Model config is logged twice on startup

Currently, the model config is logged twice during startup: 1. via `AutoConfig.from_pretrained` 2. via `AutoTokenizer.from_pretrained` -> `AutoConfig.from_pretrained` Should there be a state variable that prevents the logging of the same config twice? This happens for example with all example scripts: Example log when running `run_clm.py`: ``` File "examples/language-modeling/run_clm.py", line 444, in <module> main() File "examples/language-modeling/run_clm.py", line 275, in main config = AutoConfig.from_pretrained(model_args.model_name_or_path, **config_kwargs) File "/mnt/nvme1/code/huggingface/transformers-gpt-neo-nan/src/transformers/models/auto/configuration_auto.py", line 401, in from_pretrained return config_class.from_dict(config_dict, **kwargs) File "/mnt/nvme1/code/huggingface/transformers-gpt-neo-nan/src/transformers/configuration_utils.py", line 526, in from_dict traceback.print_stack() [INFO|configuration_utils.py:527] 2021-04-06 21:16:04,999 >> Model config GPT2Config { "_num_labels": 1, "activation_function": "gelu_new", "architectures": [ "GPT2LMHeadModel" ], "attn_pdrop": 0.1, "bos_token_id": 50256, "embd_pdrop": 0.1, "eos_token_id": 50256, "gradient_checkpointing": false, "id2label": { "0": "LABEL_0" }, "initializer_range": 0.02, "label2id": { "LABEL_0": 0 }, "layer_norm_epsilon": 1e-05, "model_type": "gpt2", "n_ctx": 1024, "n_embd": 768, "n_head": 12, "n_inner": null, "n_layer": 6, "n_positions": 1024, "resid_pdrop": 0.1, "summary_activation": null, "summary_first_dropout": 0.1, "summary_proj_to_labels": true, "summary_type": "cls_index", "summary_use_proj": true, "task_specific_params": { "text-generation": { "do_sample": true, "max_length": 50 } }, "transformers_version": "4.5.0", "use_cache": true, "vocab_size": 50257 } [INFO|configuration_utils.py:490] 2021-04-06 21:16:05,277 >> loading configuration file https://huggingface…