more fixes

app.py

@@ -32,13 +32,16 @@ else:
     device = "cpu"
     print("GPU is not available. Using CPU.")
 
+device = "cpu"
+print("Forcing CPU")
+
 # Load the processor
 model = "allenai/Molmo-7B-D-0924"
 
 processor = AutoProcessor.from_pretrained(
     model,
     trust_remote_code=True,
-    torch_dtype='auto',
+    torch_dtype=torch.bfloat16,
     device_map='auto'
 )
 
@@ -46,7 +49,7 @@ processor = AutoProcessor.from_pretrained(
 model = AutoModelForCausalLM.from_pretrained(
     model,
     trust_remote_code=True,
-    torch_dtype='auto',
+    torch_dtype=torch.bfloat16,
     device_map='auto',
 )
     
@@ -95,7 +98,7 @@ def generate_caption(image_path, processor, model, generation_config, bits_and_b
     inputs["images"] = inputs["images"].to(torch.bfloat16)
 
     # generate output; maximum 500 new tokens; stop generation when   is generated
-    with torch.autocast(device_type=device, enabled=True, dtype="auto"):
+    with torch.autocast(device_type=device, enabled=True, dtype=torch.bfloat16):
         output = model.generate_from_batch(
             inputs,
             GenerationConfig(max_new_tokens=200, stop_strings="<|endoftext|>"),

images/6a5e7d80-1c47-4dbf-89e2-43b5749f74ed.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:508ee3059b8e9ea66544730f4e912471ed19d9a87d1c6afd3cc94786dc469eea
+size 994