Update app.py

app.py

@@ -8,11 +8,9 @@ import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
 
 DESCRIPTION = """\
-# Gemma 2 2B JPN IT
+# magro-7b
 
-Gemma-2-JPN は Gemma 2 2B を日本語で fine-tune したものです。Gemma 2 の英語での性能と同レベルの性能で日本語をサポートします。
-
-(Gemma-2-JPN is a Gemma 2 2B model fine-tuned on Japanese text. It supports the Japanese language at the same level of performance as English-only queries on Gemma 2.)
+日本語ai
 """
 
 MAX_MAX_NEW_TOKENS = 2048
@@ -21,7 +19,7 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-model_id = "google/gemma-2-2b-jpn-it"
+model_id = "Sakalti/Magro-7B"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(
     model_id,
@@ -37,7 +35,7 @@ def generate(
     message: str,
     chat_history: list[dict],
     max_new_tokens: int = 1024,
-    temperature: float = 0.6,
+    temperature: float = 0.7,
     top_p: float = 0.9,
     top_k: int = 50,
     repetition_penalty: float = 1.2,