commit
ab1f57a394
1 changed files with 2 additions and 0 deletions
@ -0,0 +1,2 @@ |
|||||
|
<br>DeepSeek open-sourced DeepSeek-R1, an LLM fine-tuned with [support learning](http://wiki.lexserve.co.ke) (RL) to enhance reasoning capability. DeepSeek-R1 attains results on par with [OpenAI's](https://www.arztsucheonline.de) o1 design on several criteria, [including](https://www.personal-social.com) MATH-500 and [SWE-bench](https://git.spitkov.hu).<br> |
||||
|
<br>DeepSeek-R1 is based on DeepSeek-V3, a mixture of professionals (MoE) design just recently open-sourced by [DeepSeek](https://git.chartsoft.cn). This base design is fine-tuned using Group Relative Policy Optimization (GRPO), [bio.rogstecnologia.com.br](https://bio.rogstecnologia.com.br/bagjanine969) a reasoning-oriented variant of RL. The research group also carried out understanding distillation from DeepSeek-R1 to [open-source Qwen](http://www.grainfather.co.nz) and [hb9lc.org](https://www.hb9lc.org/wiki/index.php/User:Margareta19E) Llama designs and released several [versions](http://47.92.109.2308080) of each |
Loading…
Reference in new issue