Update Readme

README.md

@@ -51,7 +51,6 @@ model-index:
 - **Sparse retrieval**: Generates interpretable sparse vectors for efficient search
 - **Lightweight**: Only 14M parameters for fast inference - even on CPU.
 - **Ready-to-use**: Compatible with the `light-splade` package and `transformers` package.
-
 ## Model Details
 
 ### Architecture
@@ -67,6 +66,7 @@ This model is based on a compact BERT architecture optimized for Japanese text p
 - **Max Sequence Length**: 2048 tokens
 - **Vocabulary Size**: 32,768 tokens (same as [tohoku-nlp/bert-base-japanese-v3](https://huggingface.co/tohoku-nlp/bert-base-japanese-v3))
 
+
 ### Tokenization
 
 - **Tokenizer**: `BertJapaneseTokenizer` from Hugging Face Transformers
@@ -127,8 +127,11 @@ from light_splade.models.splade import SpladeEncoder
 encoder = SpladeEncoder(model_path="bizreach-inc/light-splade-japanese-14M")
 
 # Tokenize input text
-text = "日本の首都は東京です。"
-token_outputs = encoder.tokenizer(text, return_tensors="pt")
+corpus = [
+    "日本の首都は東京です。",
+    "大阪万博は2025年に開催されます。"
+]
+token_outputs = encoder.tokenizer(corpus, padding=True, return_tensors="pt")
 
 # Generate sparse representation
 with torch.inference_mode():
@@ -138,20 +141,62 @@ with torch.inference_mode():
     )
 
 print(sparse_vecs[0])
+print(sparse_vecs[1])
+```
+
+### Usage Example with `transformers` package only
+
+```python
+import torch
+from transformers import AutoTokenizer, AutoModelForMaskedLM
+
+
+def dense_to_sparse(dense: torch.tensor, idx2token: dict[int, str]) -> list[dict[str, float]]:
+    rows, cols = dense.nonzero(as_tuple=True)
+    rows = rows.tolist()
+    cols = cols.tolist()
+    weights = dense[rows, cols].tolist()
+
+    sparse_vecs = [{} for _ in range(dense.size(0))]
+    for row, col, weight in zip(rows, cols, weights):
+        sparse_vecs[row][idx2token[col]] = round(weight, 2)
+
+    for i in range(len(sparse_vecs)):
+        sparse_vecs[i] = dict(sorted(sparse_vecs[i].items(), key=lambda x: x[1], reverse=True))
+    return sparse_vecs
+
+
+MODEL_PATH = "bizreach-inc/light-splade-japanese-14M"
+device = "cuda" if torch.cuda.is_available() else "cpu"
+transformer = AutoModelForMaskedLM.from_pretrained(MODEL_PATH).to(device)
+tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
+idx2token = {idx: token for token, idx in tokenizer.get_vocab().items()}
+
+corpus = [
+    "日本の首都は東京です。",
+    "大阪万博は2025年に開催されます。"
+]
+token_outputs = tokenizer(corpus, padding=True, return_tensors="pt")
+attention_mask = token_outputs["attention_mask"].to(device)
+token_outputs = {key: value.to(device) for key, value in token_outputs.items()}
+
+with torch.inference_mode():
+    outputs = transformer(**token_outputs)
+    dense, _ = torch.max(
+        torch.log(1 + torch.relu(outputs.logits)) * attention_mask.unsqueeze(-1),
+        dim=1,
+    )
+sparse_vecs = dense_to_sparse(dense, idx2token)
+
+print(sparse_vecs[0])
+print(sparse_vecs[1])
 ```
 
 **Example Output:**
 
 ```python
-{
-    '東京': 2.07, '首都': 1.89, '日本': 1.81, '都': 1.51,
-    '江戸': 1.51, '都市': 1.42, '駐日': 1.37, 'ニッポン': 1.2,
-    '北京': 0.93, '所在': 0.91, '大阪': 0.78, 'ロンドン': 0.68,
-    '天皇': 0.6, 'ソウル': 0.54, '京都': 0.53, '明治': 0.52,
-    '札幌': 0.5, '省': 0.48, 'ジャ': 0.45, 'シティ': 0.42,
-    'ステート': 0.42, '市': 0.41, '国': 0.39
-    # ... additional terms
-}
+{'首都': 1.83, '日本': 1.71, '東京': 1.64, '富士': 1.07, '地名': 0.91, '都市': 0.8, '##都': 0.77, '領土': 0.73, 'ゼロ': 0.72, 'シティ': 0.71, '位置': 0.58, '大国': 0.58, '本社': 0.55, 'とうきょう': 0.54, '中心': 0.53, '外来': 0.45, '京都': 0.41...}
+{'202': 1.61, '大阪': 1.5, '##5': 1.33, '開催': 1.15, '万博': 1.11, '東京': 1.07, '関西': 1.06, '神戸': 0.99, '京都': 0.96, '年': 0.91, 'まつり': 0.91, '203': 0.9, '月': 0.89, '日程': 0.84, '207': 0.67, 'オリンピック': 0.57, '主催': 0.57...}
 ```
 
 
@@ -173,7 +218,7 @@ print(sparse_vecs[0])
 If you use this model in your research, please cite:
 
 ```bibtex
-@misc{light-splade-japanese-14m,
+@misc{light-splade-japanese-14M,
   title={Light SPLADE Japanese 14M},
   author={Bizreach Inc.},
   year={2025},