Unverified50% confidenceFactExact time
EgoOmniWorld contains 24,166 QA groups, including 20,284 training samples and 3,882 test samples.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEgoDex's held-out test set contains 3,243 episodes covering 111 types of tasks.61% similarUnverifiedThe author's fine-tuning used the QLoRA method, with LoRA rank/alpha of 32/64, a learning rate of 1e-4, batch size 1 with gradient accumulation of 16, AdamW 8-bit optimizer, and training for 1 epoch.60% similarUnverifiedReddit社区的大规模基准测试基于BeeLlama.cpp v0.4.0，在Qwen 3.6 27B和Gemma 4 31B两款模型上累计跑了413组配置对比，其中Qwen模型238组、Gemma模型175组58% similarUnverifiedIn the benchmark, the Dyna-Q algorithm obtained an average of approximately 0.62 badges.58% similarUnverifiedQwen3.6 35B模型包含256个专家，每个Token仅激活8个路由专家和1个共享专家，实际推理时激活参数约为3B57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1024008API
curl https://kongchang.com/api/v1/knowledge/claims/1024008MCP
get_claim(id=1024008)