Action Learning on ALFRED (Q2)
89.5AccuracyRAG + Direct
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| RAG + DirectAvg Latency=1017.83 ms, Avg Output Tokens=02026.05 | 89.5 | 85.7 | 18.6 | 30.57 | |
| CAG + DirectAvg Latency=1027.44 ms, Avg Output Tokens=02026.05 | 89.4 | 84 | 18.9 | 30.86 | |
| Affordance Routing (REIS)Avg Latency=406.21 ms, Avg Output Tokens=02026.05 | 89.4 | 85.2 | 18.5 | 30.4 | |
| Base + DirectAvg Latency=327.56 ms, Avg Output Tokens=02026.05 | 88.4 | 0 | 0 | — | |
| Base + ReasoningAvg Latency=43.59 sec, Avg Output Tokens=216.82026.05 | 88.4 | 0 | 0 | — |