You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Baichuan2-7B-Base微调报错 AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'
#395
Open
qingchen177 opened this issue
Apr 3, 2024
· 1 comment
报错:AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'
改用Baichuan2-13B-Base直接报内存溢出:
torch.cudatorch.cuda..OutOfMemoryErrorOutOfMemoryError: : CUDA out of memory. Tried to allocate 1.92 GiB (GPU 0; 23.65 GiB total capacity; 15.90 GiB already allocated; 1.04 GiB free; 21.20 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFCUDA out of memory. Tried to allocate 1.92 GiB (GPU 1; 23.65 GiB total capacity; 15.90 GiB already allocated; 1.73 GiB free; 21.39 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
output = input.matmul(weight.t())
三张4090跑的,求解
The text was updated successfully, but these errors were encountered:
命令:
deepspeed --hostfile=$hostfile fine-tune.py
--report_to "none"
--data_path "data/test.json"
--model_name_or_path "/data/models/Baichuan2-7B-Base"
--output_dir "output"
--model_max_length 512
--num_train_epochs 4
--per_device_train_batch_size 16
--gradient_accumulation_steps 1
--save_strategy epoch
--learning_rate 2e-5
--lr_scheduler_type constant
--adam_beta1 0.9
--adam_beta2 0.98
--adam_epsilon 1e-8
--max_grad_norm 1.0
--weight_decay 1e-4
--warmup_ratio 0.0
--logging_steps 1
--gradient_checkpointing True
--deepspeed ds_config.json
--bf16 True
--tf32 True
--use_lora True
报错:AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'AttributeError: 'BaichuanConfig' object has no attribute 'z_loss_weight'
改用Baichuan2-13B-Base直接报内存溢出:
torch.cudatorch.cuda..OutOfMemoryErrorOutOfMemoryError: : CUDA out of memory. Tried to allocate 1.92 GiB (GPU 0; 23.65 GiB total capacity; 15.90 GiB already allocated; 1.04 GiB free; 21.20 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFCUDA out of memory. Tried to allocate 1.92 GiB (GPU 1; 23.65 GiB total capacity; 15.90 GiB already allocated; 1.73 GiB free; 21.39 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
三张4090跑的,求解
The text was updated successfully, but these errors were encountered: