Add trained job-title NER model

README.md

@@ -1,3 +1,54 @@
----
-license: mit
----
+---
+language: en
+license: apache-2.0
+tags:
+  - ner
+  - token-classification
+  - job-title
+  - bert
+base_model: dslim/bert-base-NER
+---
+
+# Job Title NER
+
+Fine-tuned from [dslim/bert-base-NER](https://huggingface.co/dslim/bert-base-NER) to extract
+a single job title span from a job description.
+
+| Label | Meaning |
+|---|---|
+| `B-JOB_TITLE` | First token of the job title |
+| `I-JOB_TITLE` | Continuation token of the job title |
+| `O` | Not part of a job title |
+
+The model returns an empty sequence when no job title is detectable in the text.
+
+## Usage
+
+```python
+from transformers import pipeline
+
+pipe = pipeline(
+    "ner",
+    model="./job_title_ner_model",
+    aggregation_strategy="simple",
+)
+
+text = """
+We are seeking a Senior Data Engineer to join our platform team.
+You will design and maintain our data pipelines at scale.
+"""
+
+results = pipe(text)
+job_titles = [r for r in results if r["entity_group"] == "JOB_TITLE"]
+print(job_titles)
+# [{ 'entity_group': 'JOB_TITLE', 'score': 0.99, 'word': 'Senior Data Engineer', ... }]
+```
+
+## Training
+
+- **Base model:** `dslim/bert-base-NER`
+- **Dataset:** ~137K job descriptions scraped from job boards
+- **Positives:** up to 30 000 examples with a labeled title span
+- **Negatives:** up to 10 000 examples with all-O labels (no title found)
+- **Label scheme:** BIO — O / B-JOB_TITLE / I-JOB_TITLE
+- **Max sequence length:** 384 tokens

config.json

@@ -0,0 +1,40 @@
+{
+  "_num_labels": 9,
+  "add_cross_attention": false,
+  "architectures": [
+    "BertForTokenClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": null,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "eos_token_id": null,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "O",
+    "1": "B-JOB_TITLE",
+    "2": "I-JOB_TITLE"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "is_decoder": false,
+  "label2id": {
+    "B-JOB_TITLE": 1,
+    "I-JOB_TITLE": 2,
+    "O": 0
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.8.0",
+  "type_vocab_size": 2,
+  "use_cache": false,
+  "vocab_size": 28996
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:299e0d380d6f49574ff4db36447d7ce8fbae2dd9e37289483bea62340ac4d688
+size 430911284

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "backend": "tokenizers",
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "is_local": false,
+  "local_files_only": false,
+  "mask_token": "[MASK]",
+  "max_len": 512,
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}