Commit History
-
Add trust_remote_code wrapper (custom AutoModel + pipeline), base_model metadata, weights, and expanded model card
Verified signature- Adds configuration_marketing_classifier.py / modeling_marketing_classifier.py / pipeline_marketing_classifier.py so the model loads via AutoModel.from_pretrained(..., trust_remote_code=True) and pipeline('text-classification', ..., trust_remote_code=True), enabling the native 'Use this model' widget. - Adds base_model: Snowflake/snowflake-arctic-embed-m-v2.0 to the model card so the Hub model tree links back to the base encoder. - Adds model.safetensors and pytorch_model.bin (BMse checkpoint) plus the encoder's tokenizer files. - Expands the model card with Use Case, Intended Use, and Deployment sections, and a real captured pipeline output example. - No changes to model.py / infer.py / train.py / dataset.py / filter.py / metrics.py / scoring_utils.py -- the wrapper is a pure additive repackaging, verified to produce byte-identical scores to the original MarketingClassifier class on the same input.b3a38d0
