macanv / BERT-BiLSTM-CRF-NER

Tensorflow solution of NER task Using BiLSTM-CRF model with Google BERT Fine-tuning And private Server services
https://github.com/macanv/BERT-BiLSMT-CRF-NER
4.67k stars 1.25k forks source link

显存几乎沾满,利用率却很低 #390

Open zhishui3 opened 3 years ago

zhishui3 commented 3 years ago

不管数据量有多大,显存本身24000MB 多 占用了23000MB多(并且这个数量不变),请问这是怎么回事啊? 用了各种方式都不管用:1、tf.data.TFRecordDataset.cache(),2、tf.data.TFRecordDataset.shard,3、把tf_record 分成多份读取,4、epoch 和batch_size 分别改成 1和16, 最后只能用 session_config.gpu_options.per_process_gpu_memory_fraction @macanv 求大佬指点!

sharipp132 commented 1 year ago

俺也一样

Anthony-BGST commented 11 months ago

请问你验证用了多久?