Upload 10 files

1_Pooling/config.json

@@ -0,0 +1,7 @@
+{
+  "word_embedding_dimension": 384,
+  "pooling_mode_cls_token": true,
+  "pooling_mode_mean_tokens": false,
+  "pooling_mode_max_tokens": false,
+  "pooling_mode_mean_sqrt_len_tokens": false
+}

README.md

@@ -1,3 +1,73 @@
----
-license: mit
----
+---
+language:
+- ru
+- en
+
+pipeline_tag: sentence-similarity
+
+tags:
+- russian
+- pretraining
+- embeddings
+- feature-extraction
+- sentence-similarity
+- sentence-transformers
+- transformers
+- mteb
+
+datasets:
+- IlyaGusev/gazeta
+- zloelias/lenta-ru
+- HuggingFaceFW/fineweb-2
+- HuggingFaceFW/fineweb
+
+license: mit
+
+base_model: sergeyzh/rubert-mini-uncased
+
+---
+
+Модель BERT для задач симметричного сравнения запросов (query). Получена дистилляцией эмбеддингов русских и английских текстов с учётом префикса [Qwen/Qwen3-Embedding-8B](https://huggingface.co/Qwen/Qwen3-Embedding-8B). Модель принадлежит к виду uncased - не различает при обработке текста буквы, написанные в верхнем и нижнем регистре.
+
+Модель может использоваться для кэшировании и фильтрации запросов к LLM, Q2Q RAG.
+
+Основные характеристики модели: 
+ - размер ембеддинга - 384, 
+ - длина контекста - 512, 
+ - слоёв - 7,
+ - префиксы - не требуются. 
+
+
+
+## Использование
+
+```Python
+from sentence_transformers import SentenceTransformer
+
+model = SentenceTransformer('sergeyzh/rubert-mini-uncased-query')
+
+sentences = ["восстановление доступа", "как сбросить пароль"]
+embeddings = model.encode(sentences)
+print(model.similarity(embeddings, embeddings))
+```
+
+
+
+## Метрики
+
+Оценка качества модели для русского языка выполнена на сравнении близости пар поисковых запросов датасета [ai-forever/rubq-retrieval](https://huggingface.co/datasets/ai-forever/rubq-retrieval) с ответами референсной модели [Qwen/Qwen3-Embedding-8B](https://huggingface.co/Qwen/Qwen3-Embedding-8B):
+
+| Модель | Pearson r | Spearman ρ |
+|--------|-----------|------------|
+| [Qwen/Qwen3-Embedding-8B](https://huggingface.co/Qwen/Qwen3-Embedding-8B) | 1.000 | 1.000 |
+| [sergeyzh/rubert-large-uncased-query](https://huggingface.co/sergeyzh/rubert-large-uncased-query) | 0.850 | 0.800 |
+| [Qwen/Qwen3-Embedding-4B](https://huggingface.co/Qwen/Qwen3-Embedding-4B) | 0.845 | 0.788 |
+| [Qwen/Qwen3-Embedding-0.6B](https://huggingface.co/Qwen/Qwen3-Embedding-0.6B) | 0.716 | 0.655 |
+| [**sergeyzh/rubert-mini-uncased-query**](https://huggingface.co/sergeyzh/rubert-mini-uncased-query) | 0.714 | 0.641 |
+| [intfloat/multilingual-e5-large-instruct](https://huggingface.co/intfloat/multilingual-e5-large-instruct) | 0.707 | 0.636 |
+| [intfloat/e5-large](https://huggingface.co/intfloat/e5-large) | 0.654 | 0.570 |
+| [sergeyzh/rubert-mini-frida](https://huggingface.co/sergeyzh/rubert-mini-frida) | 0.638 | 0.539 |
+| [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) | 0.630 | 0.533 |
+| [ai-forever/FRIDA](https://huggingface.co/ai-forever/FRIDA) | 0.623 | 0.533 |
+
+

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "sergeyzh/rubert-mini-uncased-query",
+  "architectures": [
+    "BertModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "emb_size": 384,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 384,
+  "initializer_range": 0.02,
+  "intermediate_size": 768,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 7,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "transformers_version": "4.57.6",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 74272
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fbe3e01dcb76048f324eed1c99073d4157daa0a37450965eb645bc991fcd5fc
+size 148627448

modules.json

@@ -0,0 +1,20 @@
+[
+  {
+    "idx": 0,
+    "name": "0",
+    "path": "",
+    "type": "sentence_transformers.models.Transformer"
+  },
+  {
+    "idx": 1,
+    "name": "1",
+    "path": "1_Pooling",
+    "type": "sentence_transformers.models.Pooling"
+  },
+  {
+    "idx": 2,
+    "name": "2",
+    "path": "2_Normalize",
+    "type": "sentence_transformers.models.Normalize"
+  }
+]

sentence_bert_config.json

@@ -0,0 +1,4 @@
+{
+  "max_seq_length": 512,
+  "do_lower_case": false
+}

special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "max_length": 512,
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_to_multiple_of": null,
+  "pad_token": "[PAD]",
+  "pad_token_type_id": 0,
+  "padding_side": "right",
+  "sep_token": "[SEP]",
+  "stride": 0,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "truncation_side": "right",
+  "truncation_strategy": "longest_first",
+  "unk_token": "[UNK]"
+}