Views
No views yet
phase2_infini run. The checkpoint was selected at epoch 4 by lowest combined validation loss.| Metric | Validation result |
|---|---|
| Combined loss | 1.3988 |
| Next-event accuracy | 0.5575 |
| Hazard probability MAE | 0.0495 |
target_hazard > 0.5). Metrics are specific to the held-out windows used by this training run and should not be treated as broad real-world benchmarks.vocabs.json contains the exact training vocabularies. Unknown values should map to the <unk> entry and padding to <pad>.1from load_model import load_model
2
3model, vocabs, config = load_model("sirus/modernnews-event-stream")1outputs = model(
2 event_types=event_types, # [batch, sequence]
3 attention_mask=attention_mask, # [batch, sequence]
4 time_deltas=time_deltas, # [batch, sequence]
5 severities=severities, # [batch, sequence]
6 novelties=novelties, # [batch, sequence]
7 uncertainties=uncertainties, # [batch, sequence]
8 rolling_features=rolling_features, # [batch, sequence, 15]
9 topic_token_ids=topic_token_ids, # [batch, sequence, topics/event]
10 topic_token_mask=topic_token_mask, # same shape as topic_token_ids
11 location_token_ids=location_token_ids, # [batch, sequence, locations/event]
12 location_token_mask=location_token_mask, # same shape as location_token_ids
13 adjacency=adjacency, # optional [batch, batch]
14)
15
16next_event_ids = outputs.event_logits.argmax(dim=-1)
17hazard_probabilities = outputs.hazard_logits.sigmoid()MODERNNEWS_DISABLE_RWKV_EXT=1 to force the fallback.event_model.pt) and vocabularies (vocabs.json) are licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. You may share and adapt those artifacts with attribution, but you may not use them for commercial purposes. Contact the model author through Hugging Face to discuss separate commercial licensing.load_model.py and modernnews/) remains available under the MIT License. See LICENSE for the exact file-level split.