MiniMax H3 Training Progress: Turbo Time LoRA Fixes Motion Lag
ostrisai · x · 2026-08-16
The author shares current results from training the MiniMax H3 model using a 'turbo time' LoRA at 4 steps. The previous issue of lack of movement has been resolved; conversely, the results now show an overabundance of motion, indicating a significant shift in the model's dynamic capabilities.
More from Models
- llama.cpp integrates Dots3 Note model, scoring 78.4 on SWE-bench Verified — victormustar · 2026-08-16
- Post-training boosts GLM-5.3 to rival larger models, showing size isn't the bottleneck — haider1 · 2026-08-16
- Commits suggest Qwen 35B model removed, likely not releasing — Local-Cardiologist-5 · 2026-08-16
- Blind Test: Anime Girl 3D Scene Generation Across Models — Jeanodel · 2026-08-16
- 1-bit Quantized Qwen 27B Runs on 12GB VRAM at 92 tokens/s — zyxciss · 2026-08-16
- BlueBench Cyber Benchmark: GPT-5.6 Sol Leads, Kimi K3 Tops Open-Weights — vijaybolina · 2026-08-16