ylsung / vl-merging Goto Github PK
View Code? Open in Web Editor NEWPyTorch codes for the paper "An Empirical Study of Multimodal Model Merging"
PyTorch codes for the paper "An Empirical Study of Multimodal Model Merging"
Hi YiLin,
When I tried to run fine-tuning code on VQA(4 gpus):
python run.py with data_root=${data_dir} num_gpus=4 num_nodes=1 task_finetune_vqa_square_randaug_base_image384_ufo \
exp_name=ma_vqa_finetuning per_gpu_batchsize=4 batch_size=16 image_size=480 learning_rate=3e-5 \
load_path=${load_path} log_dir=${log_dir} drop_rate=0.15 max_epoch=10 ufo \
I got the Error:
ERROR - VLMo - Failed after 0:00:20!
-- Process 0 terminated with the following error:
Traceback (most recent call last):
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/torch/multiprocessing/spawn.py", line 59, in _wrap
fn(i, *args)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/pytorch_lightning/strategies/launchers/multiprocessing.py", line 138, in _wrapping_function
results = function(*args, **kwargs)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 621, in _fit_impl
self._run(model, ckpt_path=self.ckpt_path)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 994, in _run
self.strategy.setup_environment()
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/pytorch_lightning/strategies/ddp_spawn.py", line 138, in setup_environment
self.setup_distributed()
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/pytorch_lightning/strategies/ddp_spawn.py", line 174, in setup_distributed
_init_dist_connection(
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/lightning_lite/utilities/distributed.py", line 237, in _init_dist_connection
torch.distributed.init_process_group(torch_distributed_backend, rank=global_rank, world_size=world_size, **kwargs)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/torch/distributed/distributed_c10d.py", line 576, in init_process_group
store, rank, world_size = next(rendezvous_iterator)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/torch/distributed/rendezvous.py", line 229, in _env_rendezvous_handler
store = _create_c10d_store(master_addr, master_port, rank, world_size, timeout)
File "/cpfs/29cd2992fe666f2a/user/wangzekun/miniconda3/envs/vlmerge/lib/python3.8/site-packages/torch/distributed/rendezvous.py", line 157, in _create_c10d_store
return TCPStore(
RuntimeError: Address already in use
I'm sure there is no process using the port. So what's the problem with my config(?) or the code?
Hi, model merging is a new term for me. Model merging is a technique, which can achieve multi-task objective while do not need to conduct multi-task training. Is that correct?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.