Unverified50% confidenceFactExact time
EmsyAI was trained on 1.96 billion tokens, with a context window of 4096 tokens, a hidden dimension of 1024, an FFN dimension of 2816, and a vocabulary of approximately 16k
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a 170k token scenario, the attention score matrix requires approximately 54 GB of VRAM in BF16 precision (170,000 × 170,000 × 2 bytes).63% similarUnverifiedThe IMGNet model is only 10.58 MB (FP32) and was trained on the CASIA-WebFace dataset (490,000 images)62% similarUnverified开发者公布在RTX 3050 Laptop（4GB）上微调Llama-3.1-8B的实测结果为训练速度119.6 tok/s，峰值显存3.32 GB61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/912643API
curl https://kongchang.com/api/v1/knowledge/claims/912643MCP
get_claim(id=912643)