NielsRogge/transformers

Pull Requests Commits

Add Blip2 as well

NielsRogge committed 2 years ago

63a5d877

Add blip to pipeline

NielsRogge committed 2 years ago

79bb6292

[`Blip2`] Add int8 support for `blip2-flan-t5-xxl` (#21574)

younesbelkada committed 2 years ago

Verified 75a208ef

Remove more unused attributes in config classes (#21543)

ydshieh committed 2 years ago

Verified b47a1674

Added timesformer configuration (#21446)

AdiaWu committed 2 years ago

Verified 862e8e4f

Replace input_values_processing with unpack_inputs (#21502)

amyeroberts committed 2 years ago

Verified cb565901

improving contributing tests section (#21569)

Shubhamai committed 2 years ago

Verified 55712563

[deepspeed] deal with models w/o `config.hidden_size` (#21504)

stas00 committed 2 years ago

Verified 3b7ed25d

Goodbye to Blip-2 doctests (#21566)

ydshieh committed 2 years ago

Verified 4f831e66

[Tasks] Adds image captioning (#21512)

sayakpaul committed 2 years ago

Verified e2ec3089

[from_pretrained] extend `torch_dtype="auto"` to look up `config.torch_dtype` first, expand docs (#21524)

stas00 committed 2 years ago

Verified 2f550758

[Tests] Improve flax test_attention_outputs (#21486)

Shubhamai committed 2 years ago

Verified 9e40bba6

Add _mp_fn to run_mae.py for XLA testing (#21551)

steventk-g committed 2 years ago

Verified c88b11c5

[Variant] Make sure variant files are not incorrectly deleted (#21562)

patrickvonplaten committed 2 years ago

Verified b20147a3

Replace inefficient torch.sqrt taking scalar input with numpy.sqrt (#21496)

Yueming Hao committed 2 years ago

Verified 51c3f42d

Add X-MOD (#20939)

jvamvas committed 2 years ago

Verified b0d539cc

Fix stuff related to the causal_mask in CodeGen. (#21527)

GeneZC committed 2 years ago

Verified adb2503e

Remove CLI spams with Whisper FeatureExtractor (#21267)

qmeeus committed 2 years ago

Verified 5b72b341

adding a tip for deepspeed integration in multi-node environment (#21459)

izapolsk committed 2 years ago

Verified 129011c2

Added with torch.no_grad() to Camembert integration test (#21544)

katiele47 committed 2 years ago

Verified 21a2d900

[`pipeline`] A simple fix for half-precision & 8bit models (#21479)

younesbelkada committed 2 years ago

Verified f8394268

Skip failing test for now

sgugger committed 2 years ago

Verified 97d3390f

Added with torch.no_grad() to XLM-Roberta integration test (#21547)

katiele47 committed 2 years ago

Verified 23c146c3

🚨🚨🚨 Enforce single model initialization (#21431)

sgugger committed 2 years ago

Verified 04b2f13c

Fix from_pretrained API with config and state_dict (#21542)

sgugger committed 2 years ago

Verified 2020ac4b

Fix inclusion of non py files in package (#21546)

sgugger committed 2 years ago

Verified 1efe9c0b

Align BLIP-2 winit with others

sgugger committed 2 years ago

Verified 7927732f

Add BLIP-2 (#21441)

NielsRogge committed 2 years ago

Verified d7f1e7c0

fix typo in run_speech_recognition_ctc.py (#21528)

21jun committed 2 years ago

Verified b31cee67

Tag tests as slow ⌛ (#21537)

gante committed 2 years ago

Verified 0d33381f

Older