ewqr2130/alignment-handbook-zephyr-7b_ppo_5e7step_102
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Jan 19, 2024License:apache-2.0Architecture:Transformer Open Weights Cold

The ewqr2130/alignment-handbook-zephyr-7b_ppo_5e7step_102 model is a 7 billion parameter language model developed by ewqr2130. This model is based on the Zephyr architecture and has been fine-tuned using Proximal Policy Optimization (PPO) for 5e7 steps. It features a 4096-token context length. Its primary characteristic is its alignment through PPO, making it suitable for tasks requiring controlled and aligned text generation.

Loading preview...