karpathy / llm.c

LLM training in simple, raw C/CUDA
MIT License
23.31k stars 2.59k forks source link

TypeError: normal_() got an unexpected keyword argument 'generator' #723

Open StarHtimE opened 1 month ago

StarHtimE commented 1 month ago

Traceback (most recent call last): File "/root/llm.c/train_gpt2.py", line 663, in model = GPT.from_pretrained(args.model) File "/root/llm.c/train_gpt2.py", line 210, in from_pretrained model = GPT(config) File "/root/llm.c/train_gpt2.py", line 147, in init self.apply(self._init_weights) File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 897, in apply module.apply(fn) File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 897, in apply module.apply(fn) File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 898, in apply fn(self) File "/root/llm.c/train_gpt2.py", line 160, in _initweights torch.nn.init.normal(module.weight, mean=0.0, std=0.02, generator=self.initrng) TypeError: normal() got an unexpected keyword argument 'generator' image

StarHtimE commented 1 month ago

How to solve this issue?