Unverified50% confidenceReleaseExact time
Huawei open-sourced the pre-training, SFT, and post-training RL code for Pangu OpenPangu 2.0, with accompanying Ascend-native training and inference technology.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/30/2026
First Seen
Valid until: 12/29/2026
Sources
Related Entities
Related Claims
UnverifiedOpenAI introduced a self-training mechanism starting from GPT-5.3, using code and data generated by the model itself to feed back into the model for training.72% similarVerifiedThe three-stage alignment training paradigm of SFT–reward model–RLHF was first systematically proposed by OpenAI in the InstructGPT paper.69% similarUnverifiedZGCM-1 publicly releases model weights, intermediate checkpoints, complete training code, data and data recipes, and W&B training logs for the pretraining, mid-training, and post-training stages.65% similarUnverifiedOpenAI continues to rely on NVIDIA for training while choosing to use its own in-house chips on the inference side.65% similarUnverifiedLoRA fine-tuning based on pre-trained models is the most realistic technical path for individual developers.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/960702API
curl https://kongchang.com/api/v1/knowledge/claims/960702MCP
get_claim(id=960702)