ONNX FP16 export of openai/privacy-filter for browser inference

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+onnx/model_fp16.onnx.data filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,78 @@
+---
+license: apache-2.0
+base_model: openai/privacy-filter
+tags:
+  - token-classification
+  - pii-detection
+  - onnx
+  - browser
+  - privacy
+  - transformers.js
+library_name: transformers
+pipeline_tag: token-classification
+---
+
+# Privacy Filter - ONNX (FP16)
+
+FP16 ONNX export of [openai/privacy-filter](https://huggingface.co/openai/privacy-filter)
+for in-browser inference via onnxruntime-web. Detects 8 categories of personally
+identifiable information (PII) and returns BIOES token labels.
+
+## Files
+
+- `onnx/model_fp16.onnx` - graph
+- `onnx/model_fp16.onnx.data` - weights (external data, ~2.6 GB)
+- `tokenizer.json`, `tokenizer_config.json`, `special_tokens_map.json` - tokenizer
+- `config.json` - model config with the 33 BIOES label taxonomy
+- `viterbi_calibration.json` - default operating-point biases for the Viterbi decoder
+
+## Label taxonomy (33 classes)
+
+Background class `O` plus BIOES tags (`B-`, `I-`, `E-`, `S-`) for 8 span categories:
+
+- `account_number`
+- `private_address`
+- `private_date`
+- `private_email`
+- `private_person`
+- `private_phone`
+- `private_url`
+- `secret`
+
+## Usage (browser, onnxruntime-web)
+
+```javascript
+import * as ort from 'onnxruntime-web';
+
+const session = await ort.InferenceSession.create(
+  'https://huggingface.co/YOUR_REPO/resolve/main/onnx/model_fp16.onnx',
+  { executionProviders: ['webgpu', 'wasm'] }
+);
+
+// Tokenize with @huggingface/tokenizers using tokenizer.json from this repo.
+// Feed int64 input_ids and attention_mask. Output is logits [batch, seq, 33].
+// Decode with a constrained BIOES Viterbi pass using viterbi_calibration.json.
+```
+
+Full browser runner (tokenizer + ONNX + Viterbi decoder in JS) is in the
+conversion project's `web/` folder.
+
+## Export notes
+
+- Exported with `torch.onnx.export(dynamo=True)` from `transformers>=5.6.0.dev0`
+- MoE blocks (128 experts top-4) rewritten to a dense-weighted-sum form for
+  ONNX compatibility while preserving reference math
+- FP16 precision (original is BF16). Keeps int64 inputs/outputs
+- Dynamic axes on batch and sequence length. Practical browser range: 256-4096
+  tokens depending on memory
+- Parity vs reference PyTorch: 100% argmax agreement on seed prompts
+
+## License
+
+Apache 2.0, same as the base model.
+
+## Acknowledgements
+
+Base model by OpenAI. See the
+[original model card](https://huggingface.co/openai/privacy-filter) for
+training details, intended use, and limitations.

onnx/model_fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0daae3b14e399f45db0d39b9f73ee2512423a9b7317a72a8534482fec665f51f
+size 1965782

onnx/model_fp16.onnx.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7d1002bc3d04553727fc5fa6b7059d93bc5372b3b7ca9d01a0ebfb2f9366327
+size 2798967296