Text Classification
Transformers
ONNX
Safetensors
English
modernbert
agents
context-management
topic-shift
cross-encoder
text-embeddings-inference
Instructions to use danielrosehill/Agent-Task-Break-Detector-ModernBERT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use danielrosehill/Agent-Task-Break-Detector-ModernBERT with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="danielrosehill/Agent-Task-Break-Detector-ModernBERT")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("danielrosehill/Agent-Task-Break-Detector-ModernBERT") model = AutoModelForSequenceClassification.from_pretrained("danielrosehill/Agent-Task-Break-Detector-ModernBERT", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download new_task_split.json from danielrosehill/Agent-Task-Break-Detector-ModernBERT: direct link, hf CLI and curl.
- Browser
- Download file 353 Bytes
-
https://huggingface.co/danielrosehill/Agent-Task-Break-Detector-ModernBERT/resolve/main/new_task_split.json
- Command line
-
hf download hf://danielrosehill/Agent-Task-Break-Detector-ModernBERT/new_task_split.json
-
curl -L -o new_task_split.json https://huggingface.co/danielrosehill/Agent-Task-Break-Detector-ModernBERT/resolve/main/new_task_split.json
353 Bytes
| {"threshold": 0.76, "threshold_basis": "int8 ONNX, validation n=697: threshold 0.751 gives precision 0.952 / recall 0.921; rounded up. Test n=608 at 0.751: precision 0.941 / recall 0.893. Calibrated 2026-10-24.", "max_len": 512, "max_turns": 4, "turn_chars": 600, "onnx": "onnx/model_int8.onnx", "tokenizer": "onnx/tokenizer.json", "label_new_task": 1} | |