Add new SentenceTransformer model with an onnx backend

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "USER-bge-m3",
+  "_name_or_path": "deepvk/USER-bge-m3",
   "architectures": [
     "XLMRobertaModel"
   ],
@@ -21,7 +21,7 @@
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
   "torch_dtype": "float32",
-  "transformers_version": "4.41.2",
+  "transformers_version": "4.48.3",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 46166

config_sentence_transformers.json

@@ -1,10 +1,10 @@
 {
   "__version__": {
-    "sentence_transformers": "3.0.1",
-    "transformers": "4.41.2",
-    "pytorch": "2.1.0"
+    "sentence_transformers": "4.0.2",
+    "transformers": "4.48.3",
+    "pytorch": "2.5.1+cu124"
   },
   "prompts": {},
   "default_prompt_name": null,
-  "similarity_fn_name": null
+  "similarity_fn_name": "cosine"
 }

onnx/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbcdcd4a487d286e7b271ea41dad515d3f4e9b62eba33e84056646ec41cc2f18
+size 1432521616

tokenizer.json

@@ -85,8 +85,8 @@
   "pre_tokenizer": {
     "type": "Metaspace",
     "replacement": "▁",
-    "add_prefix_space": true,
-    "prepend_scheme": "always"
+    "prepend_scheme": "always",
+    "split": true
   },
   "post_processor": {
     "type": "TemplateProcessing",
@@ -172,8 +172,8 @@
   "decoder": {
     "type": "Metaspace",
     "replacement": "▁",
-    "add_prefix_space": true,
-    "prepend_scheme": "always"
+    "prepend_scheme": "always",
+    "split": true
   },
   "model": {
     "type": "Unigram",
@@ -184846,4 +184846,4 @@
     ],
     "byte_fallback": false
   }
-}
+}

tokenizer_config.json

@@ -45,6 +45,7 @@
   "clean_up_tokenization_spaces": true,
   "cls_token": "<s>",
   "eos_token": "</s>",
+  "extra_special_tokens": {},
   "mask_token": "<mask>",
   "max_length": 512,
   "model_max_length": 8192,