SimonPu/
deepcoder:latest-128k

81 4 months ago

DeepCoder-14B-Preview, a code reasoning model finetuned from Deepseek-R1-Distilled-Qwen-14B via distributed RL

... /
params
51218d00ff5d · 50B
{
"num_ctx": 131072,
"temperature": 0.6,
"top_p": 0.95
}