feat: 试了一下 internVL2 这个模型还可以

app.py

@@ -2,13 +2,23 @@ import torch
 import gradio as gr
 from transformers import AutoModel, pipeline, AutoTokenizer
 
-inference = pipeline(task="visual-question-answering")
+path = "radna/Triton-InternVL2-2B"
+model = (
+    AutoModel.from_pretrained(
+        path, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True
+    )
+    .eval()
+    .cuda()
+)
+
+tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
+inference = pipeline(task="visual-question-answering", model=model, tokenizer=tokenizer)
 
 
 def predict(input_img, questions):
     try:
         predictions = inference(question=questions, image=input_img)
-        return str(predictions[0])
+        return str(predictions)
     except Exception as e:
         # 捕获异常，并将错误信息转换为字符串
         error_message = str(e)
@@ -19,9 +29,7 @@ def predict(input_img, questions):
 gradio_app = gr.Interface(
     predict,
     inputs=[
-        gr.Image(
-            label="Select A Image", sources=["upload", "webcam"], type="pil"
-        ),
+        gr.Image(label="Select A Image", sources=["upload", "webcam"], type="pil"),
         "text",
     ],
     outputs="text",