About the 3090/4090/A10 device support

#15
by thincal - opened

What is the release plan for the other device support ? such as 3090/4090/A10, actually A100 is expensive in the inference scenario.

同问V100的开源计划,感谢

Tencent Music Entertainment Lyra Lab org

@rayleee A10 & A100 are both Ampere architecture so it can run on A10. However A10 has a much smaller GPU memory so you may encounter OOM on A10.

And we have noticed this problem, new models are on their way! (including V100)

Tencent Music Entertainment Lyra Lab org

@thincal @rayleee We have updated to a new accelerated version and removed the previous TensorRT acceleration version. The new version has undergone significant optimization at the source code level, resulting in improved performance, ease of use, and GPU compatibility. Please update and feel free to try it out.

@bigmoyan @vanewu , Awesome! Also, will the script for model convert from chatglm-6b HF be release later?

@bigmoyan @vanewu , Awesome! Also, will the script for model convert from chatglm-6b HF be release later?
同关注

@rayleee A10 & A100 are both Ampere architecture so it can run on A10. However A10 has a much smaller GPU memory so you may encounter OOM on A10.

And we have noticed this problem, new models are on their way! (including V100)

How about the 4090 and A5000 support ?

Sign up or log in to comment