Update app.py
Browse files
app.py
CHANGED
@@ -2,71 +2,51 @@ import gradio as gr
|
|
2 |
import torch
|
3 |
from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer
|
4 |
from threading import Thread
|
5 |
-
from queue import Empty
|
6 |
-
|
7 |
-
# Load the tokenizer and model
|
8 |
-
tokenizer = AutoTokenizer.from_pretrained("thrishala/mental_health_chatbot")
|
9 |
-
model = AutoModelForCausalLM.from_pretrained("thrishala/mental_health_chatbot", torch_dtype=torch.float16)
|
10 |
|
|
|
|
|
11 |
# Move model to GPU if available
|
12 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
13 |
model = model.to(device)
|
14 |
|
15 |
class StopOnTokens(StoppingCriteria):
|
16 |
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
|
17 |
-
stop_ids = [29, 0]
|
18 |
for stop_id in stop_ids:
|
19 |
if input_ids[0][-1] == stop_id:
|
20 |
return True
|
21 |
return False
|
22 |
|
23 |
def predict(message, history):
|
24 |
-
# Prepare the input history in the expected format for the model
|
25 |
history_transformer_format = list(zip(history[:-1], history[1:])) + [[message, ""]]
|
26 |
stop = StopOnTokens()
|
27 |
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
# Tokenize and prepare model inputs
|
32 |
-
model_inputs = tokenizer([messages], return_tensors="pt").to(device)
|
33 |
|
34 |
-
|
35 |
-
streamer = TextIteratorStreamer(tokenizer, timeout=
|
36 |
-
|
37 |
-
# Define generation parameters
|
38 |
generate_kwargs = dict(
|
39 |
model_inputs,
|
40 |
streamer=streamer,
|
41 |
-
max_new_tokens=
|
42 |
do_sample=True,
|
43 |
-
top_p=0.
|
44 |
-
top_k=
|
45 |
temperature=1.0,
|
46 |
num_beams=1,
|
47 |
stopping_criteria=StoppingCriteriaList([stop])
|
48 |
-
|
49 |
-
|
50 |
-
# Run the generation in a separate thread
|
51 |
t = Thread(target=model.generate, kwargs=generate_kwargs)
|
52 |
t.start()
|
53 |
|
54 |
-
# Yield generated tokens
|
55 |
partial_message = ""
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
yield partial_message
|
61 |
-
except Empty:
|
62 |
-
yield "Error: No tokens generated or generation timeout."
|
63 |
-
|
64 |
-
# Gradio interface to run the chatbot
|
65 |
-
gr.ChatInterface(predict).launch()
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
|
|
|
70 |
|
71 |
|
72 |
|
|
|
2 |
import torch
|
3 |
from transformers import AutoModelForCausalLM, AutoTokenizer, StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer
|
4 |
from threading import Thread
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
+
tokenizer = AutoTokenizer.from_pretrained("togethercomputer/RedPajama-INCITE-Chat-3B-v1")
|
7 |
+
model = AutoModelForCausalLM.from_pretrained("togethercomputer/RedPajama-INCITE-Chat-3B-v1", torch_dtype=torch.float16)
|
8 |
# Move model to GPU if available
|
9 |
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
10 |
model = model.to(device)
|
11 |
|
12 |
class StopOnTokens(StoppingCriteria):
|
13 |
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
|
14 |
+
stop_ids = [29, 0]
|
15 |
for stop_id in stop_ids:
|
16 |
if input_ids[0][-1] == stop_id:
|
17 |
return True
|
18 |
return False
|
19 |
|
20 |
def predict(message, history):
|
|
|
21 |
history_transformer_format = list(zip(history[:-1], history[1:])) + [[message, ""]]
|
22 |
stop = StopOnTokens()
|
23 |
|
24 |
+
messages = "".join(["".join(["\n<human>:"+item[0], "\n<bot>:"+item[1]])
|
25 |
+
for item in history_transformer_format])
|
|
|
|
|
|
|
26 |
|
27 |
+
model_inputs = tokenizer([messages], return_tensors="pt").to("cuda")
|
28 |
+
streamer = TextIteratorStreamer(tokenizer, timeout=10., skip_prompt=True, skip_special_tokens=True)
|
|
|
|
|
29 |
generate_kwargs = dict(
|
30 |
model_inputs,
|
31 |
streamer=streamer,
|
32 |
+
max_new_tokens=1024,
|
33 |
do_sample=True,
|
34 |
+
top_p=0.95,
|
35 |
+
top_k=1000,
|
36 |
temperature=1.0,
|
37 |
num_beams=1,
|
38 |
stopping_criteria=StoppingCriteriaList([stop])
|
39 |
+
)
|
|
|
|
|
40 |
t = Thread(target=model.generate, kwargs=generate_kwargs)
|
41 |
t.start()
|
42 |
|
|
|
43 |
partial_message = ""
|
44 |
+
for new_token in streamer:
|
45 |
+
if new_token != '<':
|
46 |
+
partial_message += new_token
|
47 |
+
yield partial_message
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
|
49 |
+
gr.ChatInterface(predict).launch()
|
50 |
|
51 |
|
52 |
|