Implement a Decoder KV Cache

Implement a Decoder KV Cache Implement a KV cache for a single-layer, single-batch Transformer decoder. The key and value of each token are integer vectors of length D. The cache is initially empty. Process Q operations: APPEND k1 ... kD v1 ... vD Append one token's key vector [k1, ..., kD] and value vector [v1, ..., vD] to the end of the KV cache. GET Print all cached keys and values. For every cached token, in insertion order, print one line in the form k1 ... kD | v1 ... vD. Print EMPTY if the cache is empty. Input Format D Q <operation 1> <operation 2> ... <operation Q> 1 <= D <= 64 1 <= Q <= 2 * 10^5 Integers in APPEND are in [-10^9, 10^9]. There is at least one GET operation. Output Format For each GET, print the complete KV cache at that point. Print an additional line containing --- between outputs of consecutive GET operations. Example Input: 2 4 APPEND 1 2 10 20 APPEND 3 4 30 40 GET GET Output: 1 2 | 10 20 3 4 | 30 40 --- 1 2 | 10 20 3 4 | 30 40 Implement efficient append and retrieval logic. Example Input 1 3 GET APPEND 5 9 GET Output EMPTY --- 5 | 9