PyTorch
llama
alignment-handbook
Generated from Trainer
MambaInLlama_0_50 / configs.yaml
Junxiong Wang
fix configs
366d1cf
raw
history blame contribute delete
450 Bytes
MambaInLlama_0_50:
prompt_template: "zephyr-7b-alpha/prompt.txt"
fn_completions: "huggingface_local_completions"
completions_kwargs:
model_name: "JunxiongWang/MambaInLlama_0_50"
model_kwargs:
torch_dtype: 'bfloat16'
max_new_tokens: 2048
temperature: 0.7
top_p: 1.0
do_sample: True
pretty_name: "Mamba 0 5 From meta-llama/Meta-Llama-3-8B-Instruct"
link: "https://huggingface.co/JunxiongWang/MambaInLlama_0_50/"