weblab-GENIAC/Tanuki-8B-dpo-v1.0
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Aug 12, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

Tanuki-8B-dpo-v1.0 is an 8 billion parameter large language model developed by weblab-GENIAC, pretrained from scratch on approximately 1.3 trillion tokens. This model is instruction-tuned using Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) for conversational applications. It excels in Japanese language tasks, demonstrating strong performance across various categories including humanities, roleplay, and writing, as evaluated on Japanese MT-Bench.

Loading preview...