{ "name": "HOLA 340M", "source_checkpoint": "/data/multi_query_deltanet/flash-linear-attention/legacy/training/exp/betae-rms-340m-canon", "architecture": "gated_deltanet", "parameters": "340M", "training_data": "SlimPajama-15B", "context_length": 2048, "cache_window": 64, "cache_chunk": 256, "cache_eviction": "betae", "cache_norm": "rms" }