HRBert-mini

README.md

@@ -0,0 +1,45 @@
+---
+
+language: ["ru", "en", "be", "bg", "uk", "ro", "kz", "tg", "tat", "sv", "sl", "sr", "uz", "es", "fi"]
+
+tags:
+
+- russian
+
+- fill-mask
+
+- pretraining
+
+- embeddings
+
+- masked-lm
+
+
+license: mit
+
+widget:
+
+- text: "[MASK] категории С"
+
+---
+
+This is a distilled Herbert model for an mlm task.
+
+
+Sentence embeddings can be produced as follows:
+
+```python
+
+# pip install transformers
+
+from transformers import pipeline
+
+
+fill_mask = pipeline(
+    "fill-mask",
+    model='RabotaRu/HRBert-mini',
+    tokenizer='RabotaRu/HRBert-mini'
+)
+
+fill_mask('требуется <mask> менеджер')
+```

config.json

@@ -0,0 +1,27 @@
+{
+  "architectures": [
+    "RobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "emb_size": 384,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 384,
+  "initializer_range": 0.02,
+  "intermediate_size": 512,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 3,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.11.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 200001
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb723ea01cb97ca3041bd5a345502a97adf43f1065b0babd687007e1b80fa4e4
+size 321258804

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "cls_token": "<cls>", "mask_token": "<mask>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"special_tokens_map_file": "./HRBert2/special_tokens_map.json", "name_or_path": "./HRBert2", "tokenizer_class": "PreTrainedTokenizerFast"}