XTuner Release V0.1.12

What's Changed

• set dev version by @LZHgrla in #281
• [Fix] Update LLaVA results by @LZHgrla in #283
• [Fix] Update LLaVA results (based on VLMEvalKit) by @LZHgrla in #285
• [Fix] Fix filter bug for test data by @LZHgrla in #293
• [Fix] Fix ConcatDataset by @LZHgrla in #298
• [Improve] Redesign the prompt_template by @LZHgrla in #294
• [Fix] Fix errors about stop_words by @LZHgrla in #313
• [Fix] Fix Mixtral LoRA setting by @LZHgrla in #312
• [Feature] Support DeepSeek-MoE by @LZHgrla in #311
• [Fix] Set torch.optim.AdamW as the default optimizer by @LZHgrla in #318
• [FIx] Fix pth_to_hf for LLaVA model by @LZHgrla in #316
• [Improve] Add demo_data examples by @LZHgrla in #278
• [Feature] Support InternLM2 by @LZHgrla in #321
• [Fix] Fix the resume of seed by @LZHgrla in #309
• [Feature] Accelerate xtuner xxx by @pppppM in #307
• [Fix] Fix InternLM2 url by @LZHgrla in #325
• [Fix] Limit the version of python, >=3.8, <3.11 by @LZHgrla in #327
• [Fix] Add trust_remote_code=True for AutoModel by @LZHgrla in #328
• [Docs] Improve README by @LZHgrla in #326
• bump verion to v0.1.12 by @pppppM in #323

Full Changelog: v0.1.11...v0.1.12