use_legacy_cache docstring; beef up tests

* reintroduce unwanted deletes

---------
Co-authored-by: default avatarTom Aarsen <Cubiegamedev@gmail.com>

* move import

* add default to model_kwargs.get('use_legacy_cache')

* correct failing test

* Apply suggestions from code review
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>

* apply PR suggestions

* fix failing test

* Apply suggestions from code review
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: default avatarTom Aarsen <37621491+tomaarsen@users.noreply.github.com>

* PR comments

* tmp commit

* add docstrings

* more tests, more docstrings, add to docs

* derp

* tmp commit

* tmp dbg

* more dbg

* fix beam search bug

* cache can be a list of tuples in some models

* fix group beam search

* all but sinkcache integration tests

* fix sink cache and add hard integration test

* now also compatible with input_embeds input

* PR comments

* add Cache support to Phi+FA2

* make fixup

---------
Co-authored-by: default avatarJoao Gante <joao@huggingface.co>
Co-authored-by: default avatarJoao Gante <joaofranciscocardosogante@gmail.com>
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>
      633215ba
  19. 04 Dec, 2023 1 commit
  21. 25 Aug, 2023 1 commit
  23. 21 Jun, 2023 1 commit
  25. 20 Jun, 2023 1 commit
  27. 18 May, 2023 1 commit
  29. 09 May, 2023 1 commit
    • Matthijs Hollemans's avatar
      audio_utils improvements (#21998) · 7f919509
      Matthijs Hollemans authored 
      * silly change to allow making a PR

* clean up doc comments

* simplify hertz_to_mel and mel_to_hertz

* fixup

* clean up power_to_db

* also add amplitude_to_db

* move functions

* clean up mel_filter_bank

* fixup

* credit librosa & torchaudio authors

* add unit tests

* tests for power_to_db and amplitude_to_db

* add mel_filter_bank tests

* rewrite STFT

* add convenience spectrogram function

* missing transpose

* fewer transposes

* add integration test to M-CTC-T

* frame length can be either window or FFT length

* rewrite stft API

* add preemphasis coefficient

* move argument

* add log option to spectrogram

* replace M-CTC-T feature extractor

* fix api thing

* replace whisper STFT

* replace whisper mel filters

* replace tvlt's stft

* allow alternate window names

* replace speecht5 stft

* fixup

* fix integration tests

* fix doc comments

* remove manual FFT length calculation

* fix docs

* go away, deprecation warnings

* combine everything into spectrogram function

* add deprecated functions back

* fixup
      7f919509
  31. 03 Apr, 2023 1 commit
  33. 30 Mar, 2023 1 commit
  35. 07 Mar, 2023 1 commit
    • Eli Simhayev's avatar
      [Time-Series] informer model (#21099) · 8abe4930
      Eli Simhayev authored 
      * added informer to gitignore

* added informer to gitignore

* WIP informer2020

* added checking that instantiate works

* added config using gluonTS by kashif

* WIP config

* adding informeConfig. need to remove FeatureEmbedder

* done InformerConfig, but need to change the names

* Done informer model init. working on enc-dec

* added things to address, after reading again enc-dec in the paper

* done modeling - checking initialization work

* added informer to gitignore

* WIP informer2020

* added checking that instantiate works

* added config using gluonTS by kashif

* WIP config

* adding informeConfig. need to remove FeatureEmbedder

* done InformerConfig, but need to change the names

* Done informer model init. working on enc-dec

* added things to address, after reading again enc-dec in the paper

* done modeling - checking initialization work

* moved enc-dec init to InformerEncoder/Decoder init

* added 'init_std' to config, now model init works!

* WIP conversion script, and added code sources

* WIP conversion script: loading original informer pth works

* WIP conversion script: change defaults in the config

* WIP conversion script: supporting Informer input embedding

* WIP conversion script: added parameters for the informer embed

* WIP conversion script: change dim_feedforward=2048

* WIP conversion script: remove unused args for loading checkpoint

* just cleaning up

* DataEmbedding removed, after thinking with Kashif

* working on forward pass

* WIP forward pass: trying to establish working batch for forward pass

* cleaning and finalizing

* adding HF names and docs

* init after cleaning works

* WIP in tests

* added docs for the informer specific args

* fix style

* undo change

* cleaning informer, now need to work only enc-dec

* initial enc-dec classes

* added encoder and decoder

* added todo

* add todos for conv_layers

* added decoder docs from vanilla

* added encoder docs from vanilla

* remove encoder decoder from the original informer

* removed AttentionLayer from the original paper

* removed TriangularCausalMask, same as decoder_attention_mask

* initial sparse attention

* use conv_layers

* fixed test_config test

* fix parenthesis when itearting zip(layers, conv_layers)

* error found in prob attention, added sizes as comments

* fix sizes

* added proposal for q_reduce indexing, and remove unused

* WIP ProbMask, and changed factor=2 for testing

* remove unused libs for this PR for creating the env

* fix checking the attn_weights.size() after bmm

* Q_reduce: changed from torch.gather to simple slicing

* WIP calculate final attn_output

* finish adding v_aggregated, attn_output ready

* changed tgt_len to u in attention_mask, need to fix the size error

* comment attention_mask for encoder, and fix if cond for v_agg

* added ProbMask support (wip), removed old original code

* finished ProbMask 😃



* Revert "remove unused libs for this PR for creating the env"

This reverts commit 11a081e09e92771e51a5d2758d53a9afb59547f0.

* fixes

* make style

* fix initial tests

* fix more tests

* dry

* make style

* remove unused files

* style

* added integration tests

* fix num_static_real_features

* fix header

* remove unused function

* fix example

* fix docs

* Update src/transformers/models/informer/configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/informer/modeling_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/informer/configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/informer/configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/informer/configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Update src/transformers/models/informer/configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* fixes for reviewer

* use prediction_length from model

* fix style

* fixed informer.mdx

* added to index

* updated readme

* undo

* make fix-copies

* typo

* fix copy

* added Informer to toctree

* in order

* fixed comments

* remove unneeded new lines in docs

* make static real and cat optional

* fix use of distil conv layers

* fixed integration test

* added checkpoint for convlayer

* make fix-copies

* updated from time series model

* make fix-copies

* copy decoder

* fix unit tests

* updated scaling config

* fix integration tests

* IGNORE_NON_TESTED

* IGNORE_NON_AUTO_CONFIGURED

* IGNORE_NON_AUTO_CONFIGURED

* updated check configs

* fix formatting

* undo change from time series

* prediction_length should not be None

* aliign with the blog: prettify ProbSparse and change attention_factor  to sampling_factor

* make style

* make fix-copies

* niels CR: update contributed by

* niels CR: update configuration_informer.py
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* niels CR: update kashif -> huggingface
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>

* niels CR: `sampling_factor` only relevant when `attention_type`=prob

* make style

* fixed U_part: added multiplication by `L_Q`

* fixed bug: remove `is not None` from `if config.distil`

* fixed test: `decoder_seq_length` to `encoder_seq_length` in cross_attentions check

* fix integration tests

* updated model hub

* do not shift as in training

* undo

* fix make-copies

* make fix-copies

* added `if prediction_length is None`

* changed `ProbSparseAttention` to `InformerProbSparseAttention`

* changed `V_sum` -> `v_mean_dim_time`

* changed `ConvLayer` to `InformerConvLayer` and fixed `super()`

* TimeSeriesTansformer->Informer in decoder's Copied from

* more descriptive in ProbSparse

* make style

* fix coped from

* Revert "added `if prediction_length is None`"

This reverts commit b4cbddfa05e3bd739b79569cd3c3b89e316f2451.

* fixed indent

* use InformerSinusoidalPositionalEmbedding

* make fix-style

* fix from #21860

* fix name

* make fix-copies

* use time series utils

* fix dec num_heads

* docstring

* added time series util doc

* _import_structure

* formatting

* changes from review

* make style

* fix docs

* fix doc

* removed NegativeLogLikelihood

---------
Co-authored-by: default avatarKashif Rasul <kashif.rasul@gmail.com>
Co-authored-by: default avatarNielsRogge <48327001+NielsRogge@users.noreply.github.com>
      8abe4930
  37. 16 Feb, 2023 1 commit
    • Arthur's avatar
      [CLAP] Add CLAP to the library (#21370) · c236a621
      Arthur authored 
      

* add model like clip

* update

* text model ok

* clap text works

* some refactor

- `CLAPVision` to `CLAPAudio`
- refactor kwargs of audio modules

* more refactor

* more refactor

* more refactor

* correct fusion

* more refactor

* new modules

* add basic processor

* fixup

* remove whisper copioed from

* audio logits match

* add doc

* correct filters mel and add maxlength

* style

* few fixes

* forward passes

* fixup

* fixup

* some clean up

* remove mels form the dictionnary

* pad after the repeat

* update padding when dsmaller

* fix padding

* style

* use swin patch merging

* use copied from swin

* processor with any tokenizer

* more copied from

* some clean up

* more refactor

* fix mel when rand_trunc

* style

* remove unused imports

* update processing

* remove image processing tests

* add testing fiel

* fixmodeling issues

* replace with `is_longer`

* clap in serialization

* more refactor

* `make fixup`

* make fixup

* fix feature extractor

* update test feature extractor

* `make fixup`

* clean up config

* more clean up

* more cleanup

* update tests

* refactor tests and inits

* removeCLAP vision config

* remove CLAP from image procssing auto and dummy vision objects

* update inits

* style

* re order classes in modeling clap

* Use roberta tokenizer as the other weights are not open sourced

* small cleaup

* remove tokenization CLAP

* processor tokenizr is roberta

* update feature extraction doc

* remove vclap from model zero shot

* update f_min and f_max to frequency_xx

* some changes

- fix modeling keys
- add `is_longer` in the forward pass
- make fixup

* make fixup

* consistent behavior ebtween rand_crop and fusion

* add numpy resize and bilinear and documentation

* move resizing to image utils

* clean feature extraction

* import resize from correct file

* resize in image transforms

* update

* style

* style

* nit

* remove unused arguments form the feature extractor

* style

* few fixes + make fixup

* oops

* fix more tests

* add zero shot audio classification pipeline

* update zeroshot classification pipeline

* fixup

* fix copies

* all CI tests pass

* make fixup + fix docs

* fix docs

* fix docs

* update tests pip;eline

* update zero shot pipeline

* update feature extraction clap

* update tokenization auto

* use nested simplify

* update pipeline tests

* Apply suggestions from code review
Co-authored-by: default avatarArthur <48595927+ArthurZucker@users.noreply.github.com>

* split in two lines

* fixes

* refactor

* clean up

* add integration tests

* update config docstring

* style

* update processor

* fix processor test

* fix feat extractor tests

* update docs

* Apply suggestions from code review
Co-authored-by: default avatarArthur <48595927+ArthurZucker@users.noreply.github.com>

* fix readmes

* fix tips

* Update src/transformers/models/auto/configuration_auto.py

* update doc and remove todo -> properly explained

* fix idx and typo

* typoe

* cleanup config

* cleanup tests, styles and doc

* ignore docstyle on image transform

* add conversion script

* remove the `clap` indx in favor of `CLAP`

* update __init

* nits

* Update src/transformers/pipelines/__init__.py

* fix bug

* clarifiy config

* fix copy

* fix init

* Apply suggestions from code review
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix model output

* fix comment

* make fixup

* make fixup

* rename to `Clap`

* replace to `Clap`

* replace to `Clap`

* repo consistency

* again repo-consistency

* make fixup

* Apply suggestions from code review
Co-authored-by: default avatarSanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* add config

* changes

* update conversion

* Apply suggestions from code review
Co-authored-by: default avatarSanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* remove unused function

* update based on code reviews

* style

* more comments

* cleanup

* clean up

* style

* apply suggestions

* Empty commit

* pipeline will be added in a different PR

* update calls to audio utils functions

* update pipeline init

* style

* style

* styling again

* use pad

* fix repo-consistency

* update utils and add doc for audio utils

* clean up resize by using torch. update inits accordingly

* style

* CLap's  tokenizer is RobertA

* add audio utils to internal toctreee

* update totctree

* style

* update documentation and normalize naming accross audio utils and feature extraction clap

* style

* clean up

* update doc and typos

* fix doctest

* update modelin code, got rid of a lot of reshaping

* style on added doc audio utils

* update modeling clap

* style

* Apply suggestions from code review
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

* docstringvariables with CLAP

* rename key

* update modeling CLAP

* update audio utils docstring

* update processing clap

* fix readmes

* fix toctree

* udpate configuration clap

* fix init

* make fixup

* fix

* fix

* update naming

* update

* update checkpoint path

* Apply suggestions from code review

* Major refactoring

* Update src/transformers/models/clap/configuration_clap.py

* merge

---------
Co-authored-by: default avataryounesbelkada <younesbelkada@gmail.com>
Co-authored-by: default avatarYounes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: default avatarSanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
      c236a621
  39. 03 Jan, 2023 1 commit
    • Konstantin Kotik's avatar
      `MinNewTokensLengthLogitsProcessor` for `.generate` method #20814 (#20892) · 367fdf33
      Konstantin Kotik authored 
      

* feat: add min new length logit processor

* test: add min new length logit processor

* docs: add MinNewTokensLengthLogitsProcessor

* feat: import MinNewTokensLengthLogitsProcessor

* fix: update pytorch dummy objects

* refactor & fix: rename attributes and var and get rid of dynamic attribute

* tests: align test with new interface

* docs: fix typo

* docs: minor clarification

* Empty-Commit

* empty commit

* run automated quality edits
Co-authored-by: default avatarJoao Gante <joao@huggingface.co>
      367fdf33
  41. 18 Nov, 2022 1 commit
  43. 15 Nov, 2022 1 commit
  45. 09 Nov, 2022 1 commit
  47. 08 Nov, 2022 1 commit
    • amyeroberts's avatar
      AutoImageProcessor (#20111) · 4eb918e6
      amyeroberts authored 
      * AutoImageProcessor skeleton

* Update references

* Add mapping in init

* Add model image processors to __init__ for importing

* Add AutoImageProcessor tests

* Fix up

* Image Processor documentation

* Remove pdb

* Update docs/source/en/model_doc/mobilevit.mdx

* Update docs

* Don't add whitespace on json files

* Remove fixtures

* Move checking model config down

* Fix up

* Add check for image processor

* Remove FeatureExtractorMixin in docstrings

* Rename model_tmpfile to config_tmpfile

* Don't make None if not in image processor map
      4eb918e6
  49. 19 Oct, 2022 2 commits
    • amyeroberts's avatar
      Image transforms add center crop (#19718) · 5041bc35
      amyeroberts authored 
      

* Add center crop to transforms library

* Return PIL images if PIL image input by default

* Fixup and add docstring

* Trigger CI

* Update src/transformers/image_transforms.py
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/image_transforms.py
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

* PR comments - move comments; unindent
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>
      5041bc35
    • GMFTBY's avatar
      Adding the state-of-the-art contrastive search decoding methods for the... · 71786b10
      GMFTBY authored 
      Adding the state-of-the-art contrastive search decoding methods for the codebase of generation_utils.py (#19477)

* add: the contrastive search for generaton_utils

* add: testing scripts for contrastive search under examples/text-generation

* update the quality of codes

* revise the docstring; make the generation_contrastive_search.py scripts;

* revise the examples/pytorch/text-generation/run_generation_contrastive_search.py to the auto-APIs format

* revise the necessary documents

* fix: revise the docstring of generation_contrastive_search.py

* Fix the code indentation

* fix: revise the nits and examples in contrastive_search docstring.

* fix the copyright

* delete generation_contrastive_search.py

* revise the logic in contrastive_search

* update the intergration test and the docstring

* run the tests over

* add the slow decorate to the contrastive_search intergrate test

* add more test

* do the style, quality, consistency checks
      71786b10
  51. 17 Oct, 2022 1 commit
  53. 12 Oct, 2022 1 commit
  55. 03 Jun, 2022 1 commit
    • Robert Dargavel Smith's avatar
      Allow from transformers import TypicalLogitsWarper (#17477) · 5c17918f
      Robert Dargavel Smith authored 
      * Allow from transformers import TypicalLogitsWarper

* Added TypicalLogitsWarper

* Allow from transformers import TypicalLogitsWarper

* Allow from transformers import TypicalLogitsWarper

* Allow from transformers import TypicalLogitsWarper

* Allow from transformers import TypicalLogitsWarper

Added TypicalLogitsWarper

Allow from transformers import TypicalLogitsWarper

Allow from transformers import TypicalLogitsWarper

Allow from transformers import TypicalLogitsWarper
      5c17918f
  57. 11 May, 2022 1 commit
  59. 12 Apr, 2022 1 commit
    • Anmol Joshi's avatar
      Moved functions to pytorch_utils.py (#16625) · a315988b
      Anmol Joshi authored 
      * Moved functions to pytorch_utils.py

* isort formatting

* Reverted tf changes

* isort, make fix-copies

* documentation fix

* Fixed Conv1D import

* Reverted research examples file

* backward compatibility for pytorch_utils

* missing import

* isort fix
      a315988b
  61. 06 Apr, 2022 1 commit
  63. 04 Apr, 2022 1 commit