We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
比如在megatron中支持max-z loss, normhead等,还有支持baichuan2中的alibi + flash attention 2的结合。 另外,因为我之前也实现了megatron中的llama模型实现,其中huggingface的llama权重中没有model.layers.X.post_attention_layernorm.bias,并且post_attention_layernorm 应该是RMSNorm,而不是LayerNorm ~
The text was updated successfully, but these errors were encountered:
huggingface上的baichuan-2的实现两天前刚刚放出,我们正在接入中,预计下周可以push到github
Sorry, something went wrong.
百川2已经支持,烦请适用或者进群技术交流
No branches or pull requests
比如在megatron中支持max-z loss, normhead等,还有支持baichuan2中的alibi + flash attention 2的结合。
另外,因为我之前也实现了megatron中的llama模型实现,其中huggingface的llama权重中没有model.layers.X.post_attention_layernorm.bias,并且post_attention_layernorm 应该是RMSNorm,而不是LayerNorm ~
The text was updated successfully, but these errors were encountered: