Spaces:
Paused
Paused
Update demo/app.py
Browse files- demo/app.py +67 -80
demo/app.py
CHANGED
@@ -4,6 +4,7 @@ import logging
|
|
4 |
import sys
|
5 |
import gradio as gr
|
6 |
import torch
|
|
|
7 |
from app_modules.utils import *
|
8 |
from app_modules.presets import *
|
9 |
from app_modules.overwrites import *
|
@@ -15,53 +16,48 @@ logging.basicConfig(
|
|
15 |
|
16 |
base_model = "decapoda-research/llama-7b-hf"
|
17 |
adapter_model = "/home/user/app/checkpoint-100"
|
18 |
-
tokenizer,
|
19 |
-
|
20 |
-
|
21 |
-
def predict(
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
|
|
|
|
|
|
32 |
return
|
33 |
|
34 |
-
inputs = generate_prompt_with_history(
|
35 |
-
text, history, tokenizer, max_length=max_context_length_tokens
|
36 |
-
)
|
37 |
if inputs is None:
|
38 |
-
yield chatbot,
|
39 |
-
return
|
40 |
else:
|
41 |
-
prompt,
|
42 |
begin_length = len(prompt)
|
43 |
-
input_ids = inputs["input_ids"][
|
44 |
torch.cuda.empty_cache()
|
45 |
-
|
|
|
|
|
|
|
|
|
46 |
with torch.no_grad():
|
47 |
-
for x in
|
48 |
-
|
49 |
-
model,
|
50 |
-
tokenizer,
|
51 |
-
stop_words=["[|Human|]", "[|AI|]"],
|
52 |
-
max_length=max_length_tokens,
|
53 |
-
temperature=temperature,
|
54 |
-
top_p=top_p,
|
55 |
-
):
|
56 |
-
if is_stop_word_or_prefix(x, ["[|Human|]", "[|AI|]"]) is False:
|
57 |
if "[|Human|]" in x:
|
58 |
-
x = x[:
|
59 |
if "[|AI|]" in x:
|
60 |
-
x = x[:
|
61 |
-
x = x.strip(
|
62 |
-
a, b
|
63 |
-
[text, convert_to_markdown(x)]
|
64 |
-
], history + [[text, x]]
|
65 |
yield a, b, "Generating..."
|
66 |
if shared_state.interrupted:
|
67 |
shared_state.recover()
|
@@ -70,40 +66,33 @@ def predict(
|
|
70 |
return
|
71 |
except:
|
72 |
pass
|
|
|
|
|
73 |
torch.cuda.empty_cache()
|
74 |
-
print(
|
75 |
-
print(x)
|
76 |
-
print("="
|
77 |
try:
|
78 |
-
yield a,
|
79 |
except:
|
80 |
pass
|
81 |
-
|
82 |
-
|
83 |
def retry(
|
84 |
-
|
85 |
-
chatbot,
|
86 |
-
history,
|
87 |
-
top_p,
|
88 |
-
temperature,
|
89 |
-
max_length_tokens,
|
90 |
-
max_context_length_tokens,
|
91 |
-
):
|
92 |
-
logging.info("Retry...")
|
93 |
-
if len(history) == 0:
|
94 |
-
yield chatbot, history, "Empty context."
|
95 |
-
return
|
96 |
-
chatbot.pop()
|
97 |
-
inputs = history.pop()[0]
|
98 |
-
for x in predict(
|
99 |
-
inputs,
|
100 |
chatbot,
|
101 |
history,
|
102 |
top_p,
|
103 |
temperature,
|
104 |
max_length_tokens,
|
105 |
max_context_length_tokens,
|
106 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
107 |
yield x
|
108 |
|
109 |
|
@@ -132,13 +121,12 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
132 |
submitBtn = gr.Button("Send")
|
133 |
with gr.Column(min_width=70, scale=1):
|
134 |
cancelBtn = gr.Button("Stop")
|
135 |
-
|
136 |
with gr.Row(scale=1):
|
137 |
emptyBtn = gr.Button(
|
138 |
"๐งน New Conversation",
|
139 |
)
|
140 |
retryBtn = gr.Button("๐ Regenerate")
|
141 |
-
delLastBtn = gr.Button("๐๏ธ Remove Last Turn")
|
142 |
with gr.Column():
|
143 |
with gr.Column(min_width=50, scale=1):
|
144 |
with gr.Tab(label="Parameter Setting"):
|
@@ -162,7 +150,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
162 |
max_length_tokens = gr.Slider(
|
163 |
minimum=0,
|
164 |
maximum=512,
|
165 |
-
value=
|
166 |
step=8,
|
167 |
interactive=True,
|
168 |
label="Max Generation Tokens",
|
@@ -206,20 +194,18 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
206 |
show_progress=True,
|
207 |
)
|
208 |
|
209 |
-
reset_args = dict(
|
210 |
-
|
|
|
|
|
211 |
# Chatbot
|
212 |
-
cancelBtn.click(cancel_outputing, [], [status_display])
|
213 |
transfer_input_args = dict(
|
214 |
-
fn=transfer_input,
|
215 |
-
inputs=[user_input],
|
216 |
-
outputs=[user_question, user_input, submitBtn, cancelBtn],
|
217 |
-
show_progress=True,
|
218 |
)
|
219 |
|
220 |
-
user_input.submit(**transfer_input_args).then(**predict_args)
|
221 |
|
222 |
-
submitBtn.click(**transfer_input_args).then(**predict_args)
|
223 |
|
224 |
emptyBtn.click(
|
225 |
reset_state,
|
@@ -228,7 +214,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
228 |
)
|
229 |
emptyBtn.click(**reset_args)
|
230 |
|
231 |
-
retryBtn.click(**retry_args)
|
232 |
|
233 |
delLastBtn.click(
|
234 |
delete_last_conversation,
|
@@ -236,11 +222,12 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
|
|
236 |
[chatbot, history, status_display],
|
237 |
show_progress=True,
|
238 |
)
|
239 |
-
|
|
|
|
|
|
|
|
|
|
|
240 |
demo.title = "Baize"
|
241 |
|
242 |
-
|
243 |
-
reload_javascript()
|
244 |
-
demo.queue(concurrency_count=CONCURRENT_COUNT).launch(
|
245 |
-
share=False, favicon_path="/home/user/app/demo/assets/favicon.ico", inbrowser=True
|
246 |
-
)
|
|
|
4 |
import sys
|
5 |
import gradio as gr
|
6 |
import torch
|
7 |
+
import gc
|
8 |
from app_modules.utils import *
|
9 |
from app_modules.presets import *
|
10 |
from app_modules.overwrites import *
|
|
|
16 |
|
17 |
base_model = "decapoda-research/llama-7b-hf"
|
18 |
adapter_model = "/home/user/app/checkpoint-100"
|
19 |
+
tokenizer,model,device = load_tokenizer_and_model(base_model,adapter_model)
|
20 |
+
|
21 |
+
total_count = 0
|
22 |
+
def predict(text,
|
23 |
+
chatbot,
|
24 |
+
history,
|
25 |
+
top_p,
|
26 |
+
temperature,
|
27 |
+
max_length_tokens,
|
28 |
+
max_context_length_tokens,):
|
29 |
+
if text=="":
|
30 |
+
yield chatbot,history,"Empty context."
|
31 |
+
return
|
32 |
+
try:
|
33 |
+
model
|
34 |
+
except:
|
35 |
+
yield [[text,"No Model Found"]],[],"No Model Found"
|
36 |
return
|
37 |
|
38 |
+
inputs = generate_prompt_with_history(text,history,tokenizer,max_length=max_context_length_tokens)
|
|
|
|
|
39 |
if inputs is None:
|
40 |
+
yield chatbot,history,"Input too long."
|
41 |
+
return
|
42 |
else:
|
43 |
+
prompt,inputs=inputs
|
44 |
begin_length = len(prompt)
|
45 |
+
input_ids = inputs["input_ids"][:,-max_context_length_tokens:].to(device)
|
46 |
torch.cuda.empty_cache()
|
47 |
+
global total_count
|
48 |
+
total_count += 1
|
49 |
+
print(total_count)
|
50 |
+
if total_count % 50 == 0 :
|
51 |
+
os.system("nvidia-smi")
|
52 |
with torch.no_grad():
|
53 |
+
for x in greedy_search(input_ids,model,tokenizer,stop_words=["[|Human|]", "[|AI|]"],max_length=max_length_tokens,temperature=temperature,top_p=top_p):
|
54 |
+
if is_stop_word_or_prefix(x,["[|Human|]", "[|AI|]"]) is False:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
55 |
if "[|Human|]" in x:
|
56 |
+
x = x[:x.index("[|Human|]")].strip()
|
57 |
if "[|AI|]" in x:
|
58 |
+
x = x[:x.index("[|AI|]")].strip()
|
59 |
+
x = x.strip()
|
60 |
+
a, b= [[y[0],convert_to_markdown(y[1])] for y in history]+[[text, convert_to_markdown(x)]],history + [[text,x]]
|
|
|
|
|
61 |
yield a, b, "Generating..."
|
62 |
if shared_state.interrupted:
|
63 |
shared_state.recover()
|
|
|
66 |
return
|
67 |
except:
|
68 |
pass
|
69 |
+
del input_ids
|
70 |
+
gc.collect()
|
71 |
torch.cuda.empty_cache()
|
72 |
+
#print(text)
|
73 |
+
#print(x)
|
74 |
+
#print("="*80)
|
75 |
try:
|
76 |
+
yield a,b,"Generate: Success"
|
77 |
except:
|
78 |
pass
|
79 |
+
|
|
|
80 |
def retry(
|
81 |
+
text,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
82 |
chatbot,
|
83 |
history,
|
84 |
top_p,
|
85 |
temperature,
|
86 |
max_length_tokens,
|
87 |
max_context_length_tokens,
|
88 |
+
):
|
89 |
+
logging.info("Retry...")
|
90 |
+
if len(history) == 0:
|
91 |
+
yield chatbot, history, f"Empty context"
|
92 |
+
return
|
93 |
+
chatbot.pop()
|
94 |
+
inputs = history.pop()[0]
|
95 |
+
for x in predict(inputs,chatbot,history,top_p,temperature,max_length_tokens,max_context_length_tokens):
|
96 |
yield x
|
97 |
|
98 |
|
|
|
121 |
submitBtn = gr.Button("Send")
|
122 |
with gr.Column(min_width=70, scale=1):
|
123 |
cancelBtn = gr.Button("Stop")
|
|
|
124 |
with gr.Row(scale=1):
|
125 |
emptyBtn = gr.Button(
|
126 |
"๐งน New Conversation",
|
127 |
)
|
128 |
retryBtn = gr.Button("๐ Regenerate")
|
129 |
+
delLastBtn = gr.Button("๐๏ธ Remove Last Turn")
|
130 |
with gr.Column():
|
131 |
with gr.Column(min_width=50, scale=1):
|
132 |
with gr.Tab(label="Parameter Setting"):
|
|
|
150 |
max_length_tokens = gr.Slider(
|
151 |
minimum=0,
|
152 |
maximum=512,
|
153 |
+
value=256,
|
154 |
step=8,
|
155 |
interactive=True,
|
156 |
label="Max Generation Tokens",
|
|
|
194 |
show_progress=True,
|
195 |
)
|
196 |
|
197 |
+
reset_args = dict(
|
198 |
+
fn=reset_textbox, inputs=[], outputs=[user_input, status_display]
|
199 |
+
)
|
200 |
+
|
201 |
# Chatbot
|
|
|
202 |
transfer_input_args = dict(
|
203 |
+
fn=transfer_input, inputs=[user_input], outputs=[user_question, user_input, submitBtn], show_progress=True
|
|
|
|
|
|
|
204 |
)
|
205 |
|
206 |
+
predict_event1 = user_input.submit(**transfer_input_args).then(**predict_args)
|
207 |
|
208 |
+
predict_event2 = submitBtn.click(**transfer_input_args).then(**predict_args)
|
209 |
|
210 |
emptyBtn.click(
|
211 |
reset_state,
|
|
|
214 |
)
|
215 |
emptyBtn.click(**reset_args)
|
216 |
|
217 |
+
predict_event3 = retryBtn.click(**retry_args)
|
218 |
|
219 |
delLastBtn.click(
|
220 |
delete_last_conversation,
|
|
|
222 |
[chatbot, history, status_display],
|
223 |
show_progress=True,
|
224 |
)
|
225 |
+
cancelBtn.click(
|
226 |
+
cancel_outputing, [], [status_display],
|
227 |
+
cancels=[
|
228 |
+
predict_event1,predict_event2,predict_event3
|
229 |
+
]
|
230 |
+
)
|
231 |
demo.title = "Baize"
|
232 |
|
233 |
+
demo.queue(concurrency_count=1).launch()
|
|
|
|
|
|
|
|