Unverified50% confidenceFactExact time
The open-source fine-tuning project "Bencao" in the traditional Chinese medicine field was fine-tuned by Harbin Institute of Technology, using Llama 7B, Chinese Llama Alpaca, and ChatGLM6B as bases, all with public data, trained for 10 epochs on A100G taking 2 hours, with 4.6K GitHub stars.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/5/2026
First Seen
Valid until: 2/3/2027
Sources
A Beginner's Guide to LLM Fine-Tuning: From Principles to Industry Implementation
bilibili百里谈技术11/5/2026
Related Entities
Related Claims
UnverifiedA systematic empirical study used nearly 200B tokens for pretraining on decoder-only LLMs ranging from 360M to 3.1B parameters, examining fourteen orthogonal training variants.65% similarUnverified智谱GLM-5.3基于743B（7430亿参数）基座进行后训练，主打编程与智能体两大核心场景65% similarUnverifiedGLM-5.2 团队大幅扩展了面向编码智能体场景的 1M 上下文训练，覆盖大规模实现、自动化研究、性能优化和复杂调试等任务类型63% similarUnverifiedPerforming supervised fine-tuning (SFT) or preference alignment training (RLHF/DPO) on existing foundation models (such as the Llama or Mistral series) requires only hundreds to thousands of high-quality task samples and can be completed with limited GPU resources.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1140087API
curl https://kongchang.com/api/v1/knowledge/claims/1140087MCP
get_claim(id=1140087)