Closed ayoub-louati closed 1 year ago
To test you can just run the convert_model command with the given parameters, for example (in local): convert_model -m roberta-base --backend onnx --seq-len 16 256 256 --task embedding --load-external-data (you can test other models, here it is only an example since the issue is based on roberta model)
Tested on cloud-dev with roberta-large. closes #138