thamaraikannan / phi3_conti_pretrain

phi3 mini 4k instruct model and perform continual pretraining using data extracted from the "Reflextion LLM" research paper.

3B

3 Pulls Updated 2 days ago

2 days ago

cc24f200feed · 7.6GB

model
llama
·
3.82B
·
F16
template
{{- range .Messages }}<|{{ .Role }}|> {{ .Content }}</s> {{ end }}<|assistant|>
params
{"stop":["<|system|>","</s>","<|user|>","<|assistant|>"]}

Readme

No readme