mmarch/text-generation-webui
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

fix lora training with alpaca_lora_4bit (#3853)

Browse source
This commit is contained in:
John Smith 2023-09-11 12:22:20 +08:00 • committed by GitHub
parent 15e9b8c915
commit cc7b7ba153
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 22 additions and 30 deletions
Download patch file Download diff file Expand all files Collapse all files

13
modules/training.py
View file

@ -269,10 +269,10 @@ def calc_trainable_parameters(model):
def do_train(lora_name: str, always_override: bool, save_steps: int, micro_batch_size: int, batch_size: int, epochs: int, learning_rate: str, lr_scheduler_type: str, lora_rank: int, lora_alpha: int, lora_dropout: float, cutoff_len: int, dataset: str, eval_dataset: str, format: str, eval_steps: int, raw_text_file: str, overlap_len: int, newline_favor_len: int, higher_rank_limit: bool, warmup_steps: int, optimizer: str, hard_cut_string: str, train_only_after: str, stop_at_loss: float, add_eos_token: bool, min_chars: int, report_to: str):
if shared.args.monkey_patch:
from monkeypatch.peft_tuners_lora_monkey_patch import (
replace_peft_model_with_gptq_lora_model
from alpaca_lora_4bit.monkeypatch.peft_tuners_lora_monkey_patch import (
replace_peft_model_with_int4_lora_model
)
replace_peft_model_with_gptq_lora_model()
replace_peft_model_with_int4_lora_model()
global WANT_INTERRUPT
WANT_INTERRUPT = False
@ -512,11 +512,12 @@ def do_train(lora_name: str, always_override: bool, save_steps: int, micro_batch
return
if shared.args.monkey_patch:
for n, m in lora_model.named_modules():
if '4bit' in str(type(m)):
from alpaca_lora_4bit.autograd_4bit import Autograd4bitQuantLinear
from alpaca_lora_4bit.models import Linear4bitLt
for _, m in lora_model.named_modules():
if isinstance(m, Autograd4bitQuantLinear) or isinstance(m, Linear4bitLt):
if m.is_v1_model:
m.zeros = m.zeros.half()
m.scales = m.scales.half()
class Tracked():