Spaces:
Paused
Paused
import pandas as pd | |
from transformers import pipeline | |
import torch | |
import re | |
model_id = "meta-llama/Meta-Llama-3-8B" | |
generator = pipeline( | |
"text-generation", | |
model="meta-llama/Meta-Llama-3-8B", | |
model_kwargs={"torch_dtype": torch.bfloat16}, | |
device=-1 | |
) | |
# Load dictionary from CSV file | |
csv_file_path = './dictionary/dictionary.csv' | |
df = pd.read_csv(csv_file_path) | |
dictionary = df['description'].tolist() | |
# Method to generate embeddings using the text generation pipeline | |
def generate_embedding(sentence): | |
# Generate a text embedding using the pipeline | |
inputs = generator.tokenizer(sentence, return_tensors='pt')['input_ids'] | |
with torch.no_grad(): | |
outputs = generator.model(inputs) | |
# Extract the embeddings from the logits | |
embeddings = outputs.logits.mean(dim=1).squeeze() | |
return embeddings | |
# Method to find the best match for the input word in the dictionary | |
def match_word(input_word, dictionary): | |
# Extract words from the input | |
words = re.findall(r'\w+', input_word.lower()) | |
# Filter dictionary based on words | |
filtered_dictionary = [desc for desc in dictionary if any(word in desc.lower() for word in words)] | |
print(f"Filtered dictionary size: {len(filtered_dictionary)}") | |
# print(f"Filtered dictionary: {filtered_dictionary}") | |
# Generate embeddings and calculate cosine similarity on the filtered dictionary | |
input_embedding = generate_embedding(input_word) | |
similarities = [] | |
for entry in filtered_dictionary: | |
entry_embedding = generate_embedding(entry) | |
similarity_score = torch.nn.functional.cosine_similarity(input_embedding, entry_embedding, dim=0).item() | |
similarities.append((entry, similarity_score)) | |
# print(similarities) | |
if similarities: | |
best_match = max(similarities, key=lambda x: x[1]) | |
return best_match if best_match[1] > 0.7 else None | |
else: | |
return None | |
# Example usage | |
input_words = [ | |
"Pepper - Habanero Pepper", "Bananas (12 lbs)", "Squash - Yellow Squash", "Cauliflower", | |
"Squash mix italian/yellow (30 lbs)", "Tomato - Roma Tomato", "Tomato - Grape Tomato", | |
"Squash - Mexican Squash", "Pepper - Bell Pepper", "Squash - Italian Squash", | |
"Pepper - Red Fresno Pepper", "Tomato - Cherry Tomato", "Pepper - Serrano Pepper", | |
"Kale ( 5 lbs)", "Tomato - Beefsteak Tomato", "Pepper - Anaheim Pepper", | |
"Banana - Burro Banana", "Squash - Butternut Squash", "Apricot ( 10 lbs)", | |
"Squash - Acorn Squash", "Tomato - Heirloom Tomato", "Pepper - Pasilla Pepper", | |
"Pepper - Jalapeno Pepper", "carrot (10 lbs )" | |
] | |
for input_word in input_words: | |
matched_entry = match_word(input_word, dictionary) | |
if matched_entry: | |
print("Input word:", input_word) | |
print("Matched entry:", matched_entry[0]) | |
print("Similarity score:", matched_entry[1]) | |
else: | |
print("Input word:", input_word) | |
print("Matched entry: None") | |
print() | |