--- license: agpl-3.0 datasets: - kalomaze/Opus_Instruct_25k - MangoHQ/Claude-Data-Anon-Killed --- Experimental Finetune of Qwen 1.5B with Kalomaze Opus Instruct first and then Claude C2 logs. Model may produce NSFW results. First run on Opus Instruct -> https://raw.githubusercontent.com/MangyMango/AxolotlConfigs/main/step1two.yaml Second run on C2 logs -> https://raw.githubusercontent.com/MangyMango/AxolotlConfigs/main/step1two.yaml Model uses Alpaca formatting ``` ### Instruction: ### Input: ### Response: ``` [Built with Axolotl](https://github.com/axolotl-ai-cloud/axolotl)