Upload folder using huggingface_hub

README.md

@@ -1,3 +1,57 @@
----
-license: mit
----
+---
+language: en
+tags:
+  - sentence-transformers
+  - embeddings
+  - semantic-search
+  - retrieval
+license: mit
+---
+
+# BiEncoder RoPE — Sentence Embedding Model
+
+A 34M parameter sentence embedding model trained from scratch using PyTorch.
+
+## Architecture
+- 6-layer Transformer encoder with RoPE positional embeddings
+- Mean pooling + L2 normalization
+- 256-dim output vectors
+
+## Training (Curriculum)
+| Phase | Dataset | Loss |
+|---|---|---|
+| 1 | all-nli | MNRLoss |
+| 2 | squad | MNRLoss |
+| 3 | msmarco-bm25 | HardNegativeLoss |
+| 4 | natural-questions | MNRLoss |
+
+## Files
+- `tokenizer/` — HuggingFace tokenizer (bert-base-uncased)
+- `pytorch/checkpoint_phase4_nq.pt` — PyTorch weights
+- `onnx/biencoder_rope.onnx` — ONNX FP32
+- `onnx/biencoder_rope_int8.onnx` — ONNX INT8 (recommended for CPU)
+
+## Usage
+```python
+import torch
+from transformers import AutoTokenizer
+
+tokenizer = AutoTokenizer.from_pretrained("your-username/your-model-name", subfolder="tokenizer")
+model = BiEncoderRoPE().to("cuda")
+model.load_state_dict(
+    torch.load("pytorch/checkpoint_phase4_nq.pt")["model_state"]
+)
+model.eval()
+
+@torch.no_grad()
+def encode(texts):
+    if isinstance(texts, str): texts = [texts]
+    enc = tokenizer(texts, padding=True, truncation=True,
+                    max_length=256, return_tensors="pt")
+    return model.encode(enc["input_ids"].cuda(), enc["attention_mask"].cuda()).cpu()
+```
+
+## Performance
+- FP32 ONNX size : 134.3 MB
+- INT8 ONNX size : 34.6 MB
+- Throughput     : ~247 sentences/sec on CPU

onnx/biencoder_rope.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b4f3959a339f71bb506da4595d23bde7358a70f4a188286ece4b9f4dcf2d004
+size 140864188

onnx/biencoder_rope_int8.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3585cf0eb03c22ce6005097068c534b6c15accef55150cc71824051524af2061
+size 36265371

pytorch/checkpoint_phase4_nq.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a94028c66ce14e6802c17667a469af03e37fd0ca0f63118fe194dde150f9c18b
+size 425475351

tokenizer/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "backend": "tokenizers",
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "is_local": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}