wasabibish/similarity-code-ai-generated

SentenceTransformer based on distilbert/distilbert-base-uncased-finetuned-sst-2-english

This is a sentence-transformers model finetuned from distilbert/distilbert-base-uncased-finetuned-sst-2-english. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: distilbert/distilbert-base-uncased-finetuned-sst-2-english
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: DistilBertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

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("wasabibish/similarity-code-ai-generated")
# Run inference
sentences = [
    'def move_zeroes(nums):\n  count = 0\n  for i in range(len(nums)):\n    if nums[i] != 0:\n      nums[count], nums[i]= nums[i], nums[count]\n      count += 1\n  for i in range(count, len(nums)):\n    nums[i] =0\n\ninput = [int(x) for x in input("Enter integers separated by spaces: ").split()]\nmove_zeroes(input)\n\nprint(input)',
    'def move_zeros_to_end(lst):\n    zero_count = 0\n    for i in range(len(lst)):\n        if lst[i] != 0:\n            lst[i], lst[zero_count] = lst[zero_count], lst[i]\n            zero_count += 1\n\n# Test cases\nlst1 = [0, 1, 0, 3, 12]\nmove_zeros_to_end(lst1)\nprint(lst1)  # Output: [1, 3, 12, 0, 0]\n\nlst2 = [0, 0, 1]\nmove_zeros_to_end(lst2)\nprint(lst2)  # Output: [1, 0, 0]\n',
    'using System;\nusing System.Collections.Generic;\n\nclass BracketChecker\n{\n    private readonly Dictionary<char, char> bracketPairs = new Dictionary<char, char>\n    {\n        { \'(\', \')\' },\n        { \'[\', \']\' },\n        { \'{\', \'}\' }\n    };\n\n    public bool CheckBalancedBrackets(string input)\n    {\n        if (string.IsNullOrEmpty(input))\n        {\n            return true;\n        }\n\n        Stack<char> stack = new Stack<char>();\n\n        foreach (char c in input)\n        {\n            if (bracketPairs.ContainsValue(c))\n            {\n                if (stack.Count == 0 || bracketPairs[stack.Peek()] != c)\n                {\n                    return false;\n                }\n                stack.Pop();\n            }\n            else if (bracketPairs.ContainsKey(c))\n            {\n                stack.Push(c);\n            }\n        }\n\n        return stack.Count == 0;\n    }\n}\n\nclass Program\n{\n    static void Main()\n    {\n        BracketChecker bracketChecker = new BracketChecker();\n\n        string input1 = "(a+[b*c]-{d/e})";\n        Console.WriteLine("Input: \\"{0}\\"", input1);\n        Console.WriteLine("Output: {0}\\n", bracketChecker.CheckBalancedBrackets(input1));\n\n        string input2 = "(a+[b*c)-{d/e}]";\n        Console.WriteLine("Input: \\"{0}\\"", input2);\n        Console.WriteLine("Output: {0}", bracketChecker.CheckBalancedBrackets(input2));\n    }\n}\n',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Semantic Similarity

• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.9
spearman_cosine	0.9014
pearson_manhattan	0.862
spearman_manhattan	0.802
pearson_euclidean	0.8685
spearman_euclidean	0.8234
pearson_dot	0.8495
spearman_dot	0.8948
pearson_max	0.9
spearman_max	0.9014

Training Details

Training Dataset

Unnamed Dataset

• Size: 302 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 302 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokens mean: 206.43 tokens max: 512 tokens	min: 27 tokens mean: 244.9 tokens max: 512 tokens	min: 0.0 mean: 0.29 max: 0.9

• Samples:

  sentence1	sentence2	score
  from django.views.generic import ListView class PersonListView(ListView): model = Person template_name = 'person_list.html' def get_queryset(self): return Person.objects.filter(birthdate__year__lte=2005)	from myapp.models import Customer # Import the Customer model from your Django app def get_customers_with_zip_code_starting_with_123(): customers = Customer.objects.filter(zip_code__startswith='123').values() # Query to filter customers with zip_code starting with '123' return list(customers) # Return a list of dictionaries for matching records	0.4
  Welcome to our website!	function createSentence(words, maxChars) { if (words.length === 0
  AAAAAA	#include #include class KMP { public: std::vector findPatternIndices(const CString& text, const CString& pattern) { std::vector indices; if (pattern.IsEmpty()

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 76 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 76 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 5 tokens mean: 216.92 tokens max: 512 tokens	min: 54 tokens mean: 254.78 tokens max: 512 tokens	min: 0.0 mean: 0.33 max: 0.9

• Samples:

  sentence1	sentence2	score
  function stripHtmlTags(str) { return str.replace(/<[^>]*>/g, ''); } const input = ' Hello World! '; const output = stripHtmlTags(input); console.log(output);	function stripHtmlTags(input) { if (!input) return ''; const tagRegex = /<[^>]*>/g; return input.replace(tagRegex, ''); }	0.6
  function getTopThreeWords($text) { // Remove punctuation and convert to lowercase $words = str_word_count(strtolower(preg_replace('/[^\p{L}\p{N}\s]/u', ' ', $text)), 1); // Count the frequency of each word $wordFrequency = array_count_values($words); // Sort the words by frequency in descending order arsort($wordFrequency); // Get the top three words $topThreeWords = array_slice($wordFrequency, 0, 3, true); // Format the output $output = []; foreach ($topThreeWords as $word => $count) { $output[] = "('$word', $count)"; } return '[' . implode(', ', $output) . ']'; } // Example usage: $inputText = "The quick brown fox jumps over the lazy dog. The dog was lazy!"; echo getTopThreeWords($inputText); ?>	function countTopWords($inputString) { // Convert the input string to lowercase and remove punctuation $cleanString = preg_replace("/[\W_]+/", " ", strtolower($inputString)); // Split the string into an array of words $words = explode(" ", $cleanString); // Count the frequency of each word $wordCount = array_count_values($words); // Sort the words by frequency in descending order arsort($wordCount); // Get the top three most common words $topWords = array_slice($wordCount, 0, 3); // Format the output as an array of tuples $output = []; foreach ($topWords as $word => $count) { $output[] = [$word, $count]; } return $output; } // Test the function with the example input $inputString = "The quick brown fox jumps over the lazy dog. The dog was lazy!"; $output = countTopWords($inputString); print_r($output); ?>	0.3
  AAAAAA	#include #include class KMP { public: std::vector findPatternIndices(const CString& text, const CString& pattern) { std::vector indices; if (pattern.IsEmpty()

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• weight_decay: 0.2
• max_steps: 100
• warmup_steps: 150

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.2
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 3.0
• max_steps: 100
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• warmup_steps: 150
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• eval_use_gather_object: False
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	loss	spearman_max
0.5263	20	0.3765	0.5421
1.0526	40	0.1518	0.5774
1.5789	60	0.0501	0.8533
2.1053	80	0.0217	0.8900
2.6316	100	0.0168	0.9014

Framework Versions

• Python: 3.9.10
• Sentence Transformers: 3.1.0
• Transformers: 4.44.2
• PyTorch: 2.4.1+cpu
• Accelerate: 0.34.2
• Datasets: 3.0.0
• Tokenizers: 0.19.1

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}