Update app.py

app.py

@@ -8,7 +8,7 @@ def ensure_file(filename, src):
         total_size = int(response.headers.get('content-length', 0))
 
         with open(filename, 'wb') as file:
-            with tqdm(total=total_size, unit='8', unit_scale=True, desc=filename, ncols=80) as progress_bar:
+            with tqdm(total=total_size, unit='B', unit_scale=True, desc=filename, ncols=80) as progress_bar:
                 for data in response.iter_content(chunk_size=1024):
                     if data:
                         file.write(data)
@@ -27,7 +27,7 @@ llm = Llama(model_path="ggml-model-q4_k.gguf", n_ctx=2048)
 def generate(image, ins='Describe the image'):
     if len(ins) < 1:
         ins = 'Describe the image'
-    image_embed = llava_image_embed_make_with_filename(ctx_clip=ctx_clip, n_threads=1, filename=image.encode('utf-8'))
+    image_embed = llava_image_embed_make_with_filename(ctx_clip=ctx_clip, n_threads=1, filename=image.encode('utf8'))
 
     n_past = ctypes.c_int(llm.n_tokens)
     n_past_p = ctypes.byref(n_past)
@@ -35,13 +35,13 @@ def generate(image, ins='Describe the image'):
     llm.n_tokens = n_past.value
     llava_image_embed_free(image_embed)
 
-    llm.eval(llm.tokenize(ins.encode('utf-8')))
+    llm.eval(llm.tokenize(ins.encode('utf8')))
 
     max_target_len = 256
     res = ''
     for i in range(max_target_len):
         t_id = llm.sample(temp=0.3)
-        t = llm.detokenize([t_id]).decode('utf-8')
+        t = llm.detokenize([t_id]).decode('utf8')
         if t == '</s>':
             break 
         res += t