Text Classification
Transformers
ONNX
Safetensors
bert
barge-in
backchannel
multilingual
experimental
text-embeddings-inference
Instructions to use HashNuke/barge-in-detector with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use HashNuke/barge-in-detector with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="HashNuke/barge-in-detector")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("HashNuke/barge-in-detector") model = AutoModelForSequenceClassification.from_pretrained("HashNuke/barge-in-detector", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Release from source 0c8099fb262cbc058872c34a6a275df5ae7dd1de
Browse files- README.md +4 -3
- release.json +1 -28
README.md
CHANGED
|
@@ -22,8 +22,8 @@ and phrase supplement, across all 14 languages.
|
|
| 22 |
| Format | Download |
|
| 23 |
| --- | --- |
|
| 24 |
| Unquantized fine-tune, FP32 PyTorch | [model.safetensors](model.safetensors) |
|
| 25 |
-
| FP32 ONNX | [onnx/model
|
| 26 |
-
| Quantized INT8 ONNX | [onnx/
|
| 27 |
|
| 28 |
Tokenizer/configuration accompany the checkpoint. The ONNX directory includes
|
| 29 |
its tokenizer and input/output metadata. Quantization applies to matrix
|
|
@@ -49,7 +49,8 @@ with torch.inference_mode():
|
|
| 49 |
p_barge_in = model(**inputs).logits.softmax(-1)[0, 1].item()
|
| 50 |
```
|
| 51 |
|
| 52 |
-
|
|
|
|
| 53 |
[source transcript contract](https://github.com/vxpipe/barge-in-detector/blob/main/docs/transcript-format.md).
|
| 54 |
ONNX Runtime CPU inference uses int64 `input_ids` and `attention_mask`; apply
|
| 55 |
softmax to the two output logits and read index 1. Serving code lives in the
|
|
|
|
| 22 |
| Format | Download |
|
| 23 |
| --- | --- |
|
| 24 |
| Unquantized fine-tune, FP32 PyTorch | [model.safetensors](model.safetensors) |
|
| 25 |
+
| FP32 ONNX | [onnx/model.onnx](onnx/model.onnx) |
|
| 26 |
+
| Quantized INT8 ONNX | [onnx/model_int8.onnx](onnx/model_int8.onnx) |
|
| 27 |
|
| 28 |
Tokenizer/configuration accompany the checkpoint. The ONNX directory includes
|
| 29 |
its tokenizer and input/output metadata. Quantization applies to matrix
|
|
|
|
| 49 |
p_barge_in = model(**inputs).logits.softmax(-1)[0, 1].item()
|
| 50 |
```
|
| 51 |
|
| 52 |
+
Use a release tag to select a version, or a full commit SHA for immutable
|
| 53 |
+
reproduction; release tags may be updated for authorized corrections. Clean inputs using the
|
| 54 |
[source transcript contract](https://github.com/vxpipe/barge-in-detector/blob/main/docs/transcript-format.md).
|
| 55 |
ONNX Runtime CPU inference uses int64 `input_ids` and `attention_mask`; apply
|
| 56 |
softmax to the two output logits and read index 1. Serving code lives in the
|
release.json
CHANGED
|
@@ -41,32 +41,5 @@
|
|
| 41 |
"tokenizer.json": "255f5e32cb320b4dc2b31e0f203832fecafe87ca4cc87c3c41eb8ccae73f441a",
|
| 42 |
"tokenizer_config.json": "13660843154d3fde0f9954ad198b5a8bbf0d21cd4fc83aa4f37678d224c8481f"
|
| 43 |
},
|
| 44 |
-
"
|
| 45 |
-
"candidates/e5-phrase-run1/LICENSE",
|
| 46 |
-
"candidates/e5-phrase-run1/NOTICE",
|
| 47 |
-
"candidates/e5-phrase-run1/README.md",
|
| 48 |
-
"candidates/e5-phrase-run1/artifacts.sha256.json",
|
| 49 |
-
"candidates/e5-phrase-run1/checkpoint/config.json",
|
| 50 |
-
"candidates/e5-phrase-run1/checkpoint/model.safetensors",
|
| 51 |
-
"candidates/e5-phrase-run1/checkpoint/tokenizer.json",
|
| 52 |
-
"candidates/e5-phrase-run1/checkpoint/tokenizer_config.json",
|
| 53 |
-
"candidates/e5-phrase-run1/onnx/metadata.json",
|
| 54 |
-
"candidates/e5-phrase-run1/onnx/model-fp32.onnx",
|
| 55 |
-
"candidates/e5-phrase-run1/onnx/model-int8.onnx",
|
| 56 |
-
"candidates/e5-phrase-run1/onnx/tokenizer.json",
|
| 57 |
-
"candidates/e5-phrase-run1/provenance.json",
|
| 58 |
-
"candidates/e5-phrase-run1/seed-20261115/config.json",
|
| 59 |
-
"candidates/e5-phrase-run1/seed-20261115/model.safetensors",
|
| 60 |
-
"candidates/e5-phrase-run1/seed-20261115/tokenizer.json",
|
| 61 |
-
"candidates/e5-phrase-run1/seed-20261115/tokenizer_config.json",
|
| 62 |
-
"candidates/e5-phrase-run1/seed-20261116/config.json",
|
| 63 |
-
"candidates/e5-phrase-run1/seed-20261116/model.safetensors",
|
| 64 |
-
"candidates/e5-phrase-run1/seed-20261116/tokenizer.json",
|
| 65 |
-
"candidates/e5-phrase-run1/seed-20261116/tokenizer_config.json",
|
| 66 |
-
"candidates/e5-phrase-run1/seed-20261117/config.json",
|
| 67 |
-
"candidates/e5-phrase-run1/seed-20261117/model.safetensors",
|
| 68 |
-
"candidates/e5-phrase-run1/seed-20261117/tokenizer.json",
|
| 69 |
-
"candidates/e5-phrase-run1/seed-20261117/tokenizer_config.json",
|
| 70 |
-
"releases/e5-small-run1.json"
|
| 71 |
-
]
|
| 72 |
}
|
|
|
|
| 41 |
"tokenizer.json": "255f5e32cb320b4dc2b31e0f203832fecafe87ca4cc87c3c41eb8ccae73f441a",
|
| 42 |
"tokenizer_config.json": "13660843154d3fde0f9954ad198b5a8bbf0d21cd4fc83aa4f37678d224c8481f"
|
| 43 |
},
|
| 44 |
+
"documentation_source_revision": "351fc62359378753ab2f7b598e9fd2fccdffc952"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
}
|