huggingface/peft

Pull Requests Commits

Prepare 0.15.1 patch (#2459)

githubnemo committed 290 days ago

Verified c42eb227

Release 0.15.0 (#2435)

BenjaminBossan committed 298 days ago

Verified b34d8a2c

Fix #2422: Modules to save with multiple adapters (#2430)

githubnemo committed 298 days ago

Verified 48e0c5de

FIX Minimal target module optimization bug w/ IA³ (#2432)

BenjaminBossan committed 300 days ago

Verified b2b34fd6

FIX AutoPeftModels never reduce embedding size (#2427)

BenjaminBossan committed 303 days ago

Verified 7320bb94

ENH: Extend the regex for rank/alpha pattern (#2419)

BenjaminBossan committed 304 days ago

Verified 2f063e63

FIX Revert optimization for LoRA scaling == 1 (#2416)

BenjaminBossan committed 306 days ago

Verified 37266c1b

TST Add missing `.eval()` calls to inference tests (#2408)

faaany committed 310 days ago

Verified 8edaae94

FIX Reset the FP32 matmul precision in tests (#2411)

BenjaminBossan committed 310 days ago

Verified e1c7e8c8

TST Enable BNB tests on XPU (#2396)

faaany committed 311 days ago

Verified 24150d0e

Trainable Tokens: Support for Weight Tying (#2399)

githubnemo committed 311 days ago

Verified 461f6426

Use new slack secret token name (#2409)

githubnemo committed 312 days ago

Verified 0a695d55

ENH Add simple script to estimate train memory (#2378)

BenjaminBossan committed 313 days ago

Verified dbae7697

FIX Bug with PeftConfig.from_pretrained (#2397)

BenjaminBossan committed 313 days ago

Verified d5f5e35c

FIX Model with nested all-linear target modules (#2391)

BenjaminBossan committed 313 days ago

Verified 1dc14169

FIX GPTQModel Lora implementation (#2404)

Qubitium committed 313 days ago

Verified 8c8bf8f1

Standalone Custom Tokens Tuner and integrated into LoRA (#2376)

githubnemo committed 319 days ago

Verified f51203f3

ENH Make hotswap error on compile optional (#2393)

BenjaminBossan committed 321 days ago

Verified 3dd26682

FIX Failing single GPU tests related to hotswap (#2385)

BenjaminBossan committed 326 days ago

Verified bf186edc

SEC Bump transformers version used in examples (#2374)

BenjaminBossan committed 326 days ago

Verified c118a6e5

CI Skip audio test on single GPU CI (#2380)

BenjaminBossan committed 327 days ago

Verified e8babb10

FIX: Avoid caching in X-LoRA generate (#2384)

BenjaminBossan committed 327 days ago

Verified 1793a953

FIX Load checkpoint from custom cache dir (#2373)

henryzhengr committed 331 days ago

Verified 1e2d6b58

ENH Hotswap preparation raises when no adapter (#2375)

BenjaminBossan committed 332 days ago

Verified 94be64dd

FIX: Small fixes to hotswapping (#2366)

BenjaminBossan committed 333 days ago

Verified 6d033600

ENH DoRA optimization for ConvNd if dropout=0. (#2371)

gslama12 committed 334 days ago

Verified 363c14e6

DOC: Explain uninitialized weights warning (#2369)

BenjaminBossan committed 335 days ago

Verified 5e03d058

DOC Fix links to BOFT in docs (#2365)

makelinux committed 338 days ago

Verified 40fe1664

Hotswap allow different alpha scalings and ranks (#2177)

BenjaminBossan committed 340 days ago

Verified eaab05e1

ENH Allow disabling input dtype casting for LoRA (#2353)

BenjaminBossan committed 341 days ago

Verified db9dd3f4

Older