Update app.py

app.py

@@ -7,10 +7,17 @@ import gradio as gr
 from gradio import FileData
 import time
 import spaces
-ckpt = "Daemontatox/PixelParse_AI"
-model = MllamaForConditionalGeneration.from_pretrained(ckpt,
-    torch_dtype=torch.bfloat16).to("cuda")
-processor = AutoProcessor.from_pretrained(ckpt)
+from unsloth import FastVisionModel
+ckpt ="Daemontatox/DocumentLlama"
+#model = MllamaForConditionalGeneration.from_pretrained(ckpt,
+    #torch_dtype=torch.bfloat16).to("cuda")
+#processor = AutoProcessor.from_pretrained(ckpt)
+model, tokenizer = FastVisionModel.from_pretrained(
+    ckpt,
+    load_in_4bit = True, # Use 4bit to reduce memory use. False for 16bit LoRA.
+    use_gradient_checkpointing = "unsloth", # True or "unsloth" for long context
+)
+
 
 
 @spaces.GPU()