Unverified50% confidenceFactExact time
Masked autoencoders (MAE) is a self-supervised pretraining framework proposed by He et al. in 2022.
1
Sources
50%
Confidence
Long-term
Relevance
10/21/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedToolformer was proposed by Meta in 2023, training models to autonomously decide when to call APIs.70% similarUnverified2021年的GLaM验证了64个专家的MoE架构在1.2万亿参数下的高效训练方法66% similarUnverifiedThe DSPy framework was released by Stanford in 2023, treating prompts as trainable parameters and automatically optimizing them through methods such as gradient descent.65% similarUnverifiedControlNet was proposed by Lvmin Zhang in 2023, injecting structural conditions into pre-trained diffusion models without modifying the original model weights.65% similarUnverifiedFill-in-the-Middle (FIM) technology was proposed by OpenAI in 2022, randomly masking middle segments of code during training to force the model to utilize both prefix and suffix simultaneously for prediction.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1079204API
curl https://kongchang.com/api/v1/knowledge/claims/1079204MCP
get_claim(id=1079204)