Segmentation fault

[KawhiZhao]

`CUDA_VISIBLE_DEVICES=0 python3 run_seq2seq.py --data_dir ./dataset/ --src_file train.tsv --model_type unilm --model_name_or_path ./torch_unilm_model/ --output_dir ./output_dir/ --max_seq_length 512 --max_position_embeddings 512 --do_train --do_lower_case --train_batch_size 32 --learning_rate 1e-5 --num_train_epochs 3
09/26/2020 16:10:34 - INFO - main - device: cuda n_gpu: 1, distributed training: False, 16-bits training: False
09/26/2020 16:10:34 - INFO - transformers.configuration_utils - loading configuration file ./torch_unilm_model/config.json
09/26/2020 16:10:34 - INFO - transformers.configuration_utils - Model config UnilmConfig {
"_num_labels": 2,
"architectures": null,
"attention_probs_dropout_prob": 0.1,
"bos_token_id": null,
"directionality": "bidi",
"do_sample": false,
"early_stopping": false,
"eos_token_id": null,
"finetuning_task": null,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"id2label": {
"0": "LABEL_0",
"1": "LABEL_1"
},
"initializer_range": 0.02,
"intermediate_size": 3072,
"is_decoder": false,
"is_encoder_decoder": false,
"label2id": {
"LABEL_0": 0,
"LABEL_1": 1
},
"layer_norm_eps": 1e-12,
"length_penalty": 1.0,
"max_length": 20,
"max_position_embeddings": 512,
"min_length": 0,
"model_type": "",
"no_repeat_ngram_size": 0,
"num_attention_heads": 12,
"num_beams": 1,
"num_hidden_layers": 12,
"num_return_sequences": 1,
"output_attentions": false,
"output_hidden_states": false,
"output_past": true,
"pad_token_id": null,
"pooler_fc_size": 768,
"pooler_num_attention_heads": 12,
"pooler_num_fc_layers": 3,
"pooler_size_per_head": 128,
"pooler_type": "first_token_transform",
"pruned_heads": {},
"repetition_penalty": 1.0,
"temperature": 1.0,
"top_k": 50,
"top_p": 1.0,
"torchscript": false,
"type_vocab_size": 6,
"use_bfloat16": false,
"vocab_size": 21128
}

09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - Model name './torch_unilm_model/' not found in model shortcut name list (unilm-large-cased, unilm-base-cased). Assuming './torch_unilm_model/' is a path, a model identifier, or url to a directory containing tokenizer files.
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - Didn't find file ./torch_unilm_model/added_tokens.json. We won't load it.
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - Didn't find file ./torch_unilm_model/special_tokens_map.json. We won't load it.
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - Didn't find file ./torch_unilm_model/tokenizer_config.json. We won't load it.
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - loading file ./torch_unilm_model/vocab.txt
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - loading file None
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - loading file None
09/26/2020 16:10:34 - INFO - transformers.tokenization_utils - loading file None
Loading Train Dataset ./dataset/
Segmentation fault (core dumped)`

[PeterPanUnderhill]

Install sentencepiece version 0.1.91 fixes the problem.