Update README.md

README.md

@@ -37,8 +37,8 @@ configs:
 ## Dataset details:-
 
 - This dataset is the final version of anchor(query)-positive(chunk) pair data w.r.t fine tuning embedding model for clinical trials dataset.
-- It includes best of both ***<u> 4 anchors-consolidated positive chunk/nctId dataset-->[first dataset](https://huggingface.co/datasets/vab46/Clinical_trials_anchor-positive-pairs_EmbeddingModel-data)</u>*** and
-  ***<u>5 anchors-3 positive chunk/nctId--->[Second dataset](https://huggingface.co/datasets/vab46/Clinical_trials_anchor-positive-pairs_EmbeddingModel-data2)</u>***.
+- It includes best of both ***<u> 4 anchors-consolidated positive chunk/nctId dataset</u>-->[first dataset](https://huggingface.co/datasets/vab46/Clinical_trials_anchor-positive-pairs_EmbeddingModel-data)*** and
+  ***<u>5 anchors-3 positive chunk/nctId</u>--->[Second dataset](https://huggingface.co/datasets/vab46/Clinical_trials_anchor-positive-pairs_EmbeddingModel-data2)***.
 - The 1st dataset(consolidated *title +summary+ inclusion criteria* chunk) suffered with pre-processing bottlenecks :-
   1. rendering huge chunks upto 15k characeters.
   2. missing on certain specific corner cases of inclusion-exclusion segregation.
@@ -59,6 +59,6 @@ configs:
 - All these along with other amend mentioned in [2nd dataset readme.md](https://huggingface.co/datasets/vab46/Clinical_trials_anchor-positive-pairs_EmbeddingModel-data2) helped
   base results to advance.
      
-  1. <u>**ndcg@10 being 0.40-0.52** on mastroyska nested dimensions 64D- 768D of "nomic-ai/nomic-embed-text-v1.5 base model"</u>.
-  2. <u>the initial ***sequential evaluator score(on 5 diff mashtroyska embeddings) ~0.405***. Both these values improved significantly(***avg @ndcg@10 & seq_eval score
-     tending 25% and 40%***) which can be observed in details in metrics at [readme.md of final model](https://huggingface.co/vab46/nomic-embed-text-v1.5_Clinical-Trials_Matryoshka_final)</u>.   
+  1. **ndcg@10 being 0.40-0.52** on mastroyska nested dimensions 64D- 768D of "nomic-ai/nomic-embed-text-v1.5 base model".
+  2. The initial ***sequential evaluator score(on 5 diff mashtroyska embeddings) ~0.405***. Both these values improved significantly(***avg @ndcg@10 & seq_eval score
+     tending 25% and 40%***) which can be observed in details in metrics at [readme.md of final model](https://huggingface.co/vab46/nomic-embed-text-v1.5_Clinical-Trials_Matryoshka_final).