Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -57,20 +57,13 @@ def generate_text(messages):
|
|
57 |
model = AutoModelForCausalLM.from_pretrained(
|
58 |
model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
|
59 |
)
|
60 |
-
text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device ) #pipeline has not to(device)
|
61 |
result = text_generator(messages, max_new_tokens=256, do_sample=True, temperature=0.7)
|
62 |
|
63 |
-
generated_output =
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
content= message.get("content", "No content found.")
|
68 |
-
return content
|
69 |
-
|
70 |
-
return "No assistant response found."
|
71 |
-
else:
|
72 |
-
return "Unexpected output format."
|
73 |
-
|
74 |
|
75 |
|
76 |
def call_generate_text(message, history):
|
@@ -80,12 +73,12 @@ def call_generate_text(message, history):
|
|
80 |
|
81 |
messages = history+[{"role":"user","content":message}]
|
82 |
try:
|
83 |
-
|
84 |
-
|
|
|
85 |
except RuntimeError as e:
|
86 |
print(f"An unexpected error occurred: {e}")
|
87 |
-
|
88 |
-
return ""
|
89 |
|
90 |
demo = gr.ChatInterface(call_generate_text,type="messages")
|
91 |
|
|
|
57 |
model = AutoModelForCausalLM.from_pretrained(
|
58 |
model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
|
59 |
)
|
60 |
+
text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device ,stream=True) #pipeline has not to(device)
|
61 |
result = text_generator(messages, max_new_tokens=256, do_sample=True, temperature=0.7)
|
62 |
|
63 |
+
generated_output = ""
|
64 |
+
for token in result:
|
65 |
+
generated_output += token["generated_token"]
|
66 |
+
yield generated_output
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
|
68 |
|
69 |
def call_generate_text(message, history):
|
|
|
73 |
|
74 |
messages = history+[{"role":"user","content":message}]
|
75 |
try:
|
76 |
+
|
77 |
+
for text in generate_text(messages):
|
78 |
+
yield text
|
79 |
except RuntimeError as e:
|
80 |
print(f"An unexpected error occurred: {e}")
|
81 |
+
yield ""
|
|
|
82 |
|
83 |
demo = gr.ChatInterface(call_generate_text,type="messages")
|
84 |
|