cbcc

Multi-dimensional classifier for central-bank communications produced by the CBCommunication training pipeline (multi_task_gist rung).

Provenance

Field	Value
Trainer	multi_task_gist
Model kind	multi_task
Encoder body	Qwen/Qwen3-Embedding-4B
Loss	cached_gist
Taxonomy version	2026-04-rev2 (sha256 e7c237aac8db66ca)
Training examples	3584
Validation examples	1809
Git commit	9d90b862 (dirty)
Created	2026-04-28T02:41:30.780109+00:00

Dimensions and labels

topic (21 classes)

• Climate change
• Crisis management
• Currency circulation and management
• Financial inclusion
• Financial stability
• Fiscal policy
• Governance
• MP - balance sheet size and asset purchase programs
• MP - credit
• MP - economic activity
• MP - exchange rate
• MP - inflation
• MP - interest rate
• MP - labor market
• MP - open market operations
• MP - reserve requirements
• Metadata
• Payment system
• Structural economic reform
• Supervision and regulation
• Technological innovation and fintech

temporal_orientation (2 classes)

• Backward-looking
• Forward-looking

audience (6 classes)

• Business Sector
• Financial Sector
• General Public
• Government
• International Stakeholders
• Metadata

sentiment (6 classes)

• Confidence-building
• Dovish
• Hawkish
• Neutral/Balanced
• Not applicable
• Risk-highlighting

Evaluation (held-out validation set)

Dimension	Macro F1
topic	0.8008
temporal_orientation	0.8957
audience	0.7590
sentiment	0.7511

Intended use

Classify sentences from central-bank speeches, press releases, and financial-stability reports along the four CBC taxonomy dimensions (topic, temporal orientation, audience, sentiment). Produced for research and policy analysis at the IMF.

Limitations

• Trained on a small labeled set; tail classes (low support) carry less reliable per-class metrics.
• Multilingual coverage depends on the encoder; bge-m3 is competent in the official UN languages but performance varies.
• Sentiment / temporal labels reflect the taxonomy decision rules in the source workbook; downstream consumers should re-read those rules before interpreting per-class deltas.

How to load

# Recommended (canonical port — Phase 1.1, 2026-04):
from cb_communication.processing.classification import load_named_classifier

# Auto-resolves the artefact via config/classifiers.toml. When the
# artefact is absent locally and ``hub_repo`` is configured, the loader
# pulls from this Hub repo automatically (private collaborators set
# their own ``HF_TOKEN``).
with load_named_classifier("cbcc") as clf:
    results = clf.classify_chunk([
        "Inflation expectations remain anchored at 2 percent.",
    ])

Alternative — explicit path-based load:

from huggingface_hub import snapshot_download
from cb_communication.processing.classification import load_classifier

local = snapshot_download(repo_id="thiagochris/cbcc")
with load_classifier(local) as clf:
    ...

Both call paths satisfy the canonical MultiTaskClassifier Protocol (see cb_communication/processing/classification/multi_task_classifier.py) — the runtime dispatches on model_kind from manifest.json.

Serving optimisations

This artefact has been optimised for inference (see serving_manifest.json). Production-relevant flags:

Flag	Value
LoRA adapters merged	yes
Weight dtype	bf16
Head format	sklearn_joblib
safetensors mmap	yes
Optimised at	2026-05-02T20:53:23.465044+00:00