Can only be used for continuous pre-training, and please follow the original context template.

An experimental llama architecture alignment model that attempts to merge yi 9b 200k and qwen3 8b into the llama architecture.

Downloads last month
2
Safetensors
Model size
17B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support