Poedator a0779b9e19 Llama: fix custom 4D masks, v2 (#30348) ...

* 4d mask fixes

* Update custom 4D mask logic

* test moved to mixin

* extra tests 4d mask

* upd 4d mask and StaticCache handling

* added Mask4DTestHard to mistral tests

* post-rebase fixes

* test fixes for StaticCache

* make fix-copies

* upd 1 after #30476

* fix common tests

* rm elif attention_mask.dim() == 4:

* tests combined, fixed, mixtral supported

* bigbird style chg reverted

* rm if attention_mask.dim() == 2

* modeling_llama formatting chg

---------

Co-authored-by: Joao Gante <joao@huggingface.co>

2024-05-13 13:46:06 +02:00

..

albert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

align

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

altclip

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

audio_spectrogram_transformer

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

auto

Fix auto tests (#30067)

2024-04-05 17:49:46 +02:00

autoformer

Add tests for batching support (#29297)

2024-03-12 17:46:19 +00:00

bark

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

bart

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

barthez

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

bartpho

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

beit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

bert

[BERT] Add support for sdpa (#28802)

2024-04-26 16:23:44 +01:00

bert_generation

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

bert_japanese

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

bertweet

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

big_bird

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

bigbird_pegasus

Do not remove half seq length in generation tests (#30016)

2024-04-19 17:32:52 +01:00

biogpt

[generate] fix breaking change for patch (#29976)

2024-04-02 09:51:45 +02:00

bit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

blenderbot

Generate: left-padding test, revisited (#29515)

2024-03-08 10:06:46 +00:00

blenderbot_small

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

blip

Blip dynamic input resolution (#30722)

2024-05-13 12:20:16 +01:00

blip_2

Blip dynamic input resolution (#30722)

2024-05-13 12:20:16 +01:00

bloom

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

bridgetower

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

bros

[tests] add the missing require_torch_multi_gpu flag (#30250)

2024-04-15 16:30:52 +01:00

byt5

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

camembert

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

canine

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

chinese_clip

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

clap

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

clip

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

clipseg

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

clvp

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

code_llama

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

codegen

Add token type ids to CodeGenTokenizer (#29265)

2024-04-17 12:19:18 +02:00

cohere

Cache: models return input cache type (#30716)

2024-05-08 18:26:34 +01:00

conditional_detr

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

convbert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

convnext

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

convnextv2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

cpm

Fix PipelineTests skip conditions (#22320)

2023-03-22 20:02:24 +01:00

cpmant

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

ctrl

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

cvt

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

data2vec

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

dbrx

Cache: models return input cache type (#30716)

2024-05-08 18:26:34 +01:00

deberta

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

deberta_v2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

decision_transformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

deformable_detr

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

deit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

depth_anything

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

deta

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

detr

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

dinat

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

dinov2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

distilbert

Fix FA2 tests (#29909)

2024-04-01 07:51:00 +00:00

dit

Update old existing feature extractor references (#24552)

2023-06-29 10:17:36 +01:00

donut

Removal of deprecated maps (#30576)

2024-05-09 14:15:56 +02:00

dpr

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

dpt

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

efficientformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

efficientnet

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

electra

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

encodec

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

encoder_decoder

Generate: missing generation config eos token setting in encoder-decoder tests (#29146)

2024-02-20 16:17:51 +00:00

ernie

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

ernie_m

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

esm

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

falcon

Support for Falcon2-11B (#30771)

2024-05-13 13:32:43 +02:00

fastspeech2_conformer

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

flaubert

Fix 29807 sinusoidal positional encodings in Flaubert, Informer and XLM (#29904)

2024-04-02 10:27:26 +02:00

flava

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

fnet

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

focalnet

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

fsmt

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

funnel

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

fuyu

Update tiny model summary file (#27388)

2023-11-23 21:00:39 +01:00

gemma

Cache: models return input cache type (#30716)

2024-05-08 18:26:34 +01:00

git

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

glpn

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

gpt2

Adding Flash Attention 2 Support for GPT2 (#29226)

2024-03-28 09:31:24 +00:00

gpt_bigcode

CI with num_hidden_layers=2 🚀🚀🚀 (#25266)

2023-08-02 20:22:36 +02:00

gpt_neo

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

gpt_neox

RoPE models: add numerical sanity-check test for RoPE scaling (#29808)

2024-03-28 11:25:50 +00:00

gpt_neox_japanese

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

gpt_sw3

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

gptj

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

gptsan_japanese

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

graphormer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

grounding_dino

[Grounding DINO] Add support for cross-attention in GroundingDinoMultiHeadAttention (#30364)

2024-04-23 09:56:14 +01:00

groupvit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

herbert

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

hubert

Fix failing tests on main due to torch 2.1 (#26607)

2023-10-05 10:27:05 +02:00

ibert

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

idefics

[ProcessingIdefics] Attention mask bug with padding (#29449)

2024-04-04 10:11:09 +01:00

idefics2

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

imagegpt

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

informer

Fix 29807 sinusoidal positional encodings in Flaubert, Informer and XLM (#29904)

2024-04-02 10:27:26 +02:00

instructblip

Blip dynamic input resolution (#30722)

2024-05-13 12:20:16 +01:00

jamba

Jamba: fix left-padding test (#30389)

2024-04-22 17:02:55 +01:00

jukebox

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

kosmos2

Remove use_square_size after loading (#30567)

2024-04-30 21:11:37 +02:00

layoutlm

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

layoutlmv2

[CI update] Try to use dockers and no cache (#29202)

2024-05-06 10:10:32 +02:00

layoutlmv3

[CI update] Try to use dockers and no cache (#29202)

2024-05-06 10:10:32 +02:00

layoutxlm

Add correct batched handling for apply_chat_template (#29222)

2024-03-20 15:50:22 +00:00

led

Do not remove half seq length in generation tests (#30016)

2024-04-19 17:32:52 +01:00

levit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

lilt

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

llama

Llama: fix custom 4D masks, v2 (#30348)

2024-05-13 13:46:06 +02:00

llava

Fix llava half precision and autocast issues (#29721)

2024-05-01 17:49:44 +01:00

llava_next

Fix llava half precision and autocast issues (#29721)

2024-05-01 17:49:44 +01:00

longformer

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

longt5

Do not remove half seq length in generation tests (#30016)

2024-04-19 17:32:52 +01:00

luke

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

lxmert

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

m2m_100

Add Flash Attention 2 to M2M100 model (#30256)

2024-04-18 10:27:58 +02:00

mamba

Mamba slow_forward gradient fix (#29563)

2024-03-27 04:52:12 +01:00

marian

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

markuplm

Add correct batched handling for apply_chat_template (#29222)

2024-03-20 15:50:22 +00:00

mask2former

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

maskformer

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

mbart

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

mbart50

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

mega

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

megatron_bert

CI with num_hidden_layers=2 🚀🚀🚀 (#25266)

2023-08-02 20:22:36 +02:00

megatron_gpt2

Move test model folders (#17034)

2022-05-03 14:42:02 +02:00

mgp_str

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

mistral

Llama: fix custom 4D masks, v2 (#30348)

2024-05-13 13:46:06 +02:00

mixtral

Enable fx tracing for Mistral (#30209)

2024-04-17 14:38:48 +05:00

mluke

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

mobilebert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mobilenet_v1

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mobilenet_v2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mobilevit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mobilevitv2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mpnet

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

mpt

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mra

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

mt5

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

musicgen

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

musicgen_melody

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

mvp

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

nat

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

nezha

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

nllb

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

nllb_moe

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

nougat

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

nystromformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

olmo

Cache: models return input cache type (#30716)

2024-05-08 18:26:34 +01:00

oneformer

Fix OneFormer post_process_instance_segmentation for panoptic tasks (#29304)

2024-03-04 11:04:49 +00:00

openai

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

opt

Proper build() methods for TF (#27794)

2023-12-14 15:17:30 +00:00

owlv2

Fix image post-processing for OWLv2 (#30686)

2024-05-09 17:02:03 +01:00

owlvit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

patchtsmixer

PatchtTST and PatchTSMixer fixes (#28083)

2024-01-29 10:09:26 +00:00

patchtst

PatchtTST and PatchTSMixer fixes (#28083)

2024-01-29 10:09:26 +00:00

pegasus

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

pegasus_x

device agnostic models testing (#27146)

2023-10-31 18:12:14 +01:00

perceiver

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

persimmon

[tests] add require_bitsandbytes marker (#30116)

2024-04-08 12:49:31 +01:00

phi

RoPE models: add numerical sanity-check test for RoPE scaling (#29808)

2024-03-28 11:25:50 +00:00

phi3

Phi-3 (#30423)

2024-04-24 17:32:09 +02:00

phobert

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

pix2struct

BLIP - fix pt-tf equivalence test (#30258)

2024-04-16 17:46:53 +01:00

plbart

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

poolformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

pop2piano

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

prophetnet

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

pvt

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

pvt_v2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

qdqbert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

qwen2

Enable fx tracing for Mistral (#30209)

2024-04-17 14:38:48 +05:00

qwen2_moe

Enable fx tracing for Mistral (#30209)

2024-04-17 14:38:48 +05:00

rag

Add dataset_revision argument to RagConfig (#29610)

2024-03-14 16:48:11 +01:00

realm

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

recurrent_gemma

Cache: models return input cache type (#30716)

2024-05-08 18:26:34 +01:00

reformer

Do not remove half seq length in generation tests (#30016)

2024-04-19 17:32:52 +01:00

regnet

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

rembert

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

resnet

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

roberta

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

roberta_prelayernorm

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

roc_bert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

roformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

rwkv

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

sam

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

seamless_m4t

Generate: consistently handle special tokens as tensors (#30624)

2024-05-09 18:01:57 +01:00

seamless_m4t_v2

Generate: consistently handle special tokens as tensors (#30624)

2024-05-09 18:01:57 +01:00

segformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

seggpt

[SegGPT] Fix seggpt image processor (#29550)

2024-04-26 19:40:12 +01:00

sew

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

sew_d

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

siglip

Add dynamic resolution input/interpolate position embedding to SigLIP (#30719)

2024-05-09 11:10:38 +01:00

speech_encoder_decoder

Generate: missing generation config eos token setting in encoder-decoder tests (#29146)

2024-02-20 16:17:51 +00:00

speech_to_text

Generate: consistently handle special tokens as tensors (#30624)

2024-05-09 18:01:57 +01:00

speech_to_text_2

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

speecht5

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

splinter

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

squeezebert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

stablelm

[StableLm] Add QK normalization and Parallel Residual Support (#29745)

2024-04-08 23:51:58 +02:00

starcoder2

Fix FA2 tests (#29909)

2024-04-01 07:51:00 +00:00

superpoint

Removal of deprecated maps (#30576)

2024-05-09 14:15:56 +02:00

swiftformer

Removal of deprecated maps (#30576)

2024-05-09 14:15:56 +02:00

swin

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

swin2sr

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

swinv2

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

switch_transformers

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

t5

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

table_transformer

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

tapas

[CI update] Try to use dockers and no cache (#29202)

2024-05-06 10:10:32 +02:00

time_series_transformer

Add tests for batching support (#29297)

2024-03-12 17:46:19 +00:00

timesformer

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

timm_backbone

Adding _tie_weights() to prediction heads to support low_cpu_mem_usage=True (#29024)

2024-05-07 11:12:21 +02:00

trocr

CI with num_hidden_layers=2 🚀🚀🚀 (#25266)

2023-08-02 20:22:36 +02:00

tvlt

fix: Replace deprecated assertEquals with assertEqual (#30241)

2024-04-15 09:36:06 +01:00

tvp

Enable instantiating model with pretrained backbone weights (#28214)

2024-01-23 11:01:50 +00:00

udop

[UDOP] Add special tokens to tokenizer (#29594)

2024-04-19 09:06:01 +02:00

umt5

Tests: replace torch.testing.assert_allclose by torch.testing.assert_close (#29915)

2024-03-28 09:53:31 +00:00

unispeech

[Wav2Vec2 and Co] Update init tests for PT 2.1 (#26494)

2023-10-03 10:52:34 +02:00

unispeech_sat

Byebye torch 1.10 (#28207)

2024-01-11 16:18:27 +01:00

univnet

Add tests for batching support (#29297)

2024-03-12 17:46:19 +00:00

upernet

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

videomae

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vilt

Encoder-decoder models: move embedding scale to nn.Module (#30410)

2024-05-01 12:33:00 +05:00

vipllava

Use text config's vocab size in testing models (#30568)

2024-05-01 12:32:45 +05:00

vision_encoder_decoder

Fix VisionEncoderDecoder Positional Arg (#29497)

2024-03-07 20:45:51 +00:00

vision_text_dual_encoder

[Styling] stylify using ruff (#27144)

2023-11-16 17:43:19 +01:00

visual_bert

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vit

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vit_hybrid

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vit_mae

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vit_msn

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

vitdet

mark test_initialization as flaky in 2 model tests (#27906)

2023-12-08 14:54:32 +01:00

vitmatte

🚨 Update image_processing_vitmatte.py (#30566)

2024-05-02 11:00:07 +01:00

vits

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

vivit

Enable dynamic resolution for vivit (#30630)

2024-05-09 11:23:39 +01:00

wav2vec2

Add sdpa and fa2 the Wav2vec2 family. (#30121)

2024-04-22 18:30:38 +01:00

wav2vec2_bert

Add new meta w2v2-conformer BERT-like model (#28165)

2024-01-18 13:37:34 +00:00

wav2vec2_conformer

device agnostic models testing (#27146)

2023-10-31 18:12:14 +01:00

wav2vec2_phoneme

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

wav2vec2_with_lm

Fix some tests using "common_voice" (#27147)

2023-10-30 15:27:15 +01:00

wavlm

Output None as attention when layer is skipped (#30597)

2024-05-02 17:25:19 +01:00

whisper

Generate: consistently handle special tokens as tensors (#30624)

2024-05-09 18:01:57 +01:00

x_clip

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

xglm

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

xlm

Fix 29807 sinusoidal positional encodings in Flaubert, Informer and XLM (#29904)

2024-04-02 10:27:26 +02:00

xlm_prophetnet

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

xlm_roberta

Adds pretrained IDs directly in the tests (#29534)

2024-03-13 14:53:27 +01:00

xlm_roberta_xl

Revert low cpu mem tie weights (#29135)

2024-02-20 12:06:46 +00:00

xlnet

Do not remove half seq length in generation tests (#30016)

2024-04-19 17:32:52 +01:00

xmod

Update all references to canonical models (#29001)

2024-02-16 08:16:58 +01:00

yolos

Fix YOLOS image processor resizing (#30436)

2024-04-24 09:50:17 +01:00

yoso

Remove static pretrained maps from the library's internals (#29112)

2024-03-25 10:33:38 +01:00

__init__.py

Move test model folders (#17034)

2022-05-03 14:42:02 +02:00