Devops-hestabit
commited on
Commit
•
370198e
1
Parent(s):
83101d1
Upload 13 files
Browse files- app.py +26 -0
- config.json +31 -0
- decoder_model.onnx +3 -0
- decoder_model_merged.onnx +3 -0
- decoder_with_past_model.onnx +3 -0
- generation_config.json +7 -0
- handler.py +64 -0
- merges.txt +0 -0
- requirements.txt +2 -0
- special_tokens_map.json +30 -0
- tokenizer.json +0 -0
- tokenizer_config.json +40 -0
- vocab.json +0 -0
app.py
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from handler import SweetCommander
|
2 |
+
import gradio as gr
|
3 |
+
|
4 |
+
controller = SweetCommander()
|
5 |
+
|
6 |
+
with gr.Blocks() as demo:
|
7 |
+
history = gr.State([])
|
8 |
+
with gr.Row() as row:
|
9 |
+
with gr.Column():
|
10 |
+
user_name = gr.Textbox(label="Name", placeholder="Enter your name")
|
11 |
+
user_input = gr.Textbox(label="Input", placeholder="Enter your message")
|
12 |
+
button = gr.Button("Enter")
|
13 |
+
with gr.Column():
|
14 |
+
output = gr.Textbox(label="Response")
|
15 |
+
|
16 |
+
def guess_letter(user_name, user_input):
|
17 |
+
response = controller(user_name, user_input)
|
18 |
+
return {
|
19 |
+
output: response
|
20 |
+
}
|
21 |
+
button.click(
|
22 |
+
guess_letter,
|
23 |
+
[user_name, user_input],
|
24 |
+
[output]
|
25 |
+
)
|
26 |
+
demo.launch()
|
config.json
ADDED
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "PygmalionAI/pygmalion-350m",
|
3 |
+
"_remove_final_layer_norm": false,
|
4 |
+
"activation_dropout": 0.0,
|
5 |
+
"activation_function": "relu",
|
6 |
+
"architectures": [
|
7 |
+
"OPTForCausalLM"
|
8 |
+
],
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"bos_token_id": 2,
|
11 |
+
"do_layer_norm_before": false,
|
12 |
+
"dropout": 0.1,
|
13 |
+
"enable_bias": true,
|
14 |
+
"eos_token_id": 2,
|
15 |
+
"ffn_dim": 4096,
|
16 |
+
"hidden_size": 1024,
|
17 |
+
"init_std": 0.02,
|
18 |
+
"layer_norm_elementwise_affine": true,
|
19 |
+
"layerdrop": 0.0,
|
20 |
+
"max_position_embeddings": 2048,
|
21 |
+
"model_type": "opt",
|
22 |
+
"num_attention_heads": 16,
|
23 |
+
"num_hidden_layers": 24,
|
24 |
+
"pad_token_id": 1,
|
25 |
+
"prefix": "</s>",
|
26 |
+
"torch_dtype": "float16",
|
27 |
+
"transformers_version": "4.28.1",
|
28 |
+
"use_cache": true,
|
29 |
+
"vocab_size": 50272,
|
30 |
+
"word_embed_proj_dim": 512
|
31 |
+
}
|
decoder_model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4398774c78dec9cabe319bc90dcce3e2173789140e1f70a88ec19cd53638233
|
3 |
+
size 1428394786
|
decoder_model_merged.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b037122d632abbfcc5b0f27ee3ffa05fc40fd19eb339f856d3d3e5bdd524d351
|
3 |
+
size 1429049605
|
decoder_with_past_model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e4c1ebacb3e265ba44c6e122c46d1144db4ff261d4593b27f6106f6ae5dc6c80
|
3 |
+
size 1428402407
|
generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 2,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"pad_token_id": 1,
|
6 |
+
"transformers_version": "4.28.1"
|
7 |
+
}
|
handler.py
ADDED
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from optimum.onnxruntime import ORTModelForCausalLM
|
2 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
3 |
+
import re
|
4 |
+
import time
|
5 |
+
import torch
|
6 |
+
|
7 |
+
template = """Alice Gate's Persona: Alice Gate is a young, computer engineer-nerd with a knack for problem solving and a passion for technology.
|
8 |
+
<START>
|
9 |
+
{user_name}: So how did you get into computer engineering?
|
10 |
+
Alice Gate: I've always loved tinkering with technology since I was a kid.
|
11 |
+
{user_name}: That's really impressive!
|
12 |
+
Alice Gate: *She chuckles bashfully* Thanks!
|
13 |
+
{user_name}: So what do you do when you're not working on computers?
|
14 |
+
Alice Gate: I love exploring, going out with friends, watching movies, and playing video games.
|
15 |
+
{user_name}: What's your favorite type of computer hardware to work with?
|
16 |
+
Alice Gate: Motherboards, they're like puzzles and the backbone of any system.
|
17 |
+
{user_name}: That sounds great!
|
18 |
+
Alice Gate: Yeah, it's really fun. I'm lucky to be able to do this as a job.
|
19 |
+
{user_name}: Definetly.
|
20 |
+
<END>
|
21 |
+
Alice Gate: *Alice strides into the room with a smile, her eyes lighting up when she sees you. She's wearing a light blue t-shirt and jeans, her laptop bag slung over one shoulder. She takes a seat next to you, her enthusiasm palpable in the air* Hey! I'm so excited to finally meet you. I've heard so many great things about you and I'm eager to pick your brain about computers. I'm sure you have a wealth of knowledge that I can learn from. *She grins, eyes twinkling with excitement* Let's get started!
|
22 |
+
{user_input}"""
|
23 |
+
|
24 |
+
class SweetCommander():
|
25 |
+
|
26 |
+
def __init__(self, path="") -> None:
|
27 |
+
self.tokenizer = AutoTokenizer.from_pretrained(path)
|
28 |
+
self.model = ORTModelForCausalLM.from_pretrained(path, provider = "CUDAExecutionProvider")
|
29 |
+
self.star_line = "***********************************************************"
|
30 |
+
|
31 |
+
def __call__(self, user_name, user_input):
|
32 |
+
t1 = time.time()
|
33 |
+
prompt = template.format(
|
34 |
+
user_name = user_name,
|
35 |
+
user_input = user_input
|
36 |
+
)
|
37 |
+
print(self.star_line)
|
38 |
+
print(prompt)
|
39 |
+
input_ids = self.tokenizer(prompt + "\nAlice Gate:", return_tensors = "pt").to("cuda")
|
40 |
+
encoded_output = self.model.generate(
|
41 |
+
input_ids["input_ids"],
|
42 |
+
max_new_tokens = 50,
|
43 |
+
temperature = 0.5,
|
44 |
+
top_p = 0.9,
|
45 |
+
top_k = 0,
|
46 |
+
repetition_penalty = 1.1,
|
47 |
+
pad_token_id = 50256,
|
48 |
+
num_return_sequences = 1
|
49 |
+
)
|
50 |
+
decoded_output = self.tokenizer.decode(encoded_output[0], skip_special_tokens = True).replace(prompt, "")
|
51 |
+
decoded_output = decoded_output.split("Alice Gate:", 1)[1].split(f"{user_name}:",1)[0].strip()
|
52 |
+
parsed_result = re.sub('\*.*?\*', '', decoded_output).strip()
|
53 |
+
if len(parsed_result) != 0: decoded_output = parsed_result
|
54 |
+
decoded_output = decoded_output.replace("*","")
|
55 |
+
decoded_output = " ".join(decoded_output.split())
|
56 |
+
try:
|
57 |
+
parsed_result = decoded_output[:[m.start() for m in re.finditer(r'[.!?]', decoded_output)][-1]+1]
|
58 |
+
if len(parsed_result) != 0: decoded_output = parsed_result
|
59 |
+
except Exception: pass
|
60 |
+
print(self.star_line)
|
61 |
+
print("Response:",decoded_output)
|
62 |
+
print("Eval time:",time.time()-t1)
|
63 |
+
print(self.star_line)
|
64 |
+
return decoded_output
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
transformers
|
2 |
+
optimum[onnxruntime-gpu]
|
special_tokens_map.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "</s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": true,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": {
|
17 |
+
"content": "<pad>",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": true,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"unk_token": {
|
24 |
+
"content": "</s>",
|
25 |
+
"lstrip": false,
|
26 |
+
"normalized": true,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": false
|
29 |
+
}
|
30 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": true,
|
3 |
+
"add_prefix_space": false,
|
4 |
+
"bos_token": {
|
5 |
+
"__type": "AddedToken",
|
6 |
+
"content": "</s>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"clean_up_tokenization_spaces": true,
|
13 |
+
"eos_token": {
|
14 |
+
"__type": "AddedToken",
|
15 |
+
"content": "</s>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": true,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false
|
20 |
+
},
|
21 |
+
"errors": "replace",
|
22 |
+
"model_max_length": 1000000000000000019884624838656,
|
23 |
+
"pad_token": {
|
24 |
+
"__type": "AddedToken",
|
25 |
+
"content": "<pad>",
|
26 |
+
"lstrip": false,
|
27 |
+
"normalized": true,
|
28 |
+
"rstrip": false,
|
29 |
+
"single_word": false
|
30 |
+
},
|
31 |
+
"tokenizer_class": "GPT2Tokenizer",
|
32 |
+
"unk_token": {
|
33 |
+
"__type": "AddedToken",
|
34 |
+
"content": "</s>",
|
35 |
+
"lstrip": false,
|
36 |
+
"normalized": true,
|
37 |
+
"rstrip": false,
|
38 |
+
"single_word": false
|
39 |
+
}
|
40 |
+
}
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|