Xun Chen

icon company Shanghai Jiao Tong University icon location China icon location Physicist

Results 1 issues of Xun Chen

cannot specify the output sequence length when using pipelines for GPT3 inference.

1 comment

I tried to load the 'damo/nlp_gpt3_text-generation_1.3B' model with pipeline for inference. The output texts have similar lengthes and cannot be changed with the `max_length` parameter passed in. ``` In [1]:...