huggingface/peft

Pull Requests Commits

add docs chatbot v0

pacman100 committed 2 years ago

f31b3544

fix crash when using torch.nn.DataParallel for LORA inference (#805)

sywangyi committed 2 years ago

Verified 7d44026d

Update docstring of PeftModel.from_pretrained (#799)

BenjaminBossan committed 2 years ago

Verified ba90047d

add lora default target module for codegen (#787)

sywangyi committed 2 years ago

Verified 10cf3a4f

Add adapter error handling (#800)

BenjaminBossan committed 2 years ago

Verified aac7722b

[`core`] PEFT refactor + introducing `inject_adapter_in_model` public method (#749)

younesbelkada committed 2 years ago

Verified ed396a69

Allow passing inputs_embeds instead of input_ids (#757)

BenjaminBossan committed 2 years ago

Verified ec267c64

Support NPU adapter loading (#772)

ji-huazhong committed 2 years ago

Verified 9b580893

Support XPU adapter loading (#737)

abhilash1910 committed 2 years ago

Verified b10a8ced

Add progressbar unload/merge (#753)

BramVanroy committed 2 years ago

Verified bfb264ad

Add tests for AdaLoRA, fix a few bugs (#734)

BenjaminBossan committed 2 years ago

Verified 702f9377

DOC: Examples for LoRA with custom models (#724)

BenjaminBossan committed 2 years ago

Verified 0e33ac1e

[`ModulesToSave`] add correct hook management for modules to save (#755)

younesbelkada committed 2 years ago

Verified e27e8834

Add btlm to officially supported LoRA (#751)

Trapper4888 committed 2 years ago

Verified ffbb6bcf

fix adalora inference issue (#745)

younesbelkada committed 2 years ago

Verified 8541b60a

Updated Example in Class:LoraModel (#672)

TianyiPeng committed 2 years ago

Verified 96c0277a

FIX: Disabling adapter works with modules_to_save (#736)

BenjaminBossan committed 2 years ago

Verified b15c1859

ENH: Warn when disabling adapters and bias != 'none' (#741)

BenjaminBossan committed 2 years ago

Verified a955ef10

Fixes warning when initializing prompt encoder (#716)

BenjaminBossan committed 2 years ago

Verified e06d94dd

[`Patch`] patch trainable params for 4bit layers (#733)

younesbelkada committed 2 years ago

Verified 1681cebf

[`Llama2`] Add disabling TP behavior (#728)

younesbelkada committed 2 years ago

Verified a09f66c8

FIX: add type information to package_data (#729)

aarnphm committed 2 years ago

Verified 1869fe6e

revert change (#731)

pacman100 committed 2 years ago

Verified 1c27e24d

fix the param count when using 4-bit bnb

pacman100 committed 2 years ago

30fd5a4c

Add falcon to officially supported LoRA & IA3 modules (#722)

younesbelkada committed 2 years ago

Verified 3040782e

Fix subfolder issue (#721)

younesbelkada committed 2 years ago

Verified 1b8b17de

Release version 0.5.0.dev0 (#717)

pacman100 committed 2 years ago

Verified 029f416f

FIX: Removes warnings about unknown pytest marker (#715)

BenjaminBossan committed 2 years ago

Verified a1953bae

better hub kwargs management (#712)

younesbelkada committed 2 years ago

Verified e90dcc4b

FEAT: Make LoRA work with custom models (#676)

BenjaminBossan committed 2 years ago

Verified 71b326db

Older