QingruZhang/AdaLoRA
AdaLoRA: Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning (ICLR 2023).
PythonMIT
Issues
- 5
- 0
Works well
#26 opened - 0
- 2
AssertionError
#20 opened by luoxindi - 0
where is the "training cost as C(P, E, Q)" in your total loss L(P, E, Q) = C(P, E, Q) + γ Pn k=1 R(Pk, Qk),
#24 opened by LiZhangMing - 1
Does 'apply_rankselector appear in the code?
#15 opened by dddfaker - 1
Lora Dreambooth?
#8 opened by Dentoty - 1
How to implement prune LoRA?
#6 opened by A11en0 - 1
The question about convergence speed
#19 opened by Lanbai-eleven - 1
Question about orthogonal regularization methods
#22 opened by fei407 - 2
Random seeds used in your experiments
#23 opened by Car-pe - 1
Questions about ranknum
#12 opened by luchaoqi - 2
- 1
When I use multi-GPU training on a single machine, the following error is reported:ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 470401) of binary: /home/sqh/miniconda3/envs/NLU/bin/python
#17 opened by Sun-9923 - 0
AssertionError: DistributedDataParallel is not needed when a module doesn't have any parameter that requires a gradient.
#18 opened by DigitalLifeYZQiu - 0
Question about the random seeds
#16 opened by SEONHOK - 10
TypeError: 'NoneType' object is not iterable
#13 opened by lcqlalala - 0
Fusion lora weights to original weights
#14 opened by vvhj - 2
How do you find the optimal parameters?
#9 opened by A11en0 - 0
How about the variance in GLUE bencharmark?
#10 opened by A11en0 - 2
question about early stopping
#3 opened by DopamineLcy - 3
- 1
- 6
- 2
from_pretrained parameters
#2 opened by fxb392 - 1