Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

stuck when inferring #63

Open
raihan0824 opened this issue Mar 14, 2023 · 1 comment
Open

stuck when inferring #63

raihan0824 opened this issue Mar 14, 2023 · 1 comment

Comments

@raihan0824
Copy link

Screen Shot 2023-03-14 at 12 31 46

I run this script
deepspeed --num_gpus 1 bloom-inference-scripts/bloom-ds-inference.py --name bigscience/bloomz-7b1 --batch_size 8
and it gets stuck just like in the picture.

Log:

(base) raihanafiandi@instance-1:~/playground/transformers-bloom-inference$ deepspeed --num_gpus 1 bloom-inference-scripts/bloom-ds-inference.py --name bigscience/bloomz-7b1 --batch_size 8
[2023-03-14 05:30:02,152] [WARNING] [runner.py:186:fetch_hostfile] Unable to find hostfile, will proceed with training with local resources only.
[2023-03-14 05:30:02,965] [INFO] [runner.py:550:main] cmd = /opt/conda/bin/python3.7 -u -m deepspeed.launcher.launch --world_info=eyJsb2NhbGhvc3QiOiBbMF19 --master_addr=127.0.0.1 --master_port=29500 --enable_each_rank_log=None bloom-inference-scripts/bloom-ds-inference.py --name bigscience/bloomz-7b1 --batch_size 8
[2023-03-14 05:30:04,255] [INFO] [launch.py:142:main] WORLD INFO DICT: {'localhost': [0]}
[2023-03-14 05:30:04,255] [INFO] [launch.py:149:main] nnodes=1, num_local_procs=1, node_rank=0
[2023-03-14 05:30:04,255] [INFO] [launch.py:161:main] global_rank_mapping=defaultdict(<class 'list'>, {'localhost': [0]})
[2023-03-14 05:30:04,255] [INFO] [launch.py:162:main] dist_world_size=1
[2023-03-14 05:30:04,255] [INFO] [launch.py:164:main] Setting CUDA_VISIBLE_DEVICES=0
[2023-03-14 05:30:05,816] [INFO] [comm.py:663:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
*** Loading the model bigscience/bloomz-7b1
Fetching 8 files: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8/8 [00:00<00:00, 104857.60it/s]
Fetching 8 files: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8/8 [00:00<00:00, 113359.57it/s]
Fetching 8 files: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8/8 [00:00<00:00, 99864.38it/s]
Fetching 8 files: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 8/8 [00:00<00:00, 32832.13it/s]
[2023-03-14 05:30:13,610] [INFO] [logging.py:77:log_dist] [Rank 0] DeepSpeed info: version=0.8.2, git-hash=unknown, git-branch=unknown
[2023-03-14 05:30:13,611] [WARNING] [config_utils.py:77:_process_deprecated_field] Config parameter mp_size is deprecated use tensor_parallel.tp_size instead
[2023-03-14 05:30:13,611] [INFO] [logging.py:77:log_dist] [Rank 0] quantize_bits = 8 mlp_extra_grouping = False, quantize_groups = 1
Installed CUDA version 11.0 does not match the version torch was compiled with 11.1 but since the APIs are compatible, accepting this combination
Using /home/raihanafiandi/.cache/torch_extensions/py37_cu111 as PyTorch extensions root...

Is it because I run other bloom models? (bloomz-7b-1)? Please help me on this. Thank you

@mayank31398
Copy link
Collaborator

I don't think thats the case. I will try to run this on my end :0

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants