mod allegro

allegro.py

@@ -8,32 +8,24 @@ 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
 
 # 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(
-        http_host="weaviate",
-        http_port=8080,
-        http_secure=False,
-        grpc_host="weaviate",
-        grpc_port=50051,
-        grpc_secure=False,
-    )
+client = weaviate.Client(
+    url="http://weaviate:8080"  # Dostosuj URL do swojego środowiska
 )
 
 collection_name = "Document"  # Zakładam, że to jest nazwa Twojej kolekcji
-result = (
-    client.query.get(collection_name, ["content"])
+response = (
+    client.query
+    .get(collection_name, ["content"])
     .with_additional(["id"])
     .do()
 )
 
-documents = [item['content'] for item in result['data']['Get'][collection_name]]
+documents = [item['content'] for item in response['data']['Get'][collection_name]]
 
 # 3️⃣ Generowanie embeddingów
 embeddings = embed_model.encode(documents)
@@ -82,7 +74,7 @@ tokenized_eval = eval_dataset.map(tokenize_function, batched=True)
 # 8️⃣ Parametry treningu
 training_args = TrainingArguments(
     output_dir="./results",
-    eval_strategy="steps",
+    evaluation_strategy="steps",
     eval_steps=500,
     save_strategy="steps",
     save_steps=500,