Zheng Yuan

icon indicating a website link https://ganjinzero.github.io

icon location Beijing icon location Ph.D. from THU Center of Statistical Science. Ex: Qwen. Focus on Alignment, Reasoning and Medical in NLP.

Results 84 comments of Zheng Yuan

训练时内存持续增长到OOM

同出现此问题

训练时内存持续增长到OOM

把**main.py**中的这一行改了一下 ```{python} config.gpu_options.per_process_gpu_memory_fraction = 0.9 ``` 就好了

Error using parse_pubmed_xml

Same issue here.

预训练模型会开放吗

暂时不会，开放时会通知

请问有数据集嘛

因为隐私规定，不能公开

请问有数据集嘛

> 请问语料库是来自CCKS的还是您自己的渠道？ 自己的

希望获得 Bert结果

由于训练数据的隐私性，暂时不能公开。如果可以公开了，会在github更新。

请问有相关的论文吗？

目前还没有，之后应该会有

How to use pretrained UmlsBERT to get embeddings for all the UMLS terms?

> Hello I want to work with UmlsBERT on the huggingface. there are two UmlsBert on the huggingface: > > https://huggingface.co/GanjinZero/UMLSBert_ENG , and cambridgeltl/SapBERT-UMLS-2020AB-all-lang-from-XLMR > > I am a beginner...

Output inconsistent for autoregressive performer

I guess the reason is from **k = create_kernel(k, is_query = False)** from FastAttention.forward, in the softmax_kernel operation, it has line **data_dash = ratio * (torch.exp(data_dash - diag_data - torch.amax(data_dash,...