DeepSeek's first generation reasoning models with comparable performance to OpenAI-o1.
7b
8b
14b
32b
70b
441.9K Pulls Updated 2 months ago
f4d24e9138dd · 148B
{
"stop": [
"<|begin▁of▁sentence|>",
"<|end▁of▁sentence|>",
"<|User|>",
"<|Assistant|>"
]
}