Spaces:

ariG23498
/

qwen25vl-api

Paused

ariG23498 HF staff commited on 14 days ago

Commit

2b1bf5b

verified ·

1 Parent(s): 60f50de

Update main.py

Files changed (1) hide show

main.py CHANGED Viewed

@@ -5,7 +5,7 @@ import torch
 app = FastAPI()
-checkpoint = "Qwen/Qwen2.5-VL-7B-Instruct"
 min_pixels = 256*28*28
 max_pixels = 1280*28*28
 processor = AutoProcessor.from_pretrained(
@@ -16,7 +16,7 @@ processor = AutoProcessor.from_pretrained(
 model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
     checkpoint,
     torch_dtype=torch.bfloat16,
-    device_map="cuda",
     # attn_implementation="flash_attention_2",
 )

 app = FastAPI()
+checkpoint = "Qwen/Qwen2.5-VL-3B-Instruct"
 min_pixels = 256*28*28
 max_pixels = 1280*28*28
 processor = AutoProcessor.from_pretrained(
 model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
     checkpoint,
     torch_dtype=torch.bfloat16,
+    device_map="auto",
     # attn_implementation="flash_attention_2",
 )