| <meta charset=utf-8><title>XOR continual-learning results</title><style>body{font:14px system-ui;margin:2rem;color:#172033}table{border-collapse:collapse}th,td{border:1px solid #cbd5e1;padding:.45rem;text-align:right}th{background:#e0f2fe}</style><h1>Independent CUDA XOR-cluster reproduction</h1><p>Means over random seeds; task 1 is evaluated after each later task.</p><table><thead><tr><th>sweep</th><th>n_later</th><th>width</th><th>after_task</th><th>activation</th><th>loss</th><th>seeds</th><th>mean_train_forgetting</th><th>mean_test_forgetting</th><th>mean_delayed_gap</th><th>mean_task1_test_error</th><th>mean_later_cumulative_loss</th></tr></thead><tbody><tr><td>sample</td><td>800</td><td>1000</td><td>2</td><td>quadratic</td><td>hinge</td><td>5</td><td>6.102621555328369e-05</td><td>0.00019200146198272705</td><td>0.0067827969789505</td><td>0.0</td><td>141.82634708285332</td></tr><tr><td>sample</td><td>800</td><td>1000</td><td>3</td><td>quadratic</td><td>hinge</td><td>5</td><td>-0.00014689266681671141</td><td>-0.00011706948280334472</td><td>0.0066816449165344235</td><td>0.0</td><td>282.31662721037867</td></tr><tr><td>sample</td><td>2500</td><td>1000</td><td>2</td><td>quadratic</td><td>hinge</td><td>5</td><td>-0.00015653669834136963</td><td>1.0517239570617675e-05</td><td>0.006818875670433044</td><td>0.0</td><td>142.0373024582863</td></tr><tr><td>sample</td><td>2500</td><td>1000</td><td>3</td><td>quadratic</td><td>hinge</td><td>5</td><td>-6.243288516998291e-05</td><td>-3.930926322937012e-06</td><td>0.006710323691368103</td><td>0.0</td><td>282.3474177002907</td></tr></tbody></table> |
Xet Storage Details
- Size:
- 1.62 kB
- Xet hash:
- f685822373ab99e11e8014bb69c38ac72c329bde8d538261dd249a16896a4dbc
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.