Remove ` (default)` from MTEB metadata

README.md

@@ -1,7 +1,6 @@
 ---
 tags:
 - mteb
-- sentence-transformers
 model-index:
 - name: NV-Embed-v2
   results:
@@ -2004,10 +2003,9 @@ model-index:
 language:
 - en
 license: cc-by-nc-4.0
-library_name: transformers
 ---
 ## Introduction
-We present NV-Embed-v2, a generalist embedding model that ranks No. 1 on the Massive Text Embedding Benchmark ([MTEB benchmark](https://huggingface.co/spaces/mteb/leaderboard))(as of Aug 30, 2024) with a score of 72.31 across 56 text embedding tasks. It also holds the No. 1 in the retrieval sub-category (a score of 62.65 across 15 tasks) in the leaderboard, which is essential to the development of RAG technology.
+We present NV-Embed-v2, a generalist embedding model that ranks No. 1 on the Massive Text Embedding Benchmark ([MTEB benchmark](https://arxiv.org/abs/2210.07316))(as of Aug 30, 2024), with 56 tasks, encompassing retrieval, reranking, classification, clustering, and semantic textual similarity tasks.
 
 NV-Embed-v2 presents several new designs, including having the LLM attend to latent vectors for better pooled embedding output, and demonstrating a two-staged instruction tuning method to enhance the accuracy of both retrieval and non-retrieval tasks. Additionally, NV-Embed-v2 incorporates a novel hard-negative mining methods that take into account the positive relevance score for better false negatives removal.
 
@@ -2049,7 +2047,7 @@ passages = [
 model = AutoModel.from_pretrained('nvidia/NV-Embed-v2', trust_remote_code=True)
 
 # get the embeddings
-max_length = 32768
+max_length = 4096
 query_embeddings = model.encode(queries, instruction=query_prefix, max_length=max_length)
 passage_embeddings = model.encode(passages, instruction=passage_prefix, max_length=max_length)
 
@@ -2064,7 +2062,7 @@ passage_embeddings = F.normalize(passage_embeddings, p=2, dim=1)
 
 scores = (query_embeddings @ passage_embeddings.T) * 100
 print(scores.tolist())
-# [[87.42693328857422, 0.46283677220344543], [0.965264618396759, 86.03721618652344]]
+# [[87.42692565917969, 0.462837278842926], [0.9652643203735352, 86.0372314453125]]
 ```
 
 
@@ -2091,7 +2089,7 @@ passages = [
 
 # load model with tokenizer
 model = SentenceTransformer('nvidia/NV-Embed-v2', trust_remote_code=True)
-model.max_seq_length = 32768
+model.max_seq_length = 4096
 model.tokenizer.padding_side="right"
 
 def add_eos(input_examples):
@@ -2114,7 +2112,7 @@ For commercial purpose, we recommend you to use the models of [NeMo Retriever Mi
 
 
 ## Correspondence to
-Chankyu Lee ([email protected]), Rajarshi Roy ([email protected]), Wei Ping ([email protected])
+Chankyu Lee ([email protected]), Wei Ping ([email protected])
 
 
 ## Citation

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "nvidia/NV-Embed-v2",
+  "_name_or_path": "nvidia/NV-Embed-v1",
   "add_eos": true,
   "add_pad_token": true,
   "architectures": [
@@ -18,7 +18,7 @@
   "model_type": "nvembed",
   "padding_side": "right",
   "text_config": {
-    "_name_or_path": "nvidia/NV-Embed-v2",
+    "_name_or_path": "nvidia/NV-Embed-v1",
     "add_cross_attention": false,
     "architectures": [
       "MistralModel"

configuration_nvembed.py

@@ -76,8 +76,6 @@ class LatentAttentionConfig(PretrainedConfig):
         self.latent_dim = latent_dim
         self.cross_dim_head = cross_dim_head
 
-        super().__init__(**kwargs)
-
 
 class BidirectionalMistralConfig(MistralConfig):
     model_type = BIDIR_MISTRAL_TYPE