Issues
- 0
- 17
when I use phi-2 model,output is “the the”
#11 opened by yuxiaoranyu - 2
请问支持中文数据微调/推理吗?
#31 opened by LianghuiGuo - 2
FileNotFoundError: [Errno 2] No such file or directory: 'prompts/alignment.txt'
#27 opened by yanghu819 - 0
llama_model: "/root/autodl-tmp/phi-new" which vision of phi is used in the stage4 training process?
#33 opened by yzc-ippl - 0
- 1
- 1
- 8
Phi-2 problem
#26 opened by dexmac221 - 0
- 1
- 0
Could not create share link.
#25 opened by chygoa - 2
- 1
Size mismatch for stage1 checkpoints
#22 opened by wangskyone - 8
- 2
On the stage 4,the Q-Former is trained, however the Q-Former should not been trained in the paper
#23 opened by TangYuan96 - 2
Error while processing
#21 opened by DaBaiTuu - 2
- 7
Attempting to unscale FP16 gradients
#17 opened by sunzhe09 - 0
能在autodl社区分享一下你们的镜像吗
#19 opened by xiaoxue-roy - 1
REC Results
#16 opened by tydia - 1
performance evaluation details
#15 opened by dylanqyuan - 1
Typo in paper
#14 opened by hu-po - 2
What model should be used here?
#7 opened by VacantHusky - 5
download from huggingface failed
#12 opened by DaBaiTuu - 1
The first stage training, lora is included.
#13 opened by zr-icu - 2
BLIP-2 / Q-Former / Benchmarks
#8 opened by Jotschi - 5
- 0
Thank you!
#6 opened by miolini - 2
'PhiAttention' object has no attribute 'q_layernorm'/ 'k_layernorm'/ 'k_postnorm'
#5 opened by zr-icu - 2
Unable to run on a remote jupyter environment
#4 opened by twilwa - 1
How long does it take to train?
#3 opened by zxti - 0
Nice Work!
#1 opened by xiaoachen98