Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -4,23 +4,13 @@ import torch
|
|
4 |
import subprocess
|
5 |
import sys
|
6 |
from threading import Thread
|
7 |
-
from transformers import
|
8 |
-
import numpy as np
|
9 |
-
|
10 |
-
from queue import Queue
|
11 |
-
from threading import Event
|
12 |
|
13 |
# Install required packages
|
14 |
-
subprocess.check_call([sys.executable, "-m", "pip", "install", "-U", "--force-reinstall", "
|
15 |
subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
|
16 |
|
17 |
-
|
18 |
-
from transformers import OlmoeForCausalLM
|
19 |
-
import librosa
|
20 |
-
# Import speech-to-speech components
|
21 |
-
from VAD.vad_handler import VADHandler
|
22 |
-
from STT.whisper_stt_handler import WhisperSTTHandler
|
23 |
-
from TTS.parler_handler import ParlerTTSHandler
|
24 |
|
25 |
model_name = "allenai/OLMoE-1B-7B-0924-Instruct"
|
26 |
|
@@ -30,12 +20,12 @@ try:
|
|
30 |
model = OlmoeForCausalLM.from_pretrained(
|
31 |
model_name,
|
32 |
trust_remote_code=True,
|
33 |
-
torch_dtype=torch.float16,
|
34 |
low_cpu_mem_usage=True,
|
35 |
device_map="auto",
|
36 |
-
_attn_implementation="flash_attention_2"
|
37 |
).to(DEVICE)
|
38 |
-
model.gradient_checkpointing_enable()
|
39 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
40 |
except Exception as e:
|
41 |
print(f"Error loading model: {e}")
|
@@ -47,32 +37,10 @@ system_prompt = ("Adopt the persona of hilariously pissed off Andrej Karpathy "
|
|
47 |
"while always answering questions in full first principles analysis type of thinking "
|
48 |
"without using any analogies and always showing full working code or output in his answers.")
|
49 |
|
50 |
-
# Setup speech-to-speech components
|
51 |
-
stop_event = Event()
|
52 |
-
should_listen = Event()
|
53 |
-
vad = VADHandler(stop_event, Queue(), Queue(), setup_args=(should_listen,))
|
54 |
-
stt = WhisperSTTHandler(stop_event, Queue(), Queue())
|
55 |
-
tts = ParlerTTSHandler(stop_event, Queue(), Queue(), setup_args=(should_listen,))
|
56 |
-
|
57 |
-
@spaces.GPU
|
58 |
-
def speech_to_text(audio):
|
59 |
-
if audio is None:
|
60 |
-
return ""
|
61 |
-
audio_np = librosa.resample(audio[1], orig_sr=audio[0], target_sr=16000)
|
62 |
-
audio_np = (audio_np * 32768).astype(np.int16)
|
63 |
-
|
64 |
-
vad_output = vad.process(audio_np)
|
65 |
-
stt_output, _ = next(stt.process(vad_output))
|
66 |
-
return stt_output
|
67 |
-
|
68 |
-
@spaces.GPU
|
69 |
-
def user(user_message, history):
|
70 |
-
return "", history + [[user_message, None]]
|
71 |
-
|
72 |
@spaces.GPU
|
73 |
-
def
|
74 |
if model is None or tokenizer is None:
|
75 |
-
yield
|
76 |
return
|
77 |
|
78 |
messages = [{"role": "system", "content": system_prompt}]
|
@@ -80,6 +48,7 @@ def bot(history, temperature, max_new_tokens):
|
|
80 |
messages.append({"role": "user", "content": user_msg})
|
81 |
if assistant_msg:
|
82 |
messages.append({"role": "assistant", "content": assistant_msg})
|
|
|
83 |
|
84 |
inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to(DEVICE)
|
85 |
|
@@ -97,53 +66,52 @@ def bot(history, temperature, max_new_tokens):
|
|
97 |
thread = Thread(target=model.generate, kwargs=generation_kwargs)
|
98 |
thread.start()
|
99 |
|
100 |
-
|
101 |
for new_text in streamer:
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
|
|
|
|
|
|
|
|
106 |
except Exception as e:
|
107 |
-
|
108 |
-
yield history
|
109 |
-
|
110 |
-
def text_to_speech(text):
|
111 |
-
audio_output = np.concatenate(list(tts.process(text)))
|
112 |
-
return (16000, audio_output)
|
113 |
|
114 |
css = """
|
115 |
#output {
|
116 |
-
height:
|
117 |
overflow: auto;
|
118 |
-
border:
|
119 |
}
|
120 |
"""
|
121 |
|
122 |
with gr.Blocks(css=css) as demo:
|
123 |
-
gr.Markdown("# Nisten's Karpathy Chatbot with OSS OLMoE (Now with
|
124 |
chatbot = gr.Chatbot(elem_id="output")
|
125 |
-
audio_input = gr.Audio(source="microphone", type="numpy")
|
126 |
-
audio_output = gr.Audio()
|
127 |
msg = gr.Textbox(label="Meow")
|
128 |
with gr.Row():
|
129 |
temperature = gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature")
|
130 |
max_new_tokens = gr.Slider(minimum=50, maximum=4000, value=2000, step=50, label="Max New Tokens")
|
131 |
clear = gr.Button("Clear")
|
132 |
|
133 |
-
def
|
134 |
-
|
135 |
-
|
136 |
-
|
137 |
-
|
|
|
|
|
|
|
|
|
|
|
138 |
|
139 |
-
audio_input.stop_recording(process_audio, [audio_input, chatbot, temperature, max_new_tokens], [chatbot, audio_output])
|
140 |
msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
|
141 |
bot, [chatbot, temperature, max_new_tokens], chatbot
|
142 |
-
).then(
|
143 |
-
lambda history: text_to_speech(history[-1][1]), chatbot, audio_output
|
144 |
)
|
145 |
clear.click(lambda: None, None, chatbot, queue=False)
|
146 |
|
147 |
if __name__ == "__main__":
|
148 |
-
demo.queue(api_open=True, max_size=10)
|
149 |
-
demo.launch(debug=True, show_api=True, share=False)
|
|
|
4 |
import subprocess
|
5 |
import sys
|
6 |
from threading import Thread
|
7 |
+
from transformers import TextIteratorStreamer
|
|
|
|
|
|
|
|
|
8 |
|
9 |
# Install required packages
|
10 |
+
subprocess.check_call([sys.executable, "-m", "pip", "install", "-U", "--force-reinstall", "--no-deps", "einops", "accelerate", "git+https://github.com/Muennighoff/transformers.git@olmoe"])
|
11 |
subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
|
12 |
|
13 |
+
from transformers import OlmoeForCausalLM, AutoTokenizer
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
|
15 |
model_name = "allenai/OLMoE-1B-7B-0924-Instruct"
|
16 |
|
|
|
20 |
model = OlmoeForCausalLM.from_pretrained(
|
21 |
model_name,
|
22 |
trust_remote_code=True,
|
23 |
+
torch_dtype=torch.float16, # Using float16 for lower precision
|
24 |
low_cpu_mem_usage=True,
|
25 |
device_map="auto",
|
26 |
+
_attn_implementation="flash_attention_2" # Enable Flash Attention 2
|
27 |
).to(DEVICE)
|
28 |
+
model.gradient_checkpointing_enable() # Enable gradient checkpointing
|
29 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
30 |
except Exception as e:
|
31 |
print(f"Error loading model: {e}")
|
|
|
37 |
"while always answering questions in full first principles analysis type of thinking "
|
38 |
"without using any analogies and always showing full working code or output in his answers.")
|
39 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
@spaces.GPU
|
41 |
+
def generate_response(message, history, temperature, max_new_tokens):
|
42 |
if model is None or tokenizer is None:
|
43 |
+
yield "Model or tokenizer not loaded properly. Please check the logs."
|
44 |
return
|
45 |
|
46 |
messages = [{"role": "system", "content": system_prompt}]
|
|
|
48 |
messages.append({"role": "user", "content": user_msg})
|
49 |
if assistant_msg:
|
50 |
messages.append({"role": "assistant", "content": assistant_msg})
|
51 |
+
messages.append({"role": "user", "content": message})
|
52 |
|
53 |
inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to(DEVICE)
|
54 |
|
|
|
66 |
thread = Thread(target=model.generate, kwargs=generation_kwargs)
|
67 |
thread.start()
|
68 |
|
69 |
+
partial_message = ""
|
70 |
for new_text in streamer:
|
71 |
+
partial_message += new_text
|
72 |
+
yield partial_message.strip()
|
73 |
+
|
74 |
+
except RuntimeError as e:
|
75 |
+
if "CUDA out of memory" in str(e):
|
76 |
+
yield "GPU memory exceeded. Try reducing the max tokens or using a smaller model."
|
77 |
+
else:
|
78 |
+
yield f"An error occurred: {str(e)}"
|
79 |
except Exception as e:
|
80 |
+
yield f"An unexpected error occurred: {str(e)}"
|
|
|
|
|
|
|
|
|
|
|
81 |
|
82 |
css = """
|
83 |
#output {
|
84 |
+
height: 1100px;
|
85 |
overflow: auto;
|
86 |
+
border: 3px solid #ccc;
|
87 |
}
|
88 |
"""
|
89 |
|
90 |
with gr.Blocks(css=css) as demo:
|
91 |
+
gr.Markdown("# Nisten's Karpathy Chatbot with OSS OLMoE (Now with Flash Attention 2 and Streaming!)")
|
92 |
chatbot = gr.Chatbot(elem_id="output")
|
|
|
|
|
93 |
msg = gr.Textbox(label="Meow")
|
94 |
with gr.Row():
|
95 |
temperature = gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature")
|
96 |
max_new_tokens = gr.Slider(minimum=50, maximum=4000, value=2000, step=50, label="Max New Tokens")
|
97 |
clear = gr.Button("Clear")
|
98 |
|
99 |
+
def user(user_message, history):
|
100 |
+
return "", history + [[user_message, None]]
|
101 |
+
|
102 |
+
def bot(history, temp, max_tokens):
|
103 |
+
user_message = history[-1][0]
|
104 |
+
bot_message = ""
|
105 |
+
for token in generate_response(user_message, history[:-1], temp, max_tokens):
|
106 |
+
bot_message = token
|
107 |
+
history[-1][1] = bot_message
|
108 |
+
yield history
|
109 |
|
|
|
110 |
msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
|
111 |
bot, [chatbot, temperature, max_new_tokens], chatbot
|
|
|
|
|
112 |
)
|
113 |
clear.click(lambda: None, None, chatbot, queue=False)
|
114 |
|
115 |
if __name__ == "__main__":
|
116 |
+
demo.queue(api_open=True, max_size=10) # Limiting queue size
|
117 |
+
demo.launch(debug=True, show_api=True, share=False) # Disabled sharing for security
|