LateOn-Code
The
LateOn-Code collection is composed of
PyLate models optimized for code retrieval. These late interaction models are first pre-trained following the methodology of
CoRNStack . These pre-trained models are then further fine-tuned on train sets of CoIR using the
nv-retriever methodology to mine hard negatives while preventing false negatives.
We started from the two best ColBERT models on the BEIR benchmark for their respective sizes. The first one,
LateOn-Code is based on in-house LateOn model, a new version of
GTE-ModernColBERT-v1 built on ModernBERT-base (also developed at LightOn). This version underwent significantly deeper training, crossing the 57 mark on BEIR, almost a 2.5-point improvement and is thus SOTA by a large margin. We'll release this base model along with training data and boilerplates in the near future, so stay tuned! The second,
LateOn-Code-edge is a smaller model based on the
edge-colbert model family from mixedbread , using the
smallest variant (Ettin-17M) for maximum efficiency. For more details on the training setup, please refer to our
blogpost .
The original
CoRNStack data in a format compatible with PyLate can be found
here while the fine-tuning data can be found
here . Training boilerplates can be found
here in the PyLate repository MTEB (Code, v1) benchmark results
Pre-trained models achieve very competitive results as the 17M model outperforms the very strong granite-embedding-small-english-r2 by an average of 1.7. This is truly impressive, as the granite model is almost three times bigger (17M vs 48M), but is also a beast on its own in the <100M parameters range. It also outperforms the larger granite variant (149M). The larger version nicely scales by improving over the performance of its little sibling by 6.5 on average.
Although the pre-training results are already very impressive given that they are mostly out-of-domain, running a proper fine-tuning using the training data of CoIR significantly boost the performance of the models. Notably, the 17M model increases from 57.50 to 66.64 (+9.14), getting pretty close to EmbeddingGemma-300M while being 17 times smaller. The larger one increases from 63.77 to 74.12 (+10.35), strongly outperforming EmbeddingGemma-300M and getting closer to strong LLM models such as Qwen3-Embedding-0.6B and C2LLM-0.5B while being much smaller.
Model Params Type Avg Apps COIR CSNet CodeEdit CodeFB MT CodeFB ST CSNet CC CSNet CodeTrans Contest CodeTrans DL CosQA StackOF QA Synth T2SQL Baseline BM25 - Lexical 44.41 4.76 40.86 49.85 59.19 68.15 53.97 60.01 47.78 34.42 18.75 70.26 24.94 Small (≤50M) granite-embedding-small-english-r2 47M Single vector 55.84 13.54 60.46 57.16 52.19 76.85 48.42 78.28 77.63 33.63 35.58 90.04 46.33 LateOn-Code-edge-pretrain 17M Multi vector 57.50 10.81 73.78 62.07 51.92 76.65 63.22 88.03 71.31 33.16 30.53 74.63 53.83 LateOn-Code-edge 17M Multi vector 66.64 26.22 81.60 62.21 74.25 87.12 79.26 87.85 75.36 37.08 40.54 85.63 62.57 Δ (fine-tune - pretrain) +9.14 +15.41 +7.82 +0.14 +22.33 +10.47 +16.04 -0.18 +4.05 +3.92 +10.01 +11.00 +8.74 Medium (100M–300M) granite-embedding-english-r2 149M Single vector 57.22 13.96 64.65 59.35 52.54 77.18 47.67 80.79 77.07 35.03 37.01 91.80 49.55 CodeRankEmbed 137M Single vector 60.47 23.45 83.20 59.98 42.61 78.10 68.89 89.50 66.43 34.49 35.17 80.53 63.27 GTE-ModernBERT 149M Single vector 71.66 57.72 83.10 55.83 86.15 86.00 93.61 88.76 72.35 37.27 43.36 91.14 64.61 embeddinggemma-300m 300M Single vector 68.76 84.39 75.54 62.10 51.42 80.26 73.71 90.15 85.51 33.52 43.60 86.47 58.42 LateOn-Code-pretrain 149M Multi vector 63.77 23.09 80.27 68.74 50.21 82.66 71.47 91.05 82.20 34.46 34.15 85.61 61.34 LateOn-Code 149M Multi vector 74.12 54.76 86.57 64.99 82.22 90.40 89.32 90.40 87.44 41.00 45.23 93.43 63.67 Δ (fine-tune - pretrain) +10.35 +31.67 +6.30 -3.75 +32.01 +7.74 +17.85 -0.65 +5.24 +6.54 +11.08 +7.82 +2.33 Large (≥500M) C2LLM-0.5B 500M Single vector 75.46 61.02 86.71 71.39 92.29 88.63 96.29 89.20 84.27 33.99 38.30 89.40 74.08 Qwen3-Embedding-0.6B 600M Single vector 75.42 75.34 84.69 64.42 90.82 86.39 91.72 91.01 86.05 31.36 36.48 89.99 76.74
Best result across all sizes is underlined . Best within each size category is bolded .
Colgrep
The LateOn-Code family model can easily be used within ColGrep, an easy-to-use search tool that give their powerful search capabilities to coding agent. It has been designed to extend grep capabilities to get the best of both world and is very effective to enhance the quality of the answer while diminishing answer time and tokens consumption. Given the performance of the very light-weight 17M model, it can easily run quickly on any computer.
Install ColGrep
1 # macOS / Linux
2 curl --proto '=https' --tlsv1.2 -LsSf https://github.com/lightonai/next-plaid/releases/latest/download/colgrep-installer.sh | sh
3
4 # Windows (PowerShell)
5 powershell -c "irm https://github.com/lightonai/next-plaid/releases/latest/download/colgrep-installer.ps1 | iex"
Search
1 # Semantic search — find code by meaning
2 colgrep "function that retries HTTP requests"
3
4 # Regex search
5 colgrep -e "async fn\s+\w+"
6
7 # Hybrid — regex narrows candidates, semantics ranks them
8 colgrep -e "Result<" "error handling" --include = "*.rs"
Install for Claude Code
colgrep --install-claude-codeChoose a Model
1 # Set the model
2 colgrep set-model lightonai/LateOn-Code # default: lightonai/LateOn-Code-edge
For more information about ColGrep, please refer to the
official documentation Model
This is a multi-vector (ColBERT-style late interaction) embedding model finetuned from
lightonai/LateOn-Code-edge-pretrain on the
apps ,
synthetictext2sql ,
cosqa ,
codefeedbackst ,
codefeedbackmt ,
stackoverflowqa ,
codetranscontest ,
codetransdl ,
CodeSearchNet_go ,
CodeSearchNet_java ,
CodeSearchNet_javascript ,
CodeSearchNet_php ,
CodeSearchNet_python ,
CodeSearchNet_ruby ,
CodeSearchNet_ccr_go ,
CodeSearchNet_ccr_java ,
CodeSearchNet_ccr_javascript ,
CodeSearchNet_ccr_php ,
CodeSearchNet_ccr_python and
CodeSearchNet_ccr_ruby datasets. It maps sentences & paragraphs to sequences of 48-dimensional dense vectors and can be used for semantic textual similarity using the MaxSim operator.
Model Details
Model Description
Model Type: Multi-vector embedding model
Document Length: 2048 tokensQuery Length: 256 tokensOutput Dimensionality: 48 tokensSimilarity Function: MaxSimTraining Datasets:
Language: English, codeLicense: Apache 2.0
Model Sources
Documentation: PyLate Documentation Repository: PyLate on GitHub Hugging Face: PyLate models on Hugging Face
Full Model Architecture
ColBERT(
(0): Transformer({'max_seq_length': 2047, 'do_lower_case': True, 'architecture': 'ModernBertModel'})
(1): Dense({'in_features': 256, 'out_features': 512, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity', 'use_residual': False})
(2): Dense({'in_features': 512, 'out_features': 48, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity', 'use_residual': False})
)Usage
Sentence Transformers
This model can be used with
Sentence Transformers as a multi-vector (ColBERT-style late interaction) retriever via the
MultiVectorEncoder:
pip install "sentence-transformers>=6.0.0"1 from sentence_transformers import MultiVectorEncoder
2
3 model = MultiVectorEncoder ( "lightonai/LateOn-Code-edge" )
4
5 query = "Which planet is known as the Red Planet?"
6 documents = [
7 "Venus is often called Earth's twin because of its similar size and proximity." ,
8 "Mars, known for its reddish appearance, is often referred to as the Red Planet." ,
9 "Jupiter, the largest planet in our solar system, has a prominent red spot." ,
10 "Saturn, famous for its rings, is sometimes mistaken for the Red Planet." ,
11 ]
12
13 query_embeddings = model . encode_query ( query )
14 document_embeddings = model . encode_document ( documents )
15 print ( query_embeddings . shape , document_embeddings [ 0 ] . shape )
16 # (12, 48) (18, 48)
17
18 # MaxSim late-interaction scoring (higher is more relevant)
19 scores = model . similarity ( query_embeddings , document_embeddings )
20 print ( scores )
21 # tensor([[4.7185, 7.8521, 7.0537, 7.3523]])
PyLate
First install the PyLate library:
Retrieval
Use this model with PyLate to index and retrieve documents. The index uses
FastPLAID for efficient similarity search.
Indexing documents
Load the ColBERT model and initialize the PLAID index, then encode and index your documents:
1 from pylate import indexes , models , retrieve
2
3 # Step 1: Load the ColBERT model
4 model = models . ColBERT (
5 model_name_or_path = "pylate_model_id" ,
6 )
7
8 # Step 2: Initialize the PLAID index
9 index = indexes . PLAID (
10 index_folder = "pylate-index" ,
11 index_name = "index" ,
12 override = True , # This overwrites the existing index if any
13 )
14
15 # Step 3: Encode the documents
16 documents_ids = [ "1" , "2" , "3" ]
17 documents = [ "document 1 text" , "document 2 text" , "document 3 text" ]
18
19 documents_embeddings = model . encode (
20 documents ,
21 batch_size = 32 ,
22 is_query = False , # Ensure that it is set to False to indicate that these are documents, not queries
23 show_progress_bar = True ,
24 )
25
26 # Step 4: Add document embeddings to the index by providing embeddings and corresponding ids
27 index . add_documents (
28 documents_ids = documents_ids ,
29 documents_embeddings = documents_embeddings ,
30 )
Note that you do not have to recreate the index and encode the documents every time. Once you have created an index and added the documents, you can re-use the index later by loading it:
1 # To load an index, simply instantiate it with the correct folder/name and without overriding it
2 index = indexes . PLAID (
3 index_folder = "pylate-index" ,
4 index_name = "index" ,
5 )
Retrieving top-k documents for queries
Once the documents are indexed, you can retrieve the top-k most relevant documents for a given set of queries.
To do so, initialize the ColBERT retriever with the index you want to search in, encode the queries and then retrieve the top-k documents to get the top matches ids and relevance scores:
1 # Step 1: Initialize the ColBERT retriever
2 retriever = retrieve . ColBERT ( index = index )
3
4 # Step 2: Encode the queries
5 queries_embeddings = model . encode (
6 [ "query for document 3" , "query for document 1" ] ,
7 batch_size = 32 ,
8 is_query = True , # # Ensure that it is set to False to indicate that these are queries
9 show_progress_bar = True ,
10 )
11
12 # Step 3: Retrieve top-k documents
13 scores = retriever . retrieve (
14 queries_embeddings = queries_embeddings ,
15 k = 10 , # Retrieve the top 10 matches for each query
16 )
Reranking
If you only want to use the ColBERT model to perform reranking on top of your first-stage retrieval pipeline without building an index, you can simply use rank function and pass the queries and documents to rerank:
1 from pylate import rank , models
2
3 queries = [
4 "query A" ,
5 "query B" ,
6 ]
7
8 documents = [
9 [ "document A" , "document B" ] ,
10 [ "document 1" , "document C" , "document B" ] ,
11 ]
12
13 documents_ids = [
14 [ 1 , 2 ] ,
15 [ 1 , 3 , 2 ] ,
16 ]
17
18 model = models . ColBERT (
19 model_name_or_path = "pylate_model_id" ,
20 )
21
22 queries_embeddings = model . encode (
23 queries ,
24 is_query = True ,
25 )
26
27 documents_embeddings = model . encode (
28 documents ,
29 is_query = False ,
30 )
31
32 reranked_documents = rank . rerank (
33 documents_ids = documents_ids ,
34 queries_embeddings = queries_embeddings ,
35 documents_embeddings = documents_embeddings ,
36 )
Evaluation
Metrics
Py Late Information Retrieval
Dataset: ['CodeSearchNetPython', 'CodeSearchNetJavascript', 'CodeSearchNetGo', 'CodeSearchNetRuby', 'CodeSearchNetJava', 'CodeSearchNetPhp']
Evaluated with pylate.evaluation.pylate_information_retrieval_evaluator.PyLateInformationRetrievalEvaluator
Metric CodeSearchNetPython CodeSearchNetJavascript CodeSearchNetGo CodeSearchNetRuby CodeSearchNetJava CodeSearchNetPhp MaxSim_accuracy@1 0.855 0.707 0.92 0.737 0.755 0.802 MaxSim_accuracy@3 0.958 0.815 0.978 0.87 0.914 0.91 MaxSim_accuracy@5 0.972 0.845 0.987 0.899 0.937 0.932 MaxSim_accuracy@10 0.98 0.877 0.991 0.921 0.951 0.953 MaxSim_precision@1 0.855 0.707 0.92 0.737 0.755 0.802 MaxSim_precision@3 0.3193 0.2717 0.326 0.29 0.3047 0.3033 MaxSim_precision@5 0.1944 0.169 0.1974 0.1798 0.1874 0.1864 MaxSim_precision@10 0.098 0.0877 0.0991 0.0921 0.0951 0.0953 MaxSim_recall@1 0.855 0.707 0.92 0.737 0.755 0.802 MaxSim_recall@3 0.958 0.815 0.978 0.87 0.914 0.91 MaxSim_recall@5 0.972 0.845 0.987 0.899 0.937 0.932 MaxSim_recall@10 0.98 0.877 0.991 0.921 0.951 0.953 MaxSim_ndcg@10 0.9244 0.7937 0.9607 0.8357 0.8655 0.8824 MaxSim_mrr@10 0.9058 0.7668 0.9505 0.8076 0.8367 0.8592 MaxSim_map@100 0.9064 0.7696 0.9508 0.8095 0.8379 0.86
Code Search Network
Dataset: CodeSearchNet_mean
Evaluated with pylate.evaluation.code_search_network_evaluator.CodeSearchNetworkEvaluator
Metric Value MaxSim_accuracy@1 0.796 MaxSim_accuracy@3 0.9075 MaxSim_accuracy@5 0.9287 MaxSim_accuracy@10 0.9455 MaxSim_precision@1 0.796 MaxSim_precision@3 0.3025 MaxSim_precision@5 0.1857 MaxSim_precision@10 0.0946 MaxSim_recall@1 0.796 MaxSim_recall@3 0.9075 MaxSim_recall@5 0.9287 MaxSim_recall@10 0.9455 MaxSim_ndcg@10 0.8771 MaxSim_mrr@10 0.8544 MaxSim_map@100 0.8557
Training Details
Training Datasets
apps
Dataset: apps at 68d15dc
Size: 4,985 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'Polycarp has $n$ different binary words. A word called binary if it contains only charact...{'document': "for _ in range(int(input())):\n n = int(input())\n mass = []\n zo = 0\n oz...{'document': "t=int(input())\nfor _ in range(t):\n n=int(input())\n l=list(map(int,input().split()))...
Loss: pylate.losses.contrastive.Contrastive
synthetictext2sql
Dataset: synthetictext2sql at 68d15dc
Size: 99,996 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'What is the total volume of timber sold by each salesperson, sorted by salesperson?', 'qu...{'document': 'SELECT salesperson_id, name, SUM(volume) as total_volume FROM timber_sales JOIN salesp...{'document': 'SELECT salesperson_id, SUM(volume) as total_volume FROM timber_sales JOIN salesperson ...
Loss: pylate.losses.contrastive.Contrastive
cosqa
Dataset: cosqa at 68d15dc
Size: 9,018 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': '1d array in char datatype in python', 'query_id': 9}{'document': 'def _convert_to_array(array_like, dtype):\n """\n Convert Matrix attribu...{'document': 'def astype(array, y):\n """A functional form of the astype method.\n\n Args:\n ...
Loss: pylate.losses.contrastive.Contrastive
codefeedbackst
Dataset: codefeedbackst at 68d15dc
Size: 125,124 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'You are tasked with implementing a Python class that extends a base class and overrides i...{'document': '```python\nclass TestsslFinding(VSFinding):\n def process_finding(self, finding):\n...{'document': '```python\nfrom googlecloudsdk.calliope import base\nfrom googlecloudsdk.api_lib.sql i...
Loss: pylate.losses.contrastive.Contrastive
codefeedbackmt
Dataset: codefeedbackmt at 68d15dc
Size: 52,941 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': "'user': Embark on a comprehensive journey through the intricate realm of quantum computin...{'document': "Regrettably, there are no standard Python libraries available for quantum computing th...{'document': "The provided code block constructs a quantum circuit with a Hadamard gate (which allow...
Loss: pylate.losses.contrastive.Contrastive
stackoverflowqa
Dataset: stackoverflowqa at 68d15dc
Size: 13,934 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'sphinxsearch-0.9 in mediawiki-1.32.0 error 2019/01/14 12:04:51 [error] 21549#21549: *3558...{'document': 'The SearchDatabase class that SphinxSearch extends was changed from REL1_31 to REL1_32...{'document': 'I was running MediaWiki 1.16.0. I upgraded to MediaWiki 1.16.2 and this resolved the ...
Loss: pylate.losses.contrastive.Contrastive
codetranscontest
Dataset: codetranscontest at 68d15dc
Size: 561 training samples
Approximate statistics based on the first 561 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'Julia set from __future__ import division\n\ncX = -0.7\ncY = 0.27015\nmaxIter = 300\n\nde...{'document': '#include <windows.h>\n#include <string>\n#include <complex>\n\nconst int BMP_SIZE = 60...{'document': '#include <windows.h>\n#include <ctime>\n#include <string>\n\nconst int BMP_SIZE = 600,...
Loss: pylate.losses.contrastive.Contrastive
codetransdl
Dataset: codetransdl at 68d15dc
Size: 564 training samples
Approximate statistics based on the first 564 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'x = tf.range(12)\ntf.size(x)\nX = tf.reshape(x, (3, 4))\ntf.zeros((2, 3, 4))\ntf.ones((2,...{'document': "x = paddle.arange(12)\nx.numel()\nX = paddle.reshape(x, (3, 4))\npaddle.zeros((2, 3, 4...{'document': 'x = torch.arange(12)\nx.numel()\nX = x.reshape(3, 4)\ntorch.zeros((2, 3, 4))\ntorch.on...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_go
Dataset: CodeSearchNet_go at 9f89bdc
Size: 166,972 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'getStringValue func getStringValue(b []rune) (int, error) {\n\tif b[0] != \'"\' {\n\t\tre...{'document': '// getStringValue will return a quoted string and the amount\n// of bytes read\n//\n//...{'document': '// stringValue returns the string value of string literal e.', 'document_id': 18454}
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_java
Dataset: CodeSearchNet_java at 9f89bdc
Size: 162,773 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'SCryptUtil.check public static boolean check(String passwd, String hashed) {\n try...{'document': 'Compare the supplied plaintext password to a hashed password.\n\n@param passwd Plai...{'document': 'Compute the the hash value for the String.\n\n@param passwd\nthe password String\n@ret...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_javascript
Dataset: CodeSearchNet_javascript at 9f89bdc
Size: 56,734 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'function (state, action) {\n return _.defaults({\n isValidating: action.isValidat...{'document': 'Update is validating result\n@param {State} state - state to update\n@param {Action} a...{'document': 'Updates state with newsletter settings submit error\nHolds information only for latest...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_php
Dataset: CodeSearchNet_php at 9f89bdc
Size: 240,327 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'BreadcrumbCollection.addOne public function addOne($title, $url, array $data = [])\n {...{'document': 'Add a breadcrumb item to collection.\n\n@param string $title\n@param string $url\n...{'document': 'Add a breadcrumb to the collection.\n\n@param string $title\n@param string $url\n@...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_python
Dataset: CodeSearchNet_python at 9f89bdc
Size: 251,063 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'AbstractElement.settext def settext(self, text, cls=\'current\'):\n """Set the tex...{'document': 'Set the text for this element.\n\n Arguments:\n text (str): The text...{'document': 'Set text value as sole Text child node of element; any existing\n Text nodes ar...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ruby
Dataset: CodeSearchNet_ruby at 9f89bdc
Size: 24,731 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'CelluloidPubsub.Reactor.handle_parsed_websocket_message def handle_parsed_websocket_messa...{'document': 'method that checks if the data is a Hash\n\n if the data is a hash then will stringify...{'document': "If the message can be parsed into a Hash it will respond to the reactor's websocket co...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_go
Dataset: CodeSearchNet_ccr_go at 9f89bdc
Size: 167,278 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'getStringValue func getStringValue(b []rune) (int, error) {\n\tif b[0] != \'"\' {\n\t\tre...{'document': ' nil {\n\t\t\t\treturn 0, err\n\t\t\t}\n\n\t\t\tb[i-1] = c\n\t\t\tb = append(b[:i], b[...{'document': '\t\t\treturn 0, "", fmt.Errorf("nothing following final escape in %q", s)\n\t\t\t}\n\t...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_java
Dataset: CodeSearchNet_ccr_java at 9f89bdc
Size: 164,900 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'SCryptUtil.check public static boolean check(String passwd, String hashed) {\n try...{'document': ' int r = (int) params >> 8 & 0xff;\n int p = (int) params & 0...{'document': '\n } catch (Exception e) {\n throw new IllegalStateException("Validity checks ...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_javascript
Dataset: CodeSearchNet_ccr_javascript at 9f89bdc
Size: 58,017 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'function (state, action) {\n return _.defaults({\n ', 'query_id': 0}{'document': ' isValidating: action.isValidating,\n lastAction: IS_VALIDATING\n }, state)\n ...{'document': ' baz: action.payload,\n };\n default:\n return state;\n }\n}', 'd...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_php
Dataset: CodeSearchNet_ccr_php at 9f89bdc
Size: 241,177 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'BreadcrumbCollection.addOne public function addOne($title, $url, array $data = [])\n {...{'document': ' return $this->addBreadcrumb(\n BreadcrumbItem::make($title, $url, $data)\n...{'document': ' $this->breadcrumbs->push(new Breadcrumb($title, $url));\n }', 'document_id': 135...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_python
Dataset: CodeSearchNet_ccr_python at 9f89bdc
Size: 251,758 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'AbstractElement.settext def settext(self, text, cls=\'current\'):\n """Set the tex...{'document': ' only one text content element of each class associated with the element.\n """...{'document': '\n Jython and has been superseded by the \'ast\' module in Python 2.6 and\n ...
Loss: pylate.losses.contrastive.Contrastive
CodeSearchNet_ccr_ruby
Dataset: CodeSearchNet_ccr_ruby at 9f89bdc
Size: 24,918 training samples
Approximate statistics based on the first 1000 samples:
query positive negative_0 negative_1 negative_2 negative_3 negative_4 negative_5 negative_6 negative_7 negative_8 negative_9 negative_10 negative_11 negative_12 negative_13 negative_14 negative_15 negative_16 negative_17 negative_18 negative_19 negative_20 negative_21 negative_22 negative_23 negative_24 negative_25 negative_26 negative_27 negative_28 negative_29 negative_30 negative_31 negative_32 negative_33 negative_34 negative_35 negative_36 negative_37 negative_38 negative_39 negative_40 negative_41 negative_42 negative_43 negative_44 negative_45 negative_46 negative_47 negative_48 negative_49 type dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict dict details
Samples:
query positive negative_0 {'query': 'CelluloidPubsub.Reactor.handle_parsed_websocket_message def handle_parsed_websocket_messa...{'document': " delegate_action(data) if data['client_action'].present?\n else\n han...{'document': ' elsif data[\'method\']\n # RPC notice.\n event = { name: data[\'method\'], ...
Loss: pylate.losses.contrastive.Contrastive
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy: stepsper_device_train_batch_size: 128per_device_eval_batch_size: 128learning_rate: 3e-05num_train_epochs: 1bf16: Truedataloader_num_workers: 8accelerator_config: {'split_batches': True, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
All Hyperparameters
Click to expand
overwrite_output_dir: Falsedo_predict: Falseeval_strategy: stepsprediction_loss_only: Trueper_device_train_batch_size: 128per_device_eval_batch_size: 128per_gpu_train_batch_size: Noneper_gpu_eval_batch_size: Nonegradient_accumulation_steps: 1eval_accumulation_steps: Nonetorch_empty_cache_steps: Nonelearning_rate: 3e-05weight_decay: 0.0adam_beta1: 0.9adam_beta2: 0.999adam_epsilon: 1e-08max_grad_norm: 1.0num_train_epochs: 1max_steps: -1lr_scheduler_type: linearlr_scheduler_kwargs: {}warmup_ratio: 0.0warmup_steps: 0log_level: passivelog_level_replica: warninglog_on_each_node: Truelogging_nan_inf_filter: Truesave_safetensors: Truesave_on_each_node: Falsesave_only_model: Falserestore_callback_states_from_checkpoint: Falseno_cuda: Falseuse_cpu: Falseuse_mps_device: Falseseed: 42data_seed: Nonejit_mode_eval: Falsebf16: Truefp16: Falsefp16_opt_level: O1half_precision_backend: autobf16_full_eval: Falsefp16_full_eval: Falsetf32: Nonelocal_rank: 0ddp_backend: Nonetpu_num_cores: Nonetpu_metrics_debug: Falsedebug: []dataloader_drop_last: Truedataloader_num_workers: 8dataloader_prefetch_factor: Nonepast_index: -1disable_tqdm: Falseremove_unused_columns: Truelabel_names: Noneload_best_model_at_end: Falseignore_data_skip: Falsefsdp: []fsdp_min_num_params: 0fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}fsdp_transformer_layer_cls_to_wrap: Noneaccelerator_config: {'split_batches': True, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}parallelism_config: Nonedeepspeed: Nonelabel_smoothing_factor: 0.0optim: adamw_torch_fusedoptim_args: Noneadafactor: Falsegroup_by_length: Falselength_column_name: lengthproject: huggingfacetrackio_space_id: trackioddp_find_unused_parameters: Noneddp_bucket_cap_mb: Noneddp_broadcast_buffers: Falsedataloader_pin_memory: Truedataloader_persistent_workers: Falseskip_memory_metrics: Trueuse_legacy_prediction_loop: Falsepush_to_hub: Falseresume_from_checkpoint: Nonehub_model_id: Nonehub_strategy: every_savehub_private_repo: Nonehub_always_push: Falsehub_revision: Nonegradient_checkpointing: Falsegradient_checkpointing_kwargs: Noneinclude_inputs_for_metrics: Falseinclude_for_metrics: []eval_do_concat_batches: Truefp16_backend: autopush_to_hub_model_id: Nonepush_to_hub_organization: Nonemp_parameters:auto_find_batch_size: Falsefull_determinism: Falsetorchdynamo: Noneray_scope: lastddp_timeout: 1800torch_compile: Falsetorch_compile_backend: Nonetorch_compile_mode: Noneinclude_tokens_per_second: Falseinclude_num_input_tokens_seen: noneftune_noise_alpha: Noneoptim_target_modules: Nonebatch_eval_metrics: Falseeval_on_start: Falseuse_liger_kernel: Falseliger_kernel_config: Noneeval_use_gather_object: Falseaverage_tokens_across_devices: Trueprompts: Nonebatch_sampler: batch_samplerrouter_mapping: {}learning_rate_mapping: {}
Training Logs
Click to expand
Epoch Step Training Loss CodeSearchNetPython_MaxSim_ndcg@10 CodeSearchNetJavascript_MaxSim_ndcg@10 CodeSearchNetGo_MaxSim_ndcg@10 CodeSearchNetRuby_MaxSim_ndcg@10 CodeSearchNetJava_MaxSim_ndcg@10 CodeSearchNetPhp_MaxSim_ndcg@10 CodeSearchNet_mean_MaxSim_ndcg@10 0.0000 1 6.4113 - - - - - - - 0.0391 1250 3.2574 - - - - - - - 0.0781 2500 19.7862 0.9377 0.7986 0.9622 0.8487 0.8837 0.8834 0.8857 0.1172 3750 4.6875 - - - - - - - 0.1562 5000 2.3691 0.9335 0.8001 0.9614 0.8435 0.8755 0.8818 0.8826 0.1953 6250 1.4007 - - - - - - - 0.2344 7500 2.5715 0.9311 0.7960 0.9611 0.8418 0.8730 0.8866 0.8816 0.2734 8750 1.5546 - - - - - - - 0.3125 10000 0.004 0.9332 0.7972 0.9620 0.8435 0.8730 0.8850 0.8823 0.3515 11250 2.2819 - - - - - - - 0.3906 12500 14.0214 0.9324 0.7986 0.9603 0.8409 0.8717 0.8855 0.8816 0.4297 13750 2.0774 - - - - - - - 0.4687 15000 1.7724 0.9272 0.7955 0.9592 0.8381 0.8733 0.8838 0.8795 0.5078 16250 3.8234 - - - - - - - 0.5468 17500 0.7029 0.9300 0.7959 0.9594 0.8371 0.8674 0.8832 0.8788 0.5859 18750 1.5763 - - - - - - - 0.6250 20000 2.3146 0.9294 0.7986 0.9589 0.8376 0.8704 0.8829 0.8796 0.6640 21250 13.784 - - - - - - - 0.7031 22500 1.4557 0.9252 0.7927 0.9617 0.8357 0.8661 0.8839 0.8775 0.7421 23750 4.973 - - - - - - - 0.7812 25000 2.206 0.9240 0.7939 0.9623 0.8354 0.8639 0.8857 0.8775 0.8203 26250 0.7343 - - - - - - - 0.8593 27500 0.727 0.9251 0.7926 0.9608 0.8362 0.8676 0.8829 0.8775 0.8984 28750 1.7905 - - - - - - - 0.9374 30000 0.7259 0.9244 0.7937 0.9607 0.8357 0.8655 0.8824 0.8771
Framework Versions
Python: 3.12.12
Sentence Transformers: 5.1.1
PyLate: 1.3.4
Transformers: 4.57.3
PyTorch: 2.9.0+cu128
Accelerate: 1.12.0
Datasets: 4.4.2
Tokenizers: 0.22.2
Citation
BibTeX
LateOn-Code
1 @misc{LateOn-Code,
2 title = {LateOn-Code: a Family of State-Of-The-Art Late Interaction Code Retrieval Models},
3 author = {Chaffin, Antoine},
4 url = {https://huggingface.co/collections/lightonai/lateon-code},
5 year = {2026}
6 }
ColGrep
1 @software{next-plaid,
2 title = {NextPlaid, ColGREP: Multi-vector search, from database to coding agents.},
3 url = {https://github.com/lightonai/next-plaid},
4 author = {Raphaël Sourty},
5 year = {2026},
6 }
CoRNStack
1 @inproceedings{DBLP:conf/iclr/SureshRXNMDJ25,
2 author = {Tarun Suresh and
3 Revanth Gangi Reddy and
4 Yifei Xu and
5 Zach Nussbaum and
6 Andriy Mulyar and
7 Brandon Duderstadt and
8 Heng Ji},
9 title = {CoRNStack: High-Quality Contrastive Data for Better Code Retrieval
10 and Reranking},
11 booktitle = {The Thirteenth International Conference on Learning Representations,
12 {ICLR} 2025, Singapore, April 24-28, 2025},
13 publisher = {OpenReview.net},
14 year = {2025},
15 url = {https://openreview.net/forum?id=iyJOUELYir},
16 timestamp = {Sun, 25 May 2025 21:25:19 +0200},
17 biburl = {https://dblp.org/rec/conf/iclr/SureshRXNMDJ25.bib},
18 bibsource = {dblp computer science bibliography, https://dblp.org}
19 }
CoIR
1 @inproceedings{li2025coir,
2 title = {Coir: A comprehensive benchmark for code information retrieval models},
3 author = {Li, Xiangyang and Dong, Kuicai and Lee, Yi Quan and Xia, Wei and Zhang, Hao and Dai, Xinyi and Wang, Yasheng and Tang, Ruiming},
4 booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
5 pages = {22074--22091},
6 year = {2025}
7 }
Sentence Transformers
1 @inproceedings{reimers-2019-sentence-bert,
2 title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
3 author = "Reimers, Nils and Gurevych, Iryna",
4 booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
5 month = "11",
6 year = "2019",
7 publisher = "Association for Computational Linguistics",
8 url = "https://arxiv.org/abs/1908.10084"
9 }
PyLate
1 @inproceedings{DBLP:conf/cikm/ChaffinS25,
2 author = {Antoine Chaffin and
3 Rapha{"{e}}l Sourty},
4 editor = {Meeyoung Cha and
5 Chanyoung Park and
6 Noseong Park and
7 Carl Yang and
8 Senjuti Basu Roy and
9 Jessie Li and
10 Jaap Kamps and
11 Kijung Shin and
12 Bryan Hooi and
13 Lifang He},
14 title = {PyLate: Flexible Training and Retrieval for Late Interaction Models},
15 booktitle = {Proceedings of the 34th {ACM} International Conference on Information
16 and Knowledge Management, {CIKM} 2025, Seoul, Republic of Korea, November
17 10-14, 2025},
18 pages = {6334--6339},
19 publisher = {{ACM}},
20 year = {2025},
21 url = {https://github.com/lightonai/pylate},
22 doi = {10.1145/3746252.3761608},
23 }
