Update app.py
Browse files
app.py
CHANGED
|
@@ -3,6 +3,7 @@ from PIL import Image
|
|
| 3 |
import gradio as gr
|
| 4 |
import spaces
|
| 5 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
|
|
|
|
| 6 |
import os
|
| 7 |
from threading import Thread
|
| 8 |
|
|
@@ -37,7 +38,7 @@ tokenizer = AutoTokenizer.from_pretrained(MODELS,trust_remote_code=True)
|
|
| 37 |
|
| 38 |
|
| 39 |
@spaces.GPU
|
| 40 |
-
def stream_chat(message: str, history: list, temperature: float,
|
| 41 |
print(f'message is - {message}')
|
| 42 |
print(f'history is - {history}')
|
| 43 |
conversation = []
|
|
@@ -48,22 +49,31 @@ def stream_chat(message: str, history: list, temperature: float, max_new_tokens:
|
|
| 48 |
print(f"Conversation is -\n{conversation}")
|
| 49 |
|
| 50 |
input_ids = tokenizer.apply_chat_template(conversation, tokenize=True, add_generation_prompt=True, return_tensors="pt", return_dict=True).to(model.device)
|
| 51 |
-
|
| 52 |
|
| 53 |
generate_kwargs = dict(
|
| 54 |
-
max_length=
|
| 55 |
-
|
| 56 |
do_sample=True,
|
| 57 |
top_k=1,
|
| 58 |
temperature=temperature,
|
| 59 |
repetition_penalty=1.2,
|
| 60 |
)
|
|
|
|
| 61 |
|
| 62 |
with torch.no_grad():
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
|
| 69 |
|
|
@@ -90,10 +100,10 @@ with gr.Blocks(css=CSS) as demo:
|
|
| 90 |
),
|
| 91 |
gr.Slider(
|
| 92 |
minimum=128,
|
| 93 |
-
maximum=
|
| 94 |
step=1,
|
| 95 |
value=1024,
|
| 96 |
-
label="Max
|
| 97 |
render=False,
|
| 98 |
),
|
| 99 |
],
|
|
|
|
| 3 |
import gradio as gr
|
| 4 |
import spaces
|
| 5 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
|
| 6 |
+
from huggingface_hub.inference._generated.types import TextGenerationStreamOutput, TextGenerationStreamOutputToken
|
| 7 |
import os
|
| 8 |
from threading import Thread
|
| 9 |
|
|
|
|
| 38 |
|
| 39 |
|
| 40 |
@spaces.GPU
|
| 41 |
+
def stream_chat(message: str, history: list, temperature: float, max_length: int):
|
| 42 |
print(f'message is - {message}')
|
| 43 |
print(f'history is - {history}')
|
| 44 |
conversation = []
|
|
|
|
| 49 |
print(f"Conversation is -\n{conversation}")
|
| 50 |
|
| 51 |
input_ids = tokenizer.apply_chat_template(conversation, tokenize=True, add_generation_prompt=True, return_tensors="pt", return_dict=True).to(model.device)
|
| 52 |
+
streamer = TextIteratorStreamer(tokenizer, timeout=60.0, skip_prompt=True, skip_special_tokens=True)
|
| 53 |
|
| 54 |
generate_kwargs = dict(
|
| 55 |
+
max_length=max_length,
|
| 56 |
+
streamer=streamer,
|
| 57 |
do_sample=True,
|
| 58 |
top_k=1,
|
| 59 |
temperature=temperature,
|
| 60 |
repetition_penalty=1.2,
|
| 61 |
)
|
| 62 |
+
gen_kwargs = {**input_ids, **generate_kwargs}
|
| 63 |
|
| 64 |
with torch.no_grad():
|
| 65 |
+
thread = Thread(target=model.generate, kwargs=gen_kwargs)
|
| 66 |
+
thread.start()
|
| 67 |
+
for next_text in streamer:
|
| 68 |
+
yield TextGenerationStreamOutput(
|
| 69 |
+
index=0,
|
| 70 |
+
token=TextGenerationStreamOutputToken(
|
| 71 |
+
id=0,
|
| 72 |
+
logprob=0,
|
| 73 |
+
text=next_text,
|
| 74 |
+
special=False,
|
| 75 |
+
)
|
| 76 |
+
)
|
| 77 |
|
| 78 |
|
| 79 |
|
|
|
|
| 100 |
),
|
| 101 |
gr.Slider(
|
| 102 |
minimum=128,
|
| 103 |
+
maximum=8192,
|
| 104 |
step=1,
|
| 105 |
value=1024,
|
| 106 |
+
label="Max Length",
|
| 107 |
render=False,
|
| 108 |
),
|
| 109 |
],
|