Is this model trained with LORA based on Qwen3-32B, or is it fully trained? Are there any hyperparameter settings for the training? Additionally, can Helpsteer3 + Qwen3-8B be used for training? Or is there a quantized version of the 32B model available?
