huggingface/optimum

Pull Requests Commits

thomasw21 committed 2 years ago

39cff2af

Add debugging code to undertand where the discrepancy comes from

thomasw21 committed 2 years ago

c809546d

WIP: fp32 forward between TP=1 and TP=2 are not exactly the same on cpu ...

thomasw21 committed 2 years ago

306299f0

thomasw21 committed 2 years ago

b38c4f97

remove transformers main in benchmark test (#330)

fxmarty committed 2 years ago

Verified ee835054

OnnxRuntime tests refactor to make them faster (#315)

michaelbenayoun committed 2 years ago

Verified db845bd1

FuseBiasInLinear transformation (#253)

michaelbenayoun committed 2 years ago

Verified d03433ea

MergeLinears set proper device and dtype (#313)

michaelbenayoun committed 2 years ago

Verified 23cbf75b

Enable possibility to not use past_key_values (#241)

echarlaix committed 2 years ago

Verified 8d64d8aa

Update clm example for ORTTrainer (#316)

JingyaHuang committed 2 years ago

Verified 840e5c7d

Add Graphcore and Intel documentation (#311)

lewtun committed 2 years ago

Verified 596000b9

Fix benchmark tests (#298)

fxmarty committed 2 years ago

Verified 553f588e

added documentation optimizedmodel (#280)

fxmarty committed 2 years ago

Verified b0a44550

Fix wrong mapping between dataset label ids and model.config.label2id in examples (#197)

fxmarty committed 2 years ago

Verified a15f3e9e

Add ORT fused adam optimizer (#295)

JingyaHuang committed 2 years ago

Verified de2a1f1d

Fix style (#309)

echarlaix committed 2 years ago

Verified 1e32570a

Replace optimizer.tokenizer by optimizer.preprocessor in examples (#304)

regisss committed 2 years ago

Verified 3c04b793

Update dockerfile (#307)

JingyaHuang committed 2 years ago

Verified 411e9e67

Multi package docs (#299)

lewtun committed 2 years ago

Verified 555f0dac

Fix bug about accessing model.config when model is None in pipeline (#301)

regisss committed 2 years ago

Verified fab5d455

Add CLM training example (#248)

carzh committed 2 years ago

Verified 7c1a6211

Add the possibility to use an authentication token when loading the calibration dataset (#289)

regisss committed 2 years ago

Verified 6911cc65

add preprocessors for benchmark (#284)

fxmarty committed 2 years ago

Verified 5c9af4e5

Fix documentation for ORTModelForImageClassification (#283)

regisss committed 2 years ago

Verified 5ac9c0d9

Add lazy imports in optimum/onnxruntime (#269)

regisss committed 2 years ago

Verified b2228503

ORT-OPTIMUM (#229)

zhijxu-MS committed 2 years ago

Verified e836c9b8

Remove duplicated key in supported pipeline task (#281)

echarlaix committed 2 years ago

Verified 4472ccb2

Update index.mdx (#235)

baskrahmer committed 2 years ago

Verified 97521dd7

Fix the number of layer to the decoder ones when creating the past key values outputs (#279)

echarlaix committed 2 years ago

Verified 2a3e888a

Correct bug in post-processing benchmark/eval (#278)

fxmarty committed 2 years ago

Verified 7ed8727e

Older