issues
search
artidoro
/
qlora
QLoRA: Efficient Finetuning of Quantized LLMs
https://arxiv.org/abs/2305.14314
MIT License
9.96k
stars
820
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
warnings.warn is not defined
#247
wangkuiyi
opened
1 year ago
0
Configure flake8, black, and pre-commit
#246
wangkuiyi
closed
11 months ago
0
how to attach LoRA modules to all layers
#245
ehartford
opened
1 year ago
1
Merge checkpoint adaptor weights with model
#244
sidracha
opened
1 year ago
0
question about quantization
#243
fxb392
opened
1 year ago
0
Multiple GPU inference
#242
Zheng392
opened
1 year ago
1
Custom dataset and special tokens
#241
BaiqingL
opened
1 year ago
0
multi gpu uneven VRAM utilization
#240
ehartford
opened
1 year ago
1
Use for ggml
#239
RonanKMcGovern
opened
1 year ago
0
how to merge
#238
ehartford
closed
1 year ago
0
Redundant `if-else`
#237
andylolu2
opened
1 year ago
0
use torch_dtype=fp16 when specified
#236
tmm1
opened
1 year ago
1
add flash attention
#235
ehartford
opened
1 year ago
12
NFC: Update qlora.py with some small cleanup
#234
njbrake
opened
1 year ago
0
print info
#233
PrisonBreak2017
closed
1 year ago
2
Error in the guanaco-7B-demo-colab.ipynb colab
#232
rain-1
opened
1 year ago
1
Update qlora.py
#231
rain-1
opened
1 year ago
1
add ZeRO to QloRA
#230
hzg0601
closed
1 year ago
0
Does finetuning need to follow the Llama 2 system prompt format?
#229
Zheng392
opened
1 year ago
1
Whenever I use QLoRA to train LLama/LLama 2 on an instruction-tuning dataset like Dolly or Alpaca I get a periodically oscillating training loss
#228
ritabratamaiti
opened
1 year ago
12
Parameters source_max_len & target_max_len
#227
usmanxia
closed
1 year ago
5
Finetuning the Llama-2-70b-chat-hf
#226
SoumitriKolavennu
opened
1 year ago
3
13B On 24GB go OOM
#225
nivibilla
closed
1 year ago
16
Notebook Code for qlora doesn't work
#224
nivibilla
closed
1 year ago
3
Expected to mark variable ready only once
#223
nivibilla
closed
1 year ago
2
Set `ddp_find_unused_parameters` to False when using distributed training
#222
aresnow1
opened
1 year ago
4
FlashAttention support?
#221
BugReporterZ
opened
1 year ago
14
Example for LLaMA2 Finetuning and Version Update
#220
artidoro
closed
1 year ago
0
Enable Qlora scripts on Intel GPUs
#219
abhilash1910
closed
1 year ago
4
Introduces load_from_disk datasets
#218
nickmitchko
opened
1 year ago
4
Updated qlora.py to fix freezing of embedding layers
#217
ffohturk
closed
1 year ago
0
Add LLaMA 2 Support
#216
artidoro
closed
1 year ago
8
Freezing embeddings and adding known issue message
#215
artidoro
closed
1 year ago
0
Adding new tokens causes performance and memory issues
#214
artidoro
opened
1 year ago
4
May be 32-bit adminw and 8-bit adminw can using together ?
#213
apachemycat
opened
1 year ago
0
Legacy tokenizer
#212
atillabasaran
opened
1 year ago
0
Is there a wandb report for the guanaco model?
#211
hafidh561
opened
1 year ago
0
LoRA for embeddings layer
#210
Andrei-Aksionov
closed
1 year ago
2
why dtype change line needs?
#209
YooSungHyun
opened
1 year ago
0
PeftModel.from_pretrained() takes 3 minutes?
#208
gptzerozero
opened
1 year ago
1
Pushing test file
#207
rajiv8544
closed
1 year ago
0
eval_loss or mmlu_loss
#206
Duncan1115
opened
1 year ago
1
How to pretrain "raw" text?
#205
SinanAkkoyun
opened
1 year ago
4
Strage error while launching the code
#204
zepmck
opened
1 year ago
2
Not able to run 8K context length even with multiple GPUs.
#203
KOVVURISATYANARAYANAREDDY
opened
1 year ago
1
Continuing training using different dataset
#202
0xymoro
closed
1 year ago
1
Why torch.float32 if args.fp16 ?
#201
guihonghao
opened
1 year ago
2
OSError: [WinError 145]
#200
ghost
opened
1 year ago
0
draft: Feat/new evals
#199
LucianoPierdona
closed
1 year ago
0
Training on 2x40GB A100s with FSDP: ValueError
#198
ffohturk
opened
1 year ago
5
Previous
Next