From 8e1f346f6ef64819d8f39ef791b27a230b023544 Mon Sep 17 00:00:00 2001
From: "l.gabrysiak" <l.gabrysiak@osadkowski.pl>
Date: Fri, 28 Feb 2025 20:58:24 +0100
Subject: [PATCH] mod allegro

---
 allegro.py | 97 ++++++++++++++++++++++--------------------------------
 1 file changed, 39 insertions(+), 58 deletions(-)

diff --git a/allegro.py b/allegro.py
index 978af6f..e2aa098 100644
--- a/allegro.py
+++ b/allegro.py
@@ -2,21 +2,15 @@ import os
 os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
 import torch
-import numpy as np
-from sentence_transformers import SentenceTransformer
-from datasets import Dataset
-from peft import LoraConfig, get_peft_model
-from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, TrainingArguments, Trainer, DataCollatorForSeq2Seq
 import weaviate
-from weaviate.client import WeaviateClient
-from weaviate.connect import ConnectionParams
+from weaviate.classes.config import Property, DataType, Configure
+from weaviate.classes.query import Query
+from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, TrainingArguments, Trainer, DataCollatorForSeq2Seq
+from datasets import Dataset
 
-# 1️⃣ Inicjalizacja modelu do embeddingów
-embed_model = SentenceTransformer("all-MiniLM-L6-v2")
-
-# 2️⃣ Połączenie z Weaviate i pobranie dokumentów
-client = WeaviateClient(
-    connection_params=ConnectionParams.from_params(
+# 1️⃣ Połączenie z bazą Weaviate
+client = weaviate.WeaviateClient(
+    connection_params=weaviate.ConnectionParams.from_params(
         http_host="weaviate",
         http_port=8080,
         http_secure=False,
@@ -26,48 +20,38 @@ client = WeaviateClient(
     )
 )
 
-collection_name = "Document"  # Zakładam, że to jest nazwa Twojej kolekcji
-result = (
-    client.query.get(collection_name, ["content"])
-    .with_additional(["id"])
-    .do()
-)
+# 2️⃣ Pobranie dokumentów z bazy Weaviate
+collection_name = "Documents"
+query = Query(collection_name).limit(1000)
+result = client.query.run(query)
 
-documents = [item['content'] for item in result['data']['Get'][collection_name]]
+documents = []
+file_names = []
 
-# 3️⃣ Generowanie embeddingów
-embeddings = embed_model.encode(documents)
+for item in result[collection_name]['objects']:
+    documents.append(item['properties']['content'])
+    file_names.append(item['properties']['fileName'])
 
-# 4️⃣ Przygotowanie danych treningowych
-def create_training_data():
-    data = {
-        "text": documents,
-        "embedding": embeddings.tolist()
-    }
-    return Dataset.from_dict(data)
+# 3️⃣ Tworzenie datasetu
+training_data = {
+    "text": documents,
+    "file_name": file_names
+}
+dataset = Dataset.from_dict(training_data)
 
-dataset = create_training_data()
-
-# Podział danych na treningowe i ewaluacyjne
+# Podział na zestaw treningowy i ewaluacyjny
 split_dataset = dataset.train_test_split(test_size=0.25)
 train_dataset = split_dataset["train"]
 eval_dataset = split_dataset["test"]
 
-# 5️⃣ Ładowanie modelu allegro/multislav-5lang
-device = "cuda" if torch.cuda.is_available() else "cpu"
+# 4️⃣ Ładowanie modelu Multislav
 model_name = "allegro/multislav-5lang"
+device = "cuda" if torch.cuda.is_available() else "cpu"
 model = AutoModelForSeq2SeqLM.from_pretrained(model_name).to(device)
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 
-# 6️⃣ Konfiguracja LoRA
-lora_config = LoraConfig(
-    r=8, lora_alpha=32, lora_dropout=0.1, bias="none", task_type="SEQ_2_SEQ_LM"
-)
-model = get_peft_model(model, lora_config)
-
-# 7️⃣ Tokenizacja danych
-max_length = 384
-
+# 5️⃣ Tokenizacja
+max_length = 512
 def tokenize_function(examples):
     return tokenizer(
         examples["text"],
@@ -79,30 +63,27 @@ def tokenize_function(examples):
 tokenized_train = train_dataset.map(tokenize_function, batched=True)
 tokenized_eval = eval_dataset.map(tokenize_function, batched=True)
 
-# 8️⃣ Parametry treningu
+# 6️⃣ Parametry treningu
 training_args = TrainingArguments(
     output_dir="./results",
-    eval_strategy="steps",
+    evaluation_strategy="steps",
     eval_steps=500,
     save_strategy="steps",
     save_steps=500,
-    learning_rate=1e-5,
-    per_device_train_batch_size=2,
-    per_device_eval_batch_size=2,
-    num_train_epochs=16,
+    learning_rate=2e-5,
+    per_device_train_batch_size=4,
+    per_device_eval_batch_size=4,
+    num_train_epochs=10,
     weight_decay=0.01,
     load_best_model_at_end=True,
     metric_for_best_model="loss",
     greater_is_better=False,
 )
 
-# 9️⃣ Data Collator
-data_collator = DataCollatorForSeq2Seq(
-    tokenizer=tokenizer,
-    model=model
-)
+# 7️⃣ Data Collator
+data_collator = DataCollatorForSeq2Seq(tokenizer=tokenizer, model=model)
 
-# 🔟 Trening modelu
+# 8️⃣ Trening
 trainer = Trainer(
     model=model,
     args=training_args,
@@ -113,8 +94,8 @@ trainer = Trainer(
 
 trainer.train()
 
-# 1️⃣1️⃣ Zapis modelu
-model.save_pretrained("./models/allegro")
-tokenizer.save_pretrained("./models/allegro")
+# 9️⃣ Zapis modelu
+model.save_pretrained("./trained_model/multislav")
+tokenizer.save_pretrained("./trained_model/multislav")
 
 print("✅ Model został wytrenowany i zapisany!")