🖼️00Pushing the Limits of KV Cache Compression · zartbotHacker News·Pushing the Limits of KV Cache Compression · zartbot·about 3 hours ago#9lNFhsUJ#zartbot#deepseek#kvcache#flash#prefill#article+2 more🧰Tag tools✨Add tagA deep dive into the DeepSeek-V4.1 Flash technical report: CED, CSA2, HSI, Single-Pass mHC, Engram and FP4 KV Cache — how the KV cache was compressed to just 890 bytes per token.15s0Read later0Read More
