| --- |
| license: mit |
| tags: |
| - coreml |
| - sentence-transformers |
| - embedding |
| - code |
| - roberta |
| base_model: microsoft/codebert-base |
| library_name: coremltools |
| pipeline_tag: feature-extraction |
| --- |
| |
| # codebert-base — CoreML (.mlpackage) |
|
|
| CoreML conversion of [microsoft/codebert-base](https://huggingface.co/microsoft/codebert-base) for native Apple Neural Engine / GPU inference on macOS and iOS. |
|
|
| ## Files |
|
|
| | File | Description | |
| |------|-------------| |
| | `model.mlpackage/` | CoreML model (FP16, flexible shapes) | |
| | `tokenizer.json` | HF fast tokenizer | |
|
|
| ## Details |
|
|
| - **Architecture**: RoBERTa (encoder-only, no token_type_ids) |
| - **Precision**: FP16 (native ANE precision) |
| - **Compute units**: `.all` — CoreML schedules across ANE, GPU, and CPU |
| - **Input shapes**: batch=1..512, seq_len=1..512 (flexible range) |
| - **Embedding dimension**: 768 |
| |
| ## Usage with cai |
| |
| ```bash |
| cai index --embed-backend swift --embed-model "rsvalerio/codebert-base-coreml" |
| ``` |
| |
| The Swift backend downloads the `.mlpackage` from this repo, compiles it to `.mlmodelc` on first run (~30-60s), and caches the compiled model for subsequent runs. |
| |
| ## Conversion |
| |
| Converted using [rsvalerio/models](https://github.com/rsvalerio/models) CI pipeline with `coremltools`. |
| |
| ```bash |
| pip install coremltools transformers torch |
| python convert.py |
| ``` |
| |
