huggingface/transformers

Pull Requests Commits

Disable head masking

Rocketknight1 committed 2 years ago

47eacad8

Fix inits, add autoclasses

Rocketknight1 committed 2 years ago

fd4c0f9c

Update tests, add dummies, make fixup

Rocketknight1 committed 2 years ago

8499b983

Add gptneo to docs

Rocketknight1 committed 2 years ago

4bd4c580

Add module imports

Rocketknight1 committed 2 years ago

869e0b5c

Avoid 'register_buffer', this is a PT method

Rocketknight1 committed 2 years ago

86a5bf59

Quick fix to the TF prefixes

Rocketknight1 committed 2 years ago

60c1b847

First commit of GPT-4 auto-port

Rocketknight1 committed 2 years ago

c8fa95b1

Pin tensorflow-text to go with tensorflow (#22362)

sgugger committed 2 years ago

Verified 6587125c

Update docker files to use official torch 2.0.0 (#22357)

ydshieh committed 2 years ago

Verified 01203475

Add Mega: Moving Average Equipped Gated Attention (#21766)

mnaylor5 committed 2 years ago

Verified 57f25f4b

Generate: Add GPTNeoX integration test (#22346)

gante committed 2 years ago

Verified 0fa46524

Fix typo in Greedy Search Description (#22345)

awinml committed 2 years ago

Verified b7960765

[HFTracer] Make embeddings ops take on the dtype of the weight (#22347)

jamesr66a committed 2 years ago

Verified c0fa2aa0

Automatically create/update tiny models (#22275)

ydshieh committed 2 years ago

Verified e8cc0255

Enable training Llama with model or pipeline parallelism (#22329)

kooshi committed 2 years ago

Verified a92e0ad2

Generate: add test for left-padding support (#22322)

gante committed 2 years ago

Verified 502fec77

Fix --bf16 option support for Neuron after PR #22300 (#22307)

jeffhataws committed 2 years ago

Verified ec9b18f6

Added type hints to TFDeiTModel (#22327)

Batese2001 committed 2 years ago

Verified aef488c5

Minor typo in pipeline FillMaskPipeline's documentation. (#22339)

SamuelLarkin committed 2 years ago

Verified 59b9351b

Fix various imports (#22281)

sgugger committed 2 years ago

Verified 506e7c63

Mention why one needs to specify max_steps in Trainer (#22333)

lhoestq committed 2 years ago

Verified 053c2153

Fixed gradient checkpoint bug for TimeSeriesTransformer (#22272)

pmollerus23 committed 2 years ago

Verified 5a9eb314

[`MBart`] Add `accelerate` support for MBart (#22309)

younesbelkada committed 2 years ago

Verified ff20f9cf

[gptj] support older pytorch version (#22325)

stas00 committed 2 years ago

Verified 61f79b29

Really fix quality due to ruff release

sgugger committed 2 years ago

Verified 80e3b363

Fix quality due to ruff release

sgugger committed 2 years ago

ef28df05

[deepspeed zero3] need `generate(synced_gpus=True, ...)` (#22242)

stas00 committed 2 years ago

Verified 73fdc8c5

Fix PipelineTests skip conditions (#22320)

ydshieh committed 2 years ago

Verified 8b05ace0

Chunkable token classification pipeline (#21771)

luccailliau committed 2 years ago

Verified d62e7d88

Older