-
Notifications
You must be signed in to change notification settings - Fork 2.3k
Pull requests: huggingface/peft
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
docs(prepare_model_for_kbit_training): note the ~0.5–1 GB CUDA reserved overhead (#3265)
#3267
opened May 26, 2026 by
Anai-Guo
Contributor
Loading…
2 tasks
Guard _convert_peft_config_moe against regex-string target_modules
#3266
opened May 26, 2026 by
Chessing234
Contributor
Loading…
FIX Transformers weight conversion undo incorrect key renaming
#3262
opened May 26, 2026 by
BenjaminBossan
Member
Loading…
Avoid TP helper imports for non-TP distributed LoRA loads
#3261
opened May 26, 2026 by
jiqing-feng
Contributor
Loading…
Return False from is_gptqmodel_available when gptqmodel isn't installed
#3255
opened May 23, 2026 by
c-tonneslan
Loading…
DOC Fix broken docstring examples in tuner model.py files
#3254
opened May 22, 2026 by
blipbyte
Contributor
Loading…
CI GPU run: Remove gptmodel from Dockerfile because of pcre
#3253
opened May 22, 2026 by
BenjaminBossan
Member
Loading…
Fail loud on save_pretrained() for unsharded LoRA tensors
#3251
opened May 21, 2026 by
akshansh47
Loading…
[BUG FIX] tuners/lora/eva: fix label mask guard using 'in' instead of hasattr for dict
#3247
opened May 20, 2026 by
Dev-X25874
Loading…
fix(tuners): use self.active_adapters (list) when guarding requires_grad for newly created sub-adapter modules
#3239
opened May 17, 2026 by
Chessing234
Contributor
Loading…
fix(tuners/lora/torchao): make get_apply_tensor_subclass optional (#3223)
#3234
opened May 13, 2026 by
Anai-Guo
Contributor
Loading…
3 tasks
Guard _convert_peft_config_moe against regex-string target_modules
#3232
opened May 12, 2026 by
1fanwang
Loading…
FIX Transformers weight conversion PEFT model prefix
#3230
opened May 12, 2026 by
BenjaminBossan
Member
Loading…
FIX BEFT merge on bias-less layers; RandLoRA save_projection default
#3227
opened May 12, 2026 by
kashif
Contributor
Loading…
fix(tuners): guard child.weight access in _replace_module for non-linear modules
#3213
opened May 4, 2026 by
Chessing234
Contributor
Loading…
fix(gralora): use _cast_input_dtype in forward instead of raw .to()
#3209
opened May 2, 2026 by
Chessing234
Contributor
Loading…
fix(pvera): use _cast_input_dtype in forward instead of raw .to()
#3208
opened May 2, 2026 by
Chessing234
Contributor
Loading…
ENH Optimize HRA forward for better memory efficiency
#3185
opened Apr 22, 2026 by
BenjaminBossan
Member
Loading…
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.