all-miniLM-L6-v4-nepali

SentenceTransformer based on jangedoo/all-MiniLM-L6-v3-nepali

This is a sentence-transformers model finetuned from jangedoo/all-MiniLM-L6-v3-nepali on the title_excerpt, ne_en, excerpt_paraphrase, nepali_triplets, stsb_en, stsb_ne, nepali_qa_9k, ecommerce, cross_lingual and sanoir_general datasets. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: jangedoo/all-MiniLM-L6-v3-nepali <!-- at revision 37d42341db65da0787d41e0f81c977ab4fa510a5 -->
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text
• Training Datasets:
  • title_excerpt
  • ne_en
  • excerpt_paraphrase
  • nepali_triplets
  • stsb_en
  • stsb_ne
  • nepali_qa_9k
  • ecommerce
  • cross_lingual
  • sanoir_general
• Languages: ne, en

<!-- - **License:** Unknown -->

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
  (1): Pooling({'embedding_dimension': 384, 'pooling_mode': 'mean', 'include_prompt': True})
  (2): Normalize({})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("jangedoo/all-miniLM-L6-v4-nepali")
# Run inference
queries = [
    'yo dictionary ko quality kasto cha ra reviews hernu hai',
]
documents = [
    'The Oxford English-Hindi Dictionary (Pocket Edition) has 1,200 pages with over 40,000 entries, priced at NPR 890. It features phonetic transcriptions and clear usage examples, making it ideal for Nepali students learning English. Users on SastoDeal have rated it 4.5 stars, praising its compact size and accurate translations for exam preparation.',
    'The Nepali Best Friend Notebook (A5, 200 pages) is made of recycled paper with a hardcover binding, costing NPR 350. It comes in six vibrant colors and is suitable for school notes or journaling. Many customers appreciate its smooth pages that prevent ink bleed-through.',
    'This Faber-Castell Watercolor Paint Set (12 colors) includes a brush and mixing palette, retailing at NPR 1,200. It is non-toxic and suitable for children aged 6 and above. Artists recommend it for its vibrant pigmentation and easy blending.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 384] [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 0.5677, -0.0084, -0.1193]])

<!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use *List how the model may foreseeably be misused and address what users ought not to do with the model.* -->

Evaluation

Metrics

Information Retrieval

• Datasets: multi_lang_news_ir, en_news_ir, ne_news_ir and nepali_qa_9k
• Evaluated with <code>InformationRetrievalEvaluator</code> with these parameters:

  {
      "query_prompt_name": "query",
      "corpus_prompt_name": "document"
  }
  

Translation

• Dataset: translation
• Evaluated with <code>TranslationEvaluator</code>

Semantic Similarity

• Datasets: stsb_en and stsb_ne
• Evaluated with <code>EmbeddingSimilarityEvaluator</code>

Information Retrieval

• Datasets: nepali_query_passage_10k, nepali_ecomm and sanoir_general
• Evaluated with <code>InformationRetrievalEvaluator</code> with these parameters:

  {
      "query_prompt_name": "query",
      "corpus_prompt_name": "document"
  }
  

<!-- ## Bias, Risks and Limitations *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.* --> <!-- ### Recommendations *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.* -->

Training Details

Training Datasets

<details><summary>title_excerpt</summary>

title_excerpt

• Dataset: title_excerpt at 1582ca1
• Size: 102,803 training samples
• Columns: <code>title</code> and <code>excerpt</code>
• Approximate statistics based on the first 100 samples:

  	title	excerpt
  type	string	string
  modality	text	text
  details	<ul><li>min: 8 tokens</li><li>mean: 28.58 tokens</li><li>max: 73 tokens</li></ul>	<ul><li>min: 22 tokens</li><li>mean: 72.08 tokens</li><li>max: 154 tokens</li></ul>

• Samples:

  title	excerpt
  <code>Electricity poles erected a year ago, but villages still without power</code>	<code>Residents of six settlements in Kumakh Rural Municipality, Salyan district, remain without electricity a year after poles were erected due to contractor negligence and transformer supply issues.</code>
  <code>नेप्सेले सार्वजनिक गर्यो धितोपत्र व्यापारीका लागि कारोबारयोग्य ११० कम्पनीको नामावली, कुन–कुन परे ?</code>	<code>नेपाल स्टक एक्सचेञ्जले धितोपत्र व्यापारीका लागि कारोबार योग्य ११० कम्पनीहरूको नामावली सार्वजनिक गरेको छ। यसमा बैंक, फाइनान्स, बीमा, जलविद्युत लगायतका कम्पनीहरू समावेश छन्।</code>
  <code>एसीसी यू-१६ इस्ट जोन कपको फाइनलमा पहिले बलिङ गर्दै नेपाल</code>	<code>एसीसी पुरुष यू-१६ इस्ट जोन कपको फाइनलमा नेपालले पहिलो बलिङ गर्ने निर्णय लिएको छ। सिंगापुरले टस जितेर ब्याटिङ रोज्दा नेपालले फिल्डिङ गर्ने भयो।</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>ne_en</summary>

ne_en

• Dataset: ne_en at e4c18f5
• Size: 3,765 training samples
• Columns: <code>title</code> and <code>translation</code>
• Approximate statistics based on the first 100 samples:

  	title	translation
  type	string	string
  modality	text	text
  details	<ul><li>min: 6 tokens</li><li>mean: 28.26 tokens</li><li>max: 68 tokens</li></ul>	<ul><li>min: 3 tokens</li><li>mean: 25.85 tokens</li><li>max: 64 tokens</li></ul>

• Samples:

  title	translation
  <code>FWLD pushes for inclusive, non-discriminatory citizenship law</code>	<code>FWLD समावेशी, गैर-भेदभावरहित नागरिकता कानूनको लागि जोड दिन्छ</code>
  <code><br> Clouded leopard: A vanishing jewel of the forest </code>	<code><br> बादलयुक्त चितुवा: जंगलको हराउने रत्न </code>
  <code>सिटिजन्स सदाबहार इकाइमा आवेदन म्याद थप</code>	<code>Application deadline extended to Citizens Sadabahar Unit</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>excerpt_paraphrase</summary>

excerpt_paraphrase

• Dataset: excerpt_paraphrase at b521e5a
• Size: 549 training samples
• Columns: <code>sentence1</code> and <code>sentence2</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2
  type	string	string
  modality	text	text
  details	<ul><li>min: 27 tokens</li><li>mean: 73.15 tokens</li><li>max: 149 tokens</li></ul>	<ul><li>min: 32 tokens</li><li>mean: 77.3 tokens</li><li>max: 160 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>भरतपुर कारागारमा मौसम परिवर्तन र वर्षातका कारण २२ कैदीबन्दीलाई रुघा र ज्वरोको समस्या देखिएको छ र उनीहरूको छुट्टै उपचार भैरहेको छ।</code>	<code>मौसममा आएको बदलाव र वर्षाका कारण भरतपुर जेलमा २२ जना कैदीबन्दीलाई रुघाखोकी र ज्वरोले सताएको छ, जसका लागि उनीहरूको विशेष उपचार भइरहेको छ।</code>
  <code>Heavy to very heavy rainfall is likely today in six provinces of Nepal due to the monsoon trough near its average position, with thunder and lightning expected in several regions.</code>	<code>Intense precipitation is anticipated this day across six Nepalese provinces, owing to the monsoon trough maintaining its usual course, and electrical storms are predicted in various areas.</code>
  <code>भोजपुरको रामप्रसाद राई गाउँपालिका–६ बैकुण्ठेका पाँच सामुदायिक विद्यालयका ४३८ विद्यार्थीलाई पोसाक वितरण गरिएको छ। वडाले विद्यालयबीच एकरुपता ल्याउने लक्ष्यले पोसाक, जुत्ता र टिसर्ट वितरण गरेको हो।</code>	<code>रामप्रसाद राई गाउँपालिका–६, भोजपुरको बैकुण्ठेस्थित पाँच सामुदायिक विद्यालयका ४३८ जना विद्यार्थीहरूलाई वडाले विद्यालयहरूमा एकरूपता कायम गर्नका लागि पोसाक, जुत्ता तथा टिसर्ट प्रदान गरेको छ।</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>nepali_triplets</summary>

nepali_triplets

• Dataset: nepali_triplets at bd8c184
• Size: 1,600 training samples
• Columns: <code>sentence</code>, <code>positive_sentence</code>, and <code>negative_sentence</code>
• Approximate statistics based on the first 100 samples:

  	sentence	positive_sentence	negative_sentence
  type	string	string	string
  modality	text	text	text
  details	<ul><li>min: 25 tokens</li><li>mean: 80.75 tokens</li><li>max: 177 tokens</li></ul>	<ul><li>min: 30 tokens</li><li>mean: 82.04 tokens</li><li>max: 167 tokens</li></ul>	<ul><li>min: 24 tokens</li><li>mean: 66.34 tokens</li><li>max: 143 tokens</li></ul>

• Samples:

  sentence	positive_sentence	negative_sentence
  <code>भारतका रक्षा प्रमुख जनरल अनिल चौहानले चीन, पाकिस्तान र बंगलादेशको गठबन्धनलाई भारतको सुरक्षाका लागि ठूलो खतरा भएको बताएका छन्।</code>	<code>चीन, पाकिस्तान र बंगलादेशको संयुक्त गठबन्धनलाई भारतको सुरक्षा चुनौतीको रूपमा जनरल अनिल चौहानले औंल्याएका छन्।</code>	<code>जनरल अनिल चौहानले भारतको सुरक्षाका लागि चीन र पाकिस्तानको गठबन्धन भन्दा मात्र बंगलादेशलाई ठूलो खतरा भनेका छन्।</code>
  <code>The Sarlahi District Court has granted bail to suspended Bagmati Municipality Mayor Bharat Bahadur Thapa on a Rs 5 million bond in connection with illegal forest resource extraction charges.</code>	<code>Bharat Bahadur Thapa, the suspended mayor of Bagmati Municipality, was released on bail by the Sarlahi District Court after posting a bond of Rs 5 million related to allegations of unlawful forest resource exploitation.</code>	<code>The Sarlahi District Court denied bail to Bharat Bahadur Thapa, the suspended mayor of Bagmati Municipality, in the case concerning illegal forest resource extraction.</code>
  <code>नेपाली युवा महिला फुटबल टोलीका प्रशिक्षक यामप्रसाद गुरुङले साफ यु-२० महिला च्याम्पियनसिपको उपाधि जित्ने आशा दिएका छन्। नेपालले जुलाईमा बंगलादेशमा हुने प्रतियोगितामा बलियो टोली बनाएर प्रतिस्पर्धा गर्नेछ।</code>	<code>यामप्रसाद गुरुङले साफ यु-२० महिला च्याम्पियनसिपमा नेपालको टोलीले शीर्ष स्थान हासिल गर्ने विश्वास व्यक्त गरेका छन्। आगामी जुलाईमा बंगलादेशमा आयोजना हुने प्रतियोगितामा नेपालले सशक्त टोली प्रस्तुत गर्ने तयारीमा छ।</code>	<code>नेपाली युवा महिला फुटबल टोलीले आगामी साफ यु-२० महिला च्याम्पियनसिपमा कमजोर प्रदर्शन गर्ने सम्भावना व्यक्त गरिएको छ।</code>

• Loss: <code>TripletLoss</code> with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.2
  }
  

</details> <details><summary>stsb_en</summary>

stsb_en

• Dataset: stsb_en at 4bfcc74
• Size: 5,710 training samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	score
  type	string	string	float
  modality	text	text
  details	<ul><li>min: 7 tokens</li><li>mean: 9.45 tokens</li><li>max: 14 tokens</li></ul>	<ul><li>min: 7 tokens</li><li>mean: 9.5 tokens</li><li>max: 15 tokens</li></ul>	<ul><li>min: 0.1</li><li>mean: 0.66</li><li>max: 1.0</li></ul>

• Samples:

  sentence1	sentence2	score
  <code>A plane is taking off.</code>	<code>An air plane is taking off.</code>	<code>1.0</code>
  <code>A man is playing a large flute.</code>	<code>A man is playing a flute.</code>	<code>0.76</code>
  <code>A man is spreading shreded cheese on a pizza.</code>	<code>A man is spreading shredded cheese on an uncooked pizza.</code>	<code>0.76</code>

• Loss: <code>CoSENTLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "pairwise_cos_sim"
  }
  

</details> <details><summary>stsb_ne</summary>

stsb_ne

• Dataset: stsb_ne at 4bfcc74
• Size: 5,710 training samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	score
  type	string	string	float
  modality	text	text
  details	<ul><li>min: 7 tokens</li><li>mean: 20.08 tokens</li><li>max: 34 tokens</li></ul>	<ul><li>min: 9 tokens</li><li>mean: 20.38 tokens</li><li>max: 42 tokens</li></ul>	<ul><li>min: 0.1</li><li>mean: 0.66</li><li>max: 1.0</li></ul>

• Samples:

  sentence1	sentence2	score
  <code>एउटा विमान उडिरहेको छ।</code>	<code>हवाई जहाज उडिरहेको छ।</code>	<code>1.0</code>
  <code>एउटा मान्छे ठूलो बाँसुरी बजाइरहेको छ।</code>	<code>एउटा मान्छे बाँसुरी बजाउँदै छ।</code>	<code>0.76</code>
  <code>एक मानिस पिज्जामा टुक्रा चिज फैलाउँदै छ।</code>	<code>एक जना मानिसले न पकाएको पिज्जामा टुक्रा पारेको चीज फैलाउँदै छ।</code>	<code>0.76</code>

• Loss: <code>CoSENTLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "pairwise_cos_sim"
  }
  

</details> <details><summary>nepali_qa_9k</summary>

nepali_qa_9k

• Dataset: nepali_qa_9k at 33f9def
• Size: 6,480 training samples
• Columns: <code>question</code> and <code>answer</code>
• Approximate statistics based on the first 100 samples:

  	question	answer
  type	string	string
  modality	text	text
  details	<ul><li>min: 20 tokens</li><li>mean: 45.64 tokens</li><li>max: 81 tokens</li></ul>	<ul><li>min: 53 tokens</li><li>mean: 100.33 tokens</li><li>max: 156 tokens</li></ul>

• Samples:

  question	answer
  <code>मेरो ल्यापटप धेरै ढिलो (slow) भइरहेको छ, के गर्ने?</code>	<code>ल्यापटपको प्रदर्शन सुधार गर्न, अनावश्यक सफ्टवेयर हटाउनुहोस्, RAM र प्रोसेसरको प्रयोग जाँच गर्नुहोस्, र सिस्टमलाई अपडेट गर्नुहोस्।</code>
  <code>मलाई अचानक ज्वरो आएको छ, मैले के गर्नुपर्छ?</code>	<code>यदि तपाईंलाई ज्वरो आएको छ भने, आराम गर्नुहोस्, पर्याप्त पानी पिउनुहोस्, र यदि ज्वरो धेरै छ वा असहजता महसुस हुन्छ भने तुरुन्त डाक्टरसँग परामर्श लिनुहोस्।</code>
  <code>मेरो मोबाइलको डाटा ब्याकअप कसरी गर्न सकिन्छ? विभिन्न विकल्पहरू के-के छन्?</code>	<code>तपाईंले आफ्नो मोबाइल सेटिङ्समा गएर क्लाउड सर्भिस (जस्तै Google Drive वा iCloud) प्रयोग गरेर वा विशेष ब्याकअप एपहरू मार्फत डाटा ब्याकअप गर्न सक्नुहुन्छ।</code>

• Loss: <code>CachedMultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 32,
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>ecommerce</summary>

ecommerce

• Dataset: ecommerce at 4b7dd81
• Size: 1,367 training samples
• Columns: <code>query</code>, <code>document</code>, <code>negative1</code>, <code>negative2</code>, and <code>negative3</code>
• Approximate statistics based on the first 100 samples:

  	query	document	negative1	negative2	negative3
  type	string	string	string	string	string
  modality	text	text	text	text	text
  details	<ul><li>min: 10 tokens</li><li>mean: 23.55 tokens</li><li>max: 53 tokens</li></ul>	<ul><li>min: 54 tokens</li><li>mean: 94.77 tokens</li><li>max: 166 tokens</li></ul>	<ul><li>min: 30 tokens</li><li>mean: 63.79 tokens</li><li>max: 93 tokens</li></ul>	<ul><li>min: 28 tokens</li><li>mean: 56.86 tokens</li><li>max: 88 tokens</li></ul>	<ul><li>min: 22 tokens</li><li>mean: 53.0 tokens</li><li>max: 79 tokens</li></ul>

• Samples:

  query	document	negative1	negative2	negative3
  <code>Samsung Galaxy S24 Ultra price in Nepal</code>	<code>The Samsung Galaxy S24 Ultra is now available in Nepal for NPR 1,59,999. It features a 200MP camera, S Pen support, and a powerful Snapdragon 8 Gen 3 processor. The phone comes with 12GB RAM and 256GB storage, offering a stunning 6.8-inch Dynamic AMOLED 2X display.</code>	<code>The Samsung Galaxy S24 FE is a more affordable option at NPR 75,999, with a 50MP camera and Exynos 2400 chipset. It offers 8GB RAM and 128GB storage, making it a good mid-range choice. The display is a 6.7-inch Super AMOLED with 120Hz refresh rate.</code>	<code>The Samsung Galaxy Tab S9 Ultra is a high-end tablet priced at NPR 1,49,999. It features a 14.6-inch Dynamic AMOLED 2X display and comes with the S Pen included. This tablet is ideal for creative professionals and students.</code>	<code>The Samsung 65-inch 4K Smart TV is available for NPR 1,20,000. It supports HDR10+ and has built-in Alexa. This TV offers vibrant picture quality and is perfect for home entertainment.</code>
  <code>sasto tel kun ramro fortuna ki sathi cooking oil price compare gara na</code>	<code>When comparing Fortune versus Sathi refined cooking oils, Fortune 1-liter bottle is priced at NPR 210 while Sathi's same size costs NPR 195, making Sathi a more budget-friendly option for daily use. Both are suitable for deep frying and regular cooking, but Fortune is often preferred for its lighter texture and neutral flavor. If you are looking to save on your monthly grocery bill, switching to Sathi oil can save you around NPR 15 per liter without sacrificing quality.</code>	<code>Fortune basmati rice 5kg is now available at NPR 850, which is a good deal compared to other premium brands like Kohinoor. Many customers praise its long grains and aromatic flavor, perfect for biryani and pulao. It is currently on a 10% discount for new users on our store.</code>	<code>Wai Wai noodles are a popular instant snack in Nepal, available in curry, chicken, and vegetable flavors. A single packet costs NPR 15, and bulk packs of 12 are priced at NPR 170. They are quick to prepare and loved by students and travelers.</code>	<code>Our store now offers free delivery on all orders above NPR 1000 within Kathmandu Valley. You can also earn reward points on every purchase of groceries and household items. Check out the new arrivals section for exclusive discounts this week.</code>
  <code>CDC Grade 10 Nepali textbook quality kasto cha? review ra ratings hernu hai</code>	<code>For students using the CDC Grade 10 Nepali textbook, our guide shows how to annotate and highlight key poems like 'Mero Desh' for better exam preparation. We recommend Pilot V7 pens for clean underlining and Stabile Boss highlighters in yellow, available at NPR 85 and NPR 120 per piece at Sajha Bookstore. Users report that using sticky tabs for grammar sections improves revision speed by 30%, making this a top-rated method among Kathmandu school teachers.</code>	<code>Our guide for Oxford Advanced Learner's Dictionary explains how to use the companion app to check word pronunciations and build vocabulary lists. This digital tool costs NPR 1,500 for a one-year subscription and syncs across phone and laptop. Students in Pokhara have praised its clarity for improving their English speaking skills.</code>	<code>This blog post reviews the top five ballpoint pens under NPR 200, including Linc Pentonic and Hauser XO, focusing on smooth writing and durability. We mention that the Hauser XO offers a 1.2mm tip for bold strokes, ideal for signatures. However, no Nepali textbook usage or teacher feedback is included.</code>	<code>Today's special offer: Buy three A4 notebooks from Classic Notebooks for just NPR 350 and get a free Bic pen set. Available only at our Jawalakhel outlet while stocks last. Visit our store for stationery deals on school supplies.</code>

• Loss: <code>CachedMultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 32,
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": "hard_negatives",
      "hardness_strength": 5
  }
  

</details> <details><summary>cross_lingual</summary>

cross_lingual

• Dataset: cross_lingual at 7efcacf
• Size: 8,000 training samples
• Columns: <code>source_text</code> and <code>text</code>
• Approximate statistics based on the first 100 samples:

  	source_text	text
  type	string	string
  modality	text	text
  details	<ul><li>min: 23 tokens</li><li>mean: 78.92 tokens</li><li>max: 178 tokens</li></ul>	<ul><li>min: 15 tokens</li><li>mean: 71.51 tokens</li><li>max: 184 tokens</li></ul>

• Samples:

  source_text	text
  <code>The Supreme Court of Nepal will deliver its verdict on the controversial 52 appointments to constitutional bodies made through amended ordinances, amidst multiple legal challenges.</code>	<code>Nepal ko Supreme Court le aba controversial 52 appointments ko barema verdict dine wala cha, jo constitutional bodies ma amended ordinances through bhako thiyo. Tara tyo bichama dherai legal challenges pani aairako cha.</code>
  <code>सांसद ज्ञानेन्द्रबहादुर कार्कीले नेपालको विकास मोडलमा छलफल गर्न आवश्यक रहेको र छिमेकी देशहरूको विकास अध्ययन गर्नुपर्ने बताएका छन्।</code>	<code>सांसद ज्ञानेन्द्रबहादुर कार्की कहलन कि नेपाल के विकास मोडल प छलफल जरूरी बा, आ पड़ोसी देसन के विकास के अध्ययन करल जाय के चाही।</code>
  <code>कालीमाटी तरकारी बजारमा आज काउली, गोलभेँडा र भिन्डी लगायतका तरकारीको थोक मूल्य घटेको छ भने अन्य तरकारीहरूका मूल्य स्थिर रहेका छन्।</code>	<code>Kalimati tarkari bazar ma aaj kauli, golbhenda ra bhindi lagayat ka tarkari ko wholesale price ghatya cha bhane aru tarkari haru ka price sthir raheka chan.</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>sanoir_general</summary>

sanoir_general

• Dataset: sanoir_general at 7e53250
• Size: 4,977 training samples
• Columns: <code>query</code>, <code>positive</code>, <code>negative_1</code>, <code>negative_2</code>, <code>negative_3</code>, and <code>negative_4</code>
• Approximate statistics based on the first 100 samples:

  	query	positive	negative_1	negative_2	negative_3	negative_4
  type	string	string	string	string	string	string
  modality	text	text	text	text	text	text
  details	<ul><li>min: 15 tokens</li><li>mean: 58.08 tokens</li><li>max: 160 tokens</li></ul>	<ul><li>min: 48 tokens</li><li>mean: 191.17 tokens</li><li>max: 256 tokens</li></ul>	<ul><li>min: 27 tokens</li><li>mean: 159.71 tokens</li><li>max: 237 tokens</li></ul>	<ul><li>min: 24 tokens</li><li>mean: 139.11 tokens</li><li>max: 213 tokens</li></ul>	<ul><li>min: 21 tokens</li><li>mean: 130.95 tokens</li><li>max: 215 tokens</li></ul>	<ul><li>min: 23 tokens</li><li>mean: 129.66 tokens</li><li>max: 216 tokens</li></ul>

• Samples:

  query	positive	negative_1	negative_2	negative_3	negative_4
  <code>Is it better for Nepal to switch completely to organic farming instead of using chemical fertilizers?</code>	<code>के पूर्ण रूपमा जैविक खेती (Organic Farming) अपनाउनु नै नेपालको कृषि भविष्यका लागि उत्तम विकल्प हो? हो, पूर्ण जैविक खेती अपनाउनु नै सबैभन्दा उत्तम विकल्प हो। रासायनिक मल र कृत्रिम कीटनाशकको प्रयोगले माटोको प्राकृतिक उर्वराशक्ति नष्ट गरिरहेको छ। जैविक खेतीले माटोको जीवसत्त्वलाई जोगाउँछ, वातावरणलाई प्रदूषणमुक्त राख्छ र उपभोक्तालाई स्वस्थ खाद्यान्न उपलब्ध गराउँछ। दीर्घकालीन रूपमा यो पद्धतिले माटोको स्वास्थ्य र कृषकको आम्दानी दुवैलाई सुनिश्चित गर्दछ।</code>	<code>नेपालका साना किसानहरूका लागि रासायनिक मलको प्रयोग पूर्णतया बन्द गर्नु व्यावहारिक छैन। यसले उत्पादनमा तत्काल कमी ल्याउन सक्छ। त्यसैले, रासायनिक मलको मात्रा बिस्तारै घटाउँदै लैजानुपर्छ र प्राङ्गारिक मलको प्रयोगलाई प्रोत्साहन गर्ने नीति ल्याउनुपर्छ।</code>	<code>जैविक खेती गर्दा माटोको गुणस्तर दीर्घकालीन रूपमा सुध्रिन्छ र हानिकारक विषादीको प्रयोग हुँदैन। तर, यसका लागि आवश्यक पर्ने श्रम र समय बढी लाग्ने भएकाले व्यावसायिक खेती गर्ने ठूला कृषकहरूका लागि यो चुनौतीपूर्ण हुन सक्छ।</code>	<code>सरकारले कृषि क्षेत्रमा अनुदान दिँदा केवल रासायनिक मलमा मात्र सीमित नभई जैविक मल र कीटनाशक उत्पादन गर्ने उद्योगहरूलाई पनि प्राथमिकता दिनुपर्छ। यसले गर्दा कृषकहरूले सस्तो दरमा प्राङ्गारिक सामग्री प्राप्त गर्न सक्छन्।</code>	<code>हिमाली भेगका किसानहरूले आफ्नो उत्पादन बढाउन रासायनिक मलको सट्टा स्थानीय रूपमा उपलब्ध हुने कम्पोस्ट मल र गोबर मलको प्रयोगमा बढी ध्यान दिनुपर्छ। यसले गर्दा माटोको उत्पादकत्व बढ्नेछ।</code>
  <code>मुस्ताङ जाने कच्ची बाटो हिउँदमा बर्खाको तुलनामा सुगम हुन्छ कि भन्ने claim लाई support गर्ने forum answer खोज्दैछु, जहाँ landslide र river obstruction कम हुँदा safe travel months अक्टोबरदेखि मार्चसम्म भनेर local guide को consensus छ।</code>	<code>मुस्ताङ जाने कच्ची बाटो हिउँदमा बर्खाको तुलनामा धेरै सुगम हुन्छ किनभने बर्खामा भासिने र खोलामा बगेर आउने पहिरोले मार्ग अवरुद्ध गर्छ भने हिउँदमा जमेको हिउँले तल्लो भूभाग स्थिर राख्छ र गाडी चलाउन मिल्ने अवस्था बन्छ। स्थानीय ट्रेकिङ गाइडहरू पनि अक्टोबरदेखि मार्चसम्मको यात्रालाई सुरक्षित मान्छन्।</code>	<code>हिउँदमा मुस्ताङ जाने बाटो बर्खाको भन्दा सजिलो हुन्छ किनभने त्यहाँको हावा सफा हुन्छ र पर्यटक कम हुन्छन्। तर यो भनेको मात्र मौसमको अनुभव हो, बाटोको भौतिक अवस्था भने सधैं उस्तै रहन्छ।</code>	<code>मुस्ताङको लो मन्थाङबाट जोमसोमसम्मको पैदल मार्ग हिउँदमा बर्खाको तुलनामा बढी जोखिमपूर्ण हुन्छ किनभने त्यहाँ हिउँले मार्ग ढाक्छ र स्थानीय भने यो सिजनमा यात्रा नगर्न सल्लाह दिन्छन्।</code>	<code>मुस्ताङ जाने प्रमुख सडक मार्ग बेनीबाट कागबेनी हुँदै अगाडि जान्छ। बर्खामा खोला बग्ने भएकाले सवारी चलाउन गाह्रो हुन्छ तर हिउँदमा पनि त्यही मार्ग प्रयोग हुन्छ। यसले मौसम अनुसार मर्मत आवश्यक पर्छ।</code>	<code>डोल्पा जिल्लाको दुर्गम भेगमा हिउँदमा बाटो बर्खाको भन्दा बढी सुगम हुन्छ किनभने वर्षाका कारण खसेका ढुङ्गा हटिसकेका हुन्छन् र जनजागरण कार्यक्रमले बाटो सफा राख्छ।</code>
  <code>What is the impact of nitrogen fertilizer usage on maize yields in the Terai region of Nepal?</code>	<code>नेपालको तराई क्षेत्रमा मकैको उत्पादनमा सुधार ल्याउनका लागि नाइट्रोजनको सही मात्रामा प्रयोग महत्त्वपूर्ण मानिन्छ। कृषि विज्ञहरूका अनुसार, माटो परीक्षण गरी तोकिएको मात्रामा मात्र युरिया मल प्रयोग गर्दा मकैको दानाको गुणस्तर र उत्पादन दुवैमा उल्लेख्य वृद्धि हुने गरेको छ। यसले माटोको प्राकृतिक सन्तुलन बिगार्नबाट पनि जोगाउँछ।</code>	<code>नेपालको पहाडी भेगमा धान खेती गर्दासायनिक मलको प्रयोगले उत्पादनमा वृद्धि ल्याएको देखिन्छ। अनुसन्धानले देखाएको छ कि नाइट्रोजनयुक्त मलको उचित प्रयोगले माटोको उर्वरशक्ति कायम राख्न मद्दत गर्छ, जसले गर्दा किसानहरूले कम लागतमा बढी अन्न उत्पादन गर्न सक्छन्।</code>	<code>तराईका क्षेत्रमा गहुँ उत्पादन बढाउन सिँचाइ प्रविधिमा सुधार गर्नु आवश्यक छ। हालैका अध्ययनहरूले देखाए अनुसार आधुनिक सिँचाइ प्रणाली र उन्नत बीउको संयोजनले माटोको गुणस्तर सुधार्नुका साथै उत्पादनको दरलाई पनि स्थिर राख्न सहयोग पुर्‍याउँछ।</code>	<code>जलवायु परिवर्तनका कारण नेपालको कृषि क्षेत्रमा ठूलो चुनौती थपिएको छ। बढ्दो तापक्रम र अनियमित वर्षाले गर्दा परम्परागत बालीनालीको उत्पादनमा ह्रास आउनुका साथै माटोको जैविक संरचनामा पनि नकारात्मक प्रभाव परिरहेको छ।</code>	<code>जैविक खेती (Organic farming) को विस्तारले नेपालको ग्रामीण अर्थतन्त्रलाई बलियो बनाउन सक्छ। रासायनिक विषादीको सट्टा प्राङ्गारिक मलको प्रयोग गर्दा माटोको संरचना मात्र सुध्रिँदैन, यसले उपभोक्ताको स्वास्थ्यलाई पनि प्रत्यक्ष फाइदा पुर्‍याउँछ।</code>

• Loss: <code>CachedMultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 32,
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details>

Evaluation Datasets

<details><summary>title_excerpt</summary>

title_excerpt

• Dataset: title_excerpt at 1582ca1
• Size: 20,560 evaluation samples
• Columns: <code>title</code> and <code>excerpt</code>
• Approximate statistics based on the first 100 samples:

  	title	excerpt
  type	string	string
  modality	text	text
  details	<ul><li>min: 3 tokens</li><li>mean: 28.98 tokens</li><li>max: 58 tokens</li></ul>	<ul><li>min: 31 tokens</li><li>mean: 76.45 tokens</li><li>max: 157 tokens</li></ul>

• Samples:

  title	excerpt
  <code>UML Youth Protest in Biratnagar Against Party Chairman's Arrest</code>	<code>CPN-UML youths in Biratnagar protested against the arrest of Party Chairman KP Sharma Oli, gathering at Mahendra Chowk and chanting slogans against the Home Minister.</code>
  <code>Special Court extends UML Vice-chair Bishnu Paudel's remand by three days</code>	<code>The Special Court has extended the remand of former Finance Minister and CPN-UML Vice-chair Bishnu Prasad Paudel by three days amid an ongoing money laundering investigation. Meanwhile, the Supreme Court is hearing a habeas corpus petition challenging his detention.</code>
  <code>अनावश्यक रूपमा बिरामीलाई ‘रिफर’ नगर्न अस्पताललाई स्वास्थ्य बीमा बोर्डको निर्देशन</code>	<code>स्वास्थ्य बिमा बोर्डले सबै स्वास्थ्य संस्थालाई बिरामीलाई अनावश्यक रिफर नगर्न निर्देशन दिएको छ। अनावश्यक रिफर हुँदा बिरामीलाई समस्या र अतिरिक्त परीक्षण हुन सक्छन्।</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>ne_en</summary>

ne_en

• Dataset: ne_en at e4c18f5
• Size: 753 evaluation samples
• Columns: <code>title</code> and <code>translation</code>
• Approximate statistics based on the first 100 samples:

  	title	translation
  type	string	string
  modality	text	text
  details	<ul><li>min: 6 tokens</li><li>mean: 32.12 tokens</li><li>max: 72 tokens</li></ul>	<ul><li>min: 3 tokens</li><li>mean: 23.36 tokens</li><li>max: 75 tokens</li></ul>

• Samples:

  title	translation
  <code>Sudurpaschim province sees sharp incline in Covid-19 infection</code>	<code>सुदूरपश्चिम प्रदेशमा कोभिड–१९ को संक्रमण बढेको छ</code>
  <code>खालि हुन थाल्यो पनिका, मुख्य सेवा अनलाइनबाटै</code>	<code>Even though it is empty, the main service is online</code>
  <code>Nepal-UK Tech Forum highlights digital investment opportunities</code>	<code>नेपाल-युके टेक फोरमले डिजिटल लगानीका अवसरहरू हाइलाइट गरेको छ</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>excerpt_paraphrase</summary>

excerpt_paraphrase

• Dataset: excerpt_paraphrase at b521e5a
• Size: 110 evaluation samples
• Columns: <code>sentence1</code> and <code>sentence2</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2
  type	string	string
  modality	text	text
  details	<ul><li>min: 27 tokens</li><li>mean: 71.64 tokens</li><li>max: 174 tokens</li></ul>	<ul><li>min: 31 tokens</li><li>mean: 77.55 tokens</li><li>max: 195 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Khanikhola Hydropower has received approval from SEBON to issue rights shares worth Rs 465.7 million, doubling its paid-up capital to Rs 931.4 million.</code>	<code>SEBON has granted approval to Khanikhola Hydropower for the issuance of rights shares valued at Rs 465.7 million, a move that will double its paid-up capital to Rs 931.4 million.</code>
  <code>अन्तर्राष्ट्रिय आणविक ऊर्जा एजेन्सीले इरानमा अमेरिकाको आक्रमणपछि विकिरण स्तर नबढेको पुष्टि गरेको छ।</code>	<code>अमेरिकाले इरानमा आक्रमण गरेपछि अन्तर्राष्ट्रिय आणविक ऊर्जा एजेन्सीले त्यहाँ विकिरणको स्तर नबढेको कुराको पुष्टि गरेको छ।</code>
  <code>EU foreign affairs chief Kaja Kallas urged China to stop undermining European security amid ongoing tensions over trade practices and Chinese support to Russia during the Ukraine war.</code>	<code>Kaja Kallas, the European Union's top diplomat, called on China to cease jeopardizing Europe's safety due to persistent disagreements concerning commercial activities and Beijing's backing of Moscow throughout the conflict in Ukraine.</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc",
          "doc_to_query"
      ],
      "partition_mode": "per_direction",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>nepali_triplets</summary>

nepali_triplets

• Dataset: nepali_triplets at bd8c184
• Size: 200 evaluation samples
• Columns: <code>sentence</code>, <code>positive_sentence</code>, and <code>negative_sentence</code>
• Approximate statistics based on the first 100 samples:

  	sentence	positive_sentence	negative_sentence
  type	string	string	string
  modality	text	text	text
  details	<ul><li>min: 26 tokens</li><li>mean: 79.94 tokens</li><li>max: 178 tokens</li></ul>	<ul><li>min: 25 tokens</li><li>mean: 81.61 tokens</li><li>max: 186 tokens</li></ul>	<ul><li>min: 25 tokens</li><li>mean: 62.1 tokens</li><li>max: 134 tokens</li></ul>

• Samples:

  sentence	positive_sentence	negative_sentence
  <code>रसुवामा अस्थायी प्रहरी चौकी मैलुङबाट हत्कडीसहित फरार भएका दुई थुनुवा चार घण्टाभित्रै सिम्लेबाट पक्राउ परे। उनीहरूलाई निर्माणाधीन त्रिशूली २१६ हाइड्रोपावरबाट चोरी गरेको सामानसहित पक्राउ गरिएको थियो।</code>	<code>रसुवाको सिम्लेबाट चार घण्टाभित्रै दुई थुनुवाहरूलाई हत्कडी लगाएर फरार भएको मैलुङको अस्थायी प्रहरी चौकीले पक्राउ गरेको छ, जसले त्रिशूली २१६ हाइड्रोपावर निर्माण स्थलबाट चोरी गरेका सामानहरू पनि बरामद गरियो।</code>	<code>रसुवामा अस्थायी प्रहरी चौकी मैलुङबाट फरार भएका दुई थुनुवाहरू अझै पक्राउ परेका छैनन् र उनीहरूलाई चोरीको आरोपमा अनुसन्धान जारी छ।</code>
  <code>The Forest Development Fund in Nepal, amounting to Rs 234.22 million, has remained unused for six years despite being established for forest conservation and development. The Department of Forests and Soil Conservation aims to utilize the fund once procedures are approved, with compensation from forest land use and infrastructure projects expected to replenish it.</code>	<code>Although the Forest Development Fund in Nepal, totaling Rs 234.22 million, was created to support forest conservation and development, it has not been spent for six years; the Department of Forests and Soil Conservation plans to activate its use following procedural approvals, anticipating that contributions from forest land utilization and infrastructure initiatives will restore the fund.</code>	<code>The Forest Development Fund in Nepal, amounting to Rs 234.22 million, has been fully allocated to infrastructure projects over the past six years, leaving no resources for forest conservation efforts.</code>
  <code>रसुवागढीस्थित भोटेकोसी नदीमा आएको बाढीमा २० जना बेपत्ता भएका थिए जसमा एक प्रहरी सहायक निरीक्षक लालबहादुर श्रेष्ठको मात्रै शव पहिचान भएको छ। ५ शव र ४ मानव अंग त्रिवि शिक्षण अस्पतालमा राखिएका छन्।</code>	<code>भोटेकोसी नदीमा आएको बाढीका कारण रसुवागढीमा २० जना हराइरहेका थिए, जसमा प्रहरी सहायक निरीक्षक लालबहादुर श्रेष्ठको मात्र शव पुष्टि भएको छ भने पाँच शव र चार मानव अंग त्रिवि शिक्षण अस्पतालमा सुरक्षित राखिएका छन्।</code>	<code>रसुवागढीस्थित भोटेकोसी नदीमा आएको बाढीले धेरै घरहरू भत्काएको छ, तर कुनै पनि व्यक्ति बेपत्ता भएको खबर आएको छैन।</code>

• Loss: <code>TripletLoss</code> with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.2
  }
  

</details> <details><summary>stsb_en</summary>

stsb_en

• Dataset: stsb_en at 4bfcc74
• Size: 1,498 evaluation samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	score
  type	string	string	float
  modality	text	text
  details	<ul><li>min: 7 tokens</li><li>mean: 10.04 tokens</li><li>max: 20 tokens</li></ul>	<ul><li>min: 7 tokens</li><li>mean: 9.96 tokens</li><li>max: 18 tokens</li></ul>	<ul><li>min: 0.0</li><li>mean: 0.53</li><li>max: 1.0</li></ul>

• Samples:

  sentence1	sentence2	score
  <code>A man with a hard hat is dancing.</code>	<code>A man wearing a hard hat is dancing.</code>	<code>1.0</code>
  <code>A young child is riding a horse.</code>	<code>A child is riding a horse.</code>	<code>0.95</code>
  <code>A man is feeding a mouse to a snake.</code>	<code>The man is feeding a mouse to the snake.</code>	<code>1.0</code>

• Loss: <code>CoSENTLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "pairwise_cos_sim"
  }
  

</details> <details><summary>stsb_ne</summary>

stsb_ne

• Dataset: stsb_ne at 4bfcc74
• Size: 1,498 evaluation samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	score
  type	string	string	float
  modality	text	text
  details	<ul><li>min: 7 tokens</li><li>mean: 21.72 tokens</li><li>max: 52 tokens</li></ul>	<ul><li>min: 11 tokens</li><li>mean: 21.3 tokens</li><li>max: 35 tokens</li></ul>	<ul><li>min: 0.0</li><li>mean: 0.53</li><li>max: 1.0</li></ul>

• Samples:

  sentence1	sentence2	score
  <code>कडा टोपी लगाएको मान्छे नाचिरहेको छ।</code>	<code>कडा टोपी लगाएको मान्छे नाचिरहेको छ।</code>	<code>1.0</code>
  <code>एउटा सानो बालक घोडा चढिरहेको छ।</code>	<code>एउटा बच्चा घोडा चढिरहेको छ।</code>	<code>0.95</code>
  <code>एउटा मानिसले मुसालाई सर्पलाई खुवाइरहेको छ।</code>	<code>मानिसले मुसालाई सर्पलाई खुवाइरहेको छ।</code>	<code>1.0</code>

• Loss: <code>CoSENTLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "pairwise_cos_sim"
  }
  

</details> <details><summary>nepali_qa_9k</summary>

nepali_qa_9k

• Dataset: nepali_qa_9k at 33f9def
• Size: 720 evaluation samples
• Columns: <code>question</code> and <code>answer</code>
• Approximate statistics based on the first 100 samples:

  	question	answer
  type	string	string
  modality	text	text
  details	<ul><li>min: 28 tokens</li><li>mean: 46.5 tokens</li><li>max: 74 tokens</li></ul>	<ul><li>min: 57 tokens</li><li>mean: 101.14 tokens</li><li>max: 158 tokens</li></ul>

• Samples:

  question	answer
  <code>मले गाडी चलाउँदा ट्राफिक नियम उल्लंघन गरेर जरिवाना पाएँ, त्यसलाई कसरी तिर्न सक्छु?</code>	<code>तपाईंले सरकारी वेबसाइट वा आधिकारिक यातायात कार्यालयको पोर्टल मार्फत आफ्नो ट्राफिक जरिवानाको विवरण जाँच गर्न र तिर्न सक्नुहुन्छ।</code>
  <code>मेरो बच्चाको लागि COVID-19 को खोप कहिले लगाउनु पर्छ? के यो अनिवार्य छ?</code>	<code>बच्चाहरूको लागि खोपको समय बाल स्वास्थ्य विभागको नवीनतम निर्देशन अनुसार जाँच गर्नुपर्छ। कृपया स्थानीय स्वास्थ्य कार्यालय वा आधिकारिक सरकारी वेबसाइटमा हालको खोप कार्यक्रमको जानकारी लिनुहोस्।</code>
  <code>जन्मदर्ता गराउनका लागि आवश्यक कागजातहरू के-के हुन्?</code>	<code>जन्मदर्ताका लागि सामान्यतया जन्मदर्ताको फारम, अभिभावकको नागरिकता/पहचान पत्रको प्रतिलिपि, र जन्मस्थान तथा जन्ममिति प्रमाणित गर्ने अन्य आवश्यक कागजातहरू चाहिन्छन्।</code>

• Loss: <code>CachedMultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 32,
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

</details> <details><summary>ecommerce</summary>

ecommerce

• Dataset: ecommerce at 4b7dd81
• Size: 198 evaluation samples
• Columns: <code>query</code>, <code>document</code>, <code>negative1</code>, <code>negative2</code>, and <code>negative3</code>
• Approximate statistics based on the first 100 samples:

  	query	document	negative1	negative2	negative3
  type	string	string	string	string	string
  modality	text	text	text	text	text
  details	<ul><li>min: 8 tokens</li><li>mean: 26.75 tokens</li><li>max: 65 tokens</li></ul>	<ul><li>min: 54 tokens</li><li>mean: 94.6 tokens</li><li>max: 154 tokens</li></ul>	<ul><li>min: 41 tokens</li><li>mean: 66.27 tokens</li><li>max: 103 tokens</li></ul>	<ul><li>min: 31 tokens</li><li>mean: 60.39 tokens</li><li>max: 97 tokens</li></ul>	<ul><li>min: 37 tokens</li><li>mean: 55.79 tokens</li><li>max: 88 tokens</li></ul>

• Samples:

  query	document	negative1	negative2	negative3
  <code>कुन पश्मिना राम्रो छ, हिमालयन ट्रेजरको कि पश्मिना प्यालेसको?</code>	<code>I bought the 'Hatti Dhundo' black pashmina shawl from Himalayan Treasure for NPR 3,200 and it was incredibly soft and warm. However, my friend got a similar pashmina from Pashmina Palace for NPR 2,800, but hers felt a bit thinner and less fluffy. The Himalayan Treasure one had better stitching and a more elegant finish, perfect for winter events. I'm glad I paid a bit more for the quality.</code>	<code>The 'Hatti Dhundo' pashmina from Himalayan Treasure is made from pure cashmere wool and comes in various colors. It is machine washable on a gentle cycle and dries quickly without shrinking. Many customers have praised its durability over multiple seasons.</code>	<code>I ordered a traditional Nepali dhaka topi from Kathmandu Handicrafts for NPR 500. The fabric was high-quality and the embroidery was precise, but it was a bit tight for my head size. They offered free exchanges, which was convenient.</code>	<code>These running shoes from Nike are lightweight and have great arch support. I wear them for my daily jog at Ratna Park and they've lasted over six months. The cushioning is still as good as new.</code>
  <code>Prestige Omega electric cooker specs kati watt ko cha ra kati liter ko ho?</code>	<code>The Prestige Omega 4-in-1 Electric Cooker (model: POC-45) features a 5-liter capacity stainless steel inner pot, 900W power consumption, and multi-function capabilities including rice cooking, slow cooking, steaming, and sautéing. It comes with a 24-hour digital timer, keep-warm function, and a non-stick inner lid for easy cleaning. Priced at NPR 8,500, it is available with a 1-year warranty in Nepal.</code>	<code>The Prestige Omega Induction Cooktop (model: PIC-2000) has a 2000W power output, 8 preset cooking modes, and a digital LED display with touch controls. It is compatible with all types of cookware and comes with a 2-year warranty. Price is NPR 5,200.</code>	<code>The Philips Airfryer XXL (model: HD9650/90) uses 2225W of power, has a 3-pound capacity basket, and features fat reduction technology for healthier frying. It includes 11 preset cooking programs and a keep-warm function. Available for NPR 22,000.</code>	<code>The IKEA KALLAX Shelf Unit (dimensions: 147x147 cm) is a versatile storage solution made from particleboard with a foil finish. It comes in black-brown and white colors and can be used as a room divider or bookshelf. Price is NPR 9,500.</code>
  <code>Khadi Natural Vitamin C Serum कसरी use गर्ने?</code>	<code>To get the best results from your Khadi Natural Vitamin C Face Serum, apply 3-4 drops to a clean, damp face every morning after cleansing. Gently massage it into your skin using upward motions, avoiding the eye area, and let it absorb for a minute before layering with a moisturizer. For Nepali users, this ₹450 (approx. NPR 720) serum is ideal for brightening dull skin and reducing hyperpigmentation, especially in Kathmandu's polluted environment. Always use sunscreen during the day, as Vitamin C can increase sun sensitivity.</code>	<code>The Khadi Natural Vitamin C Face Serum contains 20% stabilized Vitamin C, hyaluronic acid, and aloe vera, making it a potent antioxidant formula for all skin types. It is available in a 30ml glass bottle with a dropper and is free from parabens and sulfates. This serum is manufactured in India and has a shelf life of 24 months from the date of manufacture.</code>	<code>Are you looking for the best Vitamin C serums in Nepal? Check out our top picks including The Ordinary, Minimalist, and Khadi Natural, with prices ranging from NPR 800 to NPR 3,500. We also have a buying guide for Nepali skin types, including tips for oily and acne-prone skin. Order now for free delivery in Kathmandu and Lalitpur.</code>	<code>Get glowing skin with our new range of organic face masks made from Nepali turmeric and local honey. Each mask is handcrafted in Pokhara and costs only NPR 500, with a 100% money-back guarantee. These masks are perfect for a weekly pampering session and come in eco-friendly packaging.</code>

• Loss: <code>CachedMultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 32,
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": "hard_negatives",
      "hardness_strength": 5
  }
  

</details> <details><summary>cross_lingual</summary>

cross_lingual

• Dataset: cross_lingual at 7efcacf
• Size: 1,600 evaluation samples
• Columns: <code>source_text</code> and <code>text</code>
• Approximate statistics based on the first 100 samples:

  	source_text	text
  type	string	string
  modality	text	text
  details	<ul><li>min: 30 tokens</li><li>mean: 78.76 tokens</li><li>max: 176 tokens</li></ul>	<ul><li>min: 22 tokens</li><li>mean: 72.65 tokens</li><li>max: 175 tokens</li></ul>

• Samples: | source_text | text |

The model card is truncated. Read the rest on Hugging Face.