Spaces:
Sleeping
Sleeping
Rohan Kumar Singh
commited on
Commit
·
e2d2960
1
Parent(s):
0fcc9d3
initial commit
Browse files- .gitattributes +1 -0
- .ipynb_checkpoints/Untitled-checkpoint.ipynb +340 -0
- Untitled.ipynb +340 -0
- __pycache__/gradio.cpython-310.pyc +0 -0
- app.py +123 -0
- best-model.ckpt +3 -0
- requirements.txt +5 -0
.gitattributes
CHANGED
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
35 |
+
best-model.ckpt filter=lfs diff=lfs merge=lfs -text
|
.ipynb_checkpoints/Untitled-checkpoint.ipynb
ADDED
@@ -0,0 +1,340 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cells": [
|
3 |
+
{
|
4 |
+
"cell_type": "code",
|
5 |
+
"execution_count": 1,
|
6 |
+
"id": "e0102cb4",
|
7 |
+
"metadata": {},
|
8 |
+
"outputs": [
|
9 |
+
{
|
10 |
+
"name": "stderr",
|
11 |
+
"output_type": "stream",
|
12 |
+
"text": [
|
13 |
+
"Global seed set to 100\n"
|
14 |
+
]
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"data": {
|
18 |
+
"text/plain": [
|
19 |
+
"100"
|
20 |
+
]
|
21 |
+
},
|
22 |
+
"execution_count": 1,
|
23 |
+
"metadata": {},
|
24 |
+
"output_type": "execute_result"
|
25 |
+
}
|
26 |
+
],
|
27 |
+
"source": [
|
28 |
+
"from transformers import T5Tokenizer, T5ForConditionalGeneration \n",
|
29 |
+
"\n",
|
30 |
+
"from transformers import AdamW\n",
|
31 |
+
"import pandas as pd\n",
|
32 |
+
"import torch\n",
|
33 |
+
"import pytorch_lightning as pl\n",
|
34 |
+
"from pytorch_lightning.callbacks import ModelCheckpoint\n",
|
35 |
+
"from torch.nn.utils.rnn import pad_sequence\n",
|
36 |
+
"# from torch.utils.data import Dataset, DataLoader, random_split, RandomSampler, SequentialSampler\n",
|
37 |
+
"\n",
|
38 |
+
"pl.seed_everything(100)"
|
39 |
+
]
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"cell_type": "code",
|
43 |
+
"execution_count": 2,
|
44 |
+
"id": "1ec5ec2a",
|
45 |
+
"metadata": {},
|
46 |
+
"outputs": [],
|
47 |
+
"source": [
|
48 |
+
"MODEL_NAME='t5-base'"
|
49 |
+
]
|
50 |
+
},
|
51 |
+
{
|
52 |
+
"cell_type": "code",
|
53 |
+
"execution_count": 3,
|
54 |
+
"id": "8044c622",
|
55 |
+
"metadata": {},
|
56 |
+
"outputs": [],
|
57 |
+
"source": [
|
58 |
+
"DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
|
59 |
+
"INPUT_MAX_LEN = 128 \n",
|
60 |
+
"OUTPUT_MAX_LEN = 128"
|
61 |
+
]
|
62 |
+
},
|
63 |
+
{
|
64 |
+
"cell_type": "code",
|
65 |
+
"execution_count": 4,
|
66 |
+
"id": "6390f2de",
|
67 |
+
"metadata": {},
|
68 |
+
"outputs": [],
|
69 |
+
"source": [
|
70 |
+
"tokenizer = T5Tokenizer.from_pretrained(MODEL_NAME, model_max_length=512)"
|
71 |
+
]
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"cell_type": "code",
|
75 |
+
"execution_count": 5,
|
76 |
+
"id": "8eec35d1",
|
77 |
+
"metadata": {},
|
78 |
+
"outputs": [],
|
79 |
+
"source": [
|
80 |
+
"class T5Model(pl.LightningModule):\n",
|
81 |
+
" \n",
|
82 |
+
" def __init__(self):\n",
|
83 |
+
" super().__init__()\n",
|
84 |
+
" self.model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME, return_dict = True)\n",
|
85 |
+
"\n",
|
86 |
+
" \n",
|
87 |
+
" def forward(self, input_ids, attention_mask, labels=None):\n",
|
88 |
+
" \n",
|
89 |
+
" output = self.model(\n",
|
90 |
+
" input_ids=input_ids, \n",
|
91 |
+
" attention_mask=attention_mask, \n",
|
92 |
+
" labels=labels\n",
|
93 |
+
" )\n",
|
94 |
+
" return output.loss, output.logits\n",
|
95 |
+
" \n",
|
96 |
+
" def training_step(self, batch, batch_idx):\n",
|
97 |
+
"\n",
|
98 |
+
" input_ids = batch[\"input_ids\"]\n",
|
99 |
+
" attention_mask = batch[\"attention_mask\"]\n",
|
100 |
+
" labels= batch[\"target\"]\n",
|
101 |
+
" loss, logits = self(input_ids , attention_mask, labels)\n",
|
102 |
+
"\n",
|
103 |
+
" \n",
|
104 |
+
" self.log(\"train_loss\", loss, prog_bar=True, logger=True)\n",
|
105 |
+
"\n",
|
106 |
+
" return {'loss': loss}\n",
|
107 |
+
" \n",
|
108 |
+
" def validation_step(self, batch, batch_idx):\n",
|
109 |
+
" input_ids = batch[\"input_ids\"]\n",
|
110 |
+
" attention_mask = batch[\"attention_mask\"]\n",
|
111 |
+
" labels= batch[\"target\"]\n",
|
112 |
+
" loss, logits = self(input_ids, attention_mask, labels)\n",
|
113 |
+
"\n",
|
114 |
+
" self.log(\"val_loss\", loss, prog_bar=True, logger=True)\n",
|
115 |
+
" \n",
|
116 |
+
" return {'val_loss': loss}\n",
|
117 |
+
"\n",
|
118 |
+
" def configure_optimizers(self):\n",
|
119 |
+
" return AdamW(self.parameters(), lr=0.0001)"
|
120 |
+
]
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"cell_type": "code",
|
124 |
+
"execution_count": 6,
|
125 |
+
"id": "e9d96844",
|
126 |
+
"metadata": {},
|
127 |
+
"outputs": [
|
128 |
+
{
|
129 |
+
"name": "stderr",
|
130 |
+
"output_type": "stream",
|
131 |
+
"text": [
|
132 |
+
"Lightning automatically upgraded your loaded checkpoint from v1.9.3 to v2.0.2. To apply the upgrade to your files permanently, run `python -m pytorch_lightning.utilities.upgrade_checkpoint --file F:\\Projects & Open_source\\Chatbot_T5_kaggle\\best-model.ckpt`\n"
|
133 |
+
]
|
134 |
+
}
|
135 |
+
],
|
136 |
+
"source": [
|
137 |
+
"train_model = T5Model.load_from_checkpoint('best-model.ckpt',map_location=DEVICE)"
|
138 |
+
]
|
139 |
+
},
|
140 |
+
{
|
141 |
+
"cell_type": "code",
|
142 |
+
"execution_count": 7,
|
143 |
+
"id": "3449943f",
|
144 |
+
"metadata": {},
|
145 |
+
"outputs": [],
|
146 |
+
"source": [
|
147 |
+
"train_model.freeze()"
|
148 |
+
]
|
149 |
+
},
|
150 |
+
{
|
151 |
+
"cell_type": "code",
|
152 |
+
"execution_count": 8,
|
153 |
+
"id": "0e9f1058",
|
154 |
+
"metadata": {},
|
155 |
+
"outputs": [],
|
156 |
+
"source": [
|
157 |
+
"def generate_question(question):\n",
|
158 |
+
"\n",
|
159 |
+
" inputs_encoding = tokenizer(\n",
|
160 |
+
" question,\n",
|
161 |
+
" add_special_tokens=True,\n",
|
162 |
+
" max_length= INPUT_MAX_LEN,\n",
|
163 |
+
" padding = 'max_length',\n",
|
164 |
+
" truncation='only_first',\n",
|
165 |
+
" return_attention_mask=True,\n",
|
166 |
+
" return_tensors=\"pt\"\n",
|
167 |
+
" )\n",
|
168 |
+
"\n",
|
169 |
+
" \n",
|
170 |
+
" generate_ids = train_model.model.generate(\n",
|
171 |
+
" input_ids = inputs_encoding[\"input_ids\"],\n",
|
172 |
+
" attention_mask = inputs_encoding[\"attention_mask\"],\n",
|
173 |
+
" max_length = INPUT_MAX_LEN,\n",
|
174 |
+
" num_beams = 4,\n",
|
175 |
+
" num_return_sequences = 1,\n",
|
176 |
+
" no_repeat_ngram_size=2,\n",
|
177 |
+
" early_stopping=True,\n",
|
178 |
+
" )\n",
|
179 |
+
"\n",
|
180 |
+
" preds = [\n",
|
181 |
+
" tokenizer.decode(gen_id,\n",
|
182 |
+
" skip_special_tokens=True, \n",
|
183 |
+
" clean_up_tokenization_spaces=True)\n",
|
184 |
+
" for gen_id in generate_ids\n",
|
185 |
+
" ]\n",
|
186 |
+
"\n",
|
187 |
+
" return \"\".join(preds)\n"
|
188 |
+
]
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"cell_type": "code",
|
192 |
+
"execution_count": 9,
|
193 |
+
"id": "ee38a88c",
|
194 |
+
"metadata": {},
|
195 |
+
"outputs": [
|
196 |
+
{
|
197 |
+
"name": "stdout",
|
198 |
+
"output_type": "stream",
|
199 |
+
"text": [
|
200 |
+
"Ques: hi, how are you doing?\n",
|
201 |
+
"BOT: i'm so glad you're doing well.\n"
|
202 |
+
]
|
203 |
+
}
|
204 |
+
],
|
205 |
+
"source": [
|
206 |
+
"ques = \"hi, how are you doing?\"\n",
|
207 |
+
"print(\"Ques: \",ques)\n",
|
208 |
+
"print(\"BOT: \",generate_question(ques))"
|
209 |
+
]
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"cell_type": "code",
|
213 |
+
"execution_count": 11,
|
214 |
+
"id": "22aa4414",
|
215 |
+
"metadata": {},
|
216 |
+
"outputs": [
|
217 |
+
{
|
218 |
+
"name": "stdout",
|
219 |
+
"output_type": "stream",
|
220 |
+
"text": [
|
221 |
+
"Running on local URL: http://127.0.0.1:7861\n",
|
222 |
+
"\n",
|
223 |
+
"To create a public link, set `share=True` in `launch()`.\n"
|
224 |
+
]
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"data": {
|
228 |
+
"text/html": [
|
229 |
+
"<div><iframe src=\"http://127.0.0.1:7861/\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
|
230 |
+
],
|
231 |
+
"text/plain": [
|
232 |
+
"<IPython.core.display.HTML object>"
|
233 |
+
]
|
234 |
+
},
|
235 |
+
"metadata": {},
|
236 |
+
"output_type": "display_data"
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"data": {
|
240 |
+
"text/plain": []
|
241 |
+
},
|
242 |
+
"execution_count": 11,
|
243 |
+
"metadata": {},
|
244 |
+
"output_type": "execute_result"
|
245 |
+
}
|
246 |
+
],
|
247 |
+
"source": [
|
248 |
+
"import gradio as gr\n",
|
249 |
+
"import random\n",
|
250 |
+
"import time\n",
|
251 |
+
"\n",
|
252 |
+
"with gr.Blocks() as demo:\n",
|
253 |
+
" chatbot = gr.Chatbot()\n",
|
254 |
+
" gr.Chatbot.style(chatbot,height=400)\n",
|
255 |
+
" msg = gr.Textbox(info=\"Press \\'Enter\\' to send\")\n",
|
256 |
+
" clear = gr.Button(\"Clear\")\n",
|
257 |
+
"\n",
|
258 |
+
" def user(user_message, history):\n",
|
259 |
+
" return \"\", history + [[user_message, None]]\n",
|
260 |
+
"\n",
|
261 |
+
" def bot(history):\n",
|
262 |
+
" bot_message = generate_question(history[-1][0])\n",
|
263 |
+
" history[-1][1] = \"\"\n",
|
264 |
+
" for character in bot_message:\n",
|
265 |
+
" history[-1][1] += character\n",
|
266 |
+
" time.sleep(0.05)\n",
|
267 |
+
" yield history\n",
|
268 |
+
"\n",
|
269 |
+
" msg.submit(user, [msg, chatbot], [msg, chatbot], queue=True).then(\n",
|
270 |
+
" bot, chatbot, chatbot\n",
|
271 |
+
" )\n",
|
272 |
+
" clear.click(lambda: None, None, chatbot, queue=True)\n",
|
273 |
+
"\n",
|
274 |
+
"demo.queue(concurrency_count=2)\n",
|
275 |
+
"demo.launch()"
|
276 |
+
]
|
277 |
+
},
|
278 |
+
{
|
279 |
+
"cell_type": "code",
|
280 |
+
"execution_count": null,
|
281 |
+
"id": "fef38bdc",
|
282 |
+
"metadata": {
|
283 |
+
"scrolled": true
|
284 |
+
},
|
285 |
+
"outputs": [],
|
286 |
+
"source": [
|
287 |
+
"import gradio as gr\n",
|
288 |
+
"import random\n",
|
289 |
+
"import time\n",
|
290 |
+
"\n",
|
291 |
+
"with gr.Blocks() as demo:\n",
|
292 |
+
" chatbot = gr.Chatbot()\n",
|
293 |
+
" msg = gr.Textbox(placeholder='Got any spare time...let\\'s chat!!!')\n",
|
294 |
+
" gr.Textbox.style(msg,show_copy_button=True)\n",
|
295 |
+
" clear = gr.Button(\"Clear\")\n",
|
296 |
+
"\n",
|
297 |
+
" def respond(message, chat_history):\n",
|
298 |
+
" bot_message = generate_question(message)\n",
|
299 |
+
" bot_message = \"**\"+bot_message+\"**\"\n",
|
300 |
+
" chat_history.append((message, bot_message))\n",
|
301 |
+
" time.sleep(1)\n",
|
302 |
+
" return \"\", chat_history\n",
|
303 |
+
"\n",
|
304 |
+
" msg.submit(respond, [msg, chatbot], [msg, chatbot])\n",
|
305 |
+
" clear.click(lambda: None, None, chatbot, queue=False)\n",
|
306 |
+
"\n",
|
307 |
+
"demo.launch()"
|
308 |
+
]
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"cell_type": "code",
|
312 |
+
"execution_count": null,
|
313 |
+
"id": "a86d446a",
|
314 |
+
"metadata": {},
|
315 |
+
"outputs": [],
|
316 |
+
"source": []
|
317 |
+
}
|
318 |
+
],
|
319 |
+
"metadata": {
|
320 |
+
"kernelspec": {
|
321 |
+
"display_name": "Python 3 (ipykernel)",
|
322 |
+
"language": "python",
|
323 |
+
"name": "python3"
|
324 |
+
},
|
325 |
+
"language_info": {
|
326 |
+
"codemirror_mode": {
|
327 |
+
"name": "ipython",
|
328 |
+
"version": 3
|
329 |
+
},
|
330 |
+
"file_extension": ".py",
|
331 |
+
"mimetype": "text/x-python",
|
332 |
+
"name": "python",
|
333 |
+
"nbconvert_exporter": "python",
|
334 |
+
"pygments_lexer": "ipython3",
|
335 |
+
"version": "3.10.1"
|
336 |
+
}
|
337 |
+
},
|
338 |
+
"nbformat": 4,
|
339 |
+
"nbformat_minor": 5
|
340 |
+
}
|
Untitled.ipynb
ADDED
@@ -0,0 +1,340 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cells": [
|
3 |
+
{
|
4 |
+
"cell_type": "code",
|
5 |
+
"execution_count": 1,
|
6 |
+
"id": "e0102cb4",
|
7 |
+
"metadata": {},
|
8 |
+
"outputs": [
|
9 |
+
{
|
10 |
+
"name": "stderr",
|
11 |
+
"output_type": "stream",
|
12 |
+
"text": [
|
13 |
+
"Global seed set to 100\n"
|
14 |
+
]
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"data": {
|
18 |
+
"text/plain": [
|
19 |
+
"100"
|
20 |
+
]
|
21 |
+
},
|
22 |
+
"execution_count": 1,
|
23 |
+
"metadata": {},
|
24 |
+
"output_type": "execute_result"
|
25 |
+
}
|
26 |
+
],
|
27 |
+
"source": [
|
28 |
+
"from transformers import T5Tokenizer, T5ForConditionalGeneration \n",
|
29 |
+
"\n",
|
30 |
+
"from transformers import AdamW\n",
|
31 |
+
"import pandas as pd\n",
|
32 |
+
"import torch\n",
|
33 |
+
"import pytorch_lightning as pl\n",
|
34 |
+
"from pytorch_lightning.callbacks import ModelCheckpoint\n",
|
35 |
+
"from torch.nn.utils.rnn import pad_sequence\n",
|
36 |
+
"# from torch.utils.data import Dataset, DataLoader, random_split, RandomSampler, SequentialSampler\n",
|
37 |
+
"\n",
|
38 |
+
"pl.seed_everything(100)"
|
39 |
+
]
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"cell_type": "code",
|
43 |
+
"execution_count": 2,
|
44 |
+
"id": "1ec5ec2a",
|
45 |
+
"metadata": {},
|
46 |
+
"outputs": [],
|
47 |
+
"source": [
|
48 |
+
"MODEL_NAME='t5-base'"
|
49 |
+
]
|
50 |
+
},
|
51 |
+
{
|
52 |
+
"cell_type": "code",
|
53 |
+
"execution_count": 3,
|
54 |
+
"id": "8044c622",
|
55 |
+
"metadata": {},
|
56 |
+
"outputs": [],
|
57 |
+
"source": [
|
58 |
+
"DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
|
59 |
+
"INPUT_MAX_LEN = 128 \n",
|
60 |
+
"OUTPUT_MAX_LEN = 128"
|
61 |
+
]
|
62 |
+
},
|
63 |
+
{
|
64 |
+
"cell_type": "code",
|
65 |
+
"execution_count": 4,
|
66 |
+
"id": "6390f2de",
|
67 |
+
"metadata": {},
|
68 |
+
"outputs": [],
|
69 |
+
"source": [
|
70 |
+
"tokenizer = T5Tokenizer.from_pretrained(MODEL_NAME, model_max_length=512)"
|
71 |
+
]
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"cell_type": "code",
|
75 |
+
"execution_count": 5,
|
76 |
+
"id": "8eec35d1",
|
77 |
+
"metadata": {},
|
78 |
+
"outputs": [],
|
79 |
+
"source": [
|
80 |
+
"class T5Model(pl.LightningModule):\n",
|
81 |
+
" \n",
|
82 |
+
" def __init__(self):\n",
|
83 |
+
" super().__init__()\n",
|
84 |
+
" self.model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME, return_dict = True)\n",
|
85 |
+
"\n",
|
86 |
+
" \n",
|
87 |
+
" def forward(self, input_ids, attention_mask, labels=None):\n",
|
88 |
+
" \n",
|
89 |
+
" output = self.model(\n",
|
90 |
+
" input_ids=input_ids, \n",
|
91 |
+
" attention_mask=attention_mask, \n",
|
92 |
+
" labels=labels\n",
|
93 |
+
" )\n",
|
94 |
+
" return output.loss, output.logits\n",
|
95 |
+
" \n",
|
96 |
+
" def training_step(self, batch, batch_idx):\n",
|
97 |
+
"\n",
|
98 |
+
" input_ids = batch[\"input_ids\"]\n",
|
99 |
+
" attention_mask = batch[\"attention_mask\"]\n",
|
100 |
+
" labels= batch[\"target\"]\n",
|
101 |
+
" loss, logits = self(input_ids , attention_mask, labels)\n",
|
102 |
+
"\n",
|
103 |
+
" \n",
|
104 |
+
" self.log(\"train_loss\", loss, prog_bar=True, logger=True)\n",
|
105 |
+
"\n",
|
106 |
+
" return {'loss': loss}\n",
|
107 |
+
" \n",
|
108 |
+
" def validation_step(self, batch, batch_idx):\n",
|
109 |
+
" input_ids = batch[\"input_ids\"]\n",
|
110 |
+
" attention_mask = batch[\"attention_mask\"]\n",
|
111 |
+
" labels= batch[\"target\"]\n",
|
112 |
+
" loss, logits = self(input_ids, attention_mask, labels)\n",
|
113 |
+
"\n",
|
114 |
+
" self.log(\"val_loss\", loss, prog_bar=True, logger=True)\n",
|
115 |
+
" \n",
|
116 |
+
" return {'val_loss': loss}\n",
|
117 |
+
"\n",
|
118 |
+
" def configure_optimizers(self):\n",
|
119 |
+
" return AdamW(self.parameters(), lr=0.0001)"
|
120 |
+
]
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"cell_type": "code",
|
124 |
+
"execution_count": 6,
|
125 |
+
"id": "e9d96844",
|
126 |
+
"metadata": {},
|
127 |
+
"outputs": [
|
128 |
+
{
|
129 |
+
"name": "stderr",
|
130 |
+
"output_type": "stream",
|
131 |
+
"text": [
|
132 |
+
"Lightning automatically upgraded your loaded checkpoint from v1.9.3 to v2.0.2. To apply the upgrade to your files permanently, run `python -m pytorch_lightning.utilities.upgrade_checkpoint --file F:\\Projects & Open_source\\Chatbot_T5_kaggle\\best-model.ckpt`\n"
|
133 |
+
]
|
134 |
+
}
|
135 |
+
],
|
136 |
+
"source": [
|
137 |
+
"train_model = T5Model.load_from_checkpoint('best-model.ckpt',map_location=DEVICE)"
|
138 |
+
]
|
139 |
+
},
|
140 |
+
{
|
141 |
+
"cell_type": "code",
|
142 |
+
"execution_count": 7,
|
143 |
+
"id": "3449943f",
|
144 |
+
"metadata": {},
|
145 |
+
"outputs": [],
|
146 |
+
"source": [
|
147 |
+
"train_model.freeze()"
|
148 |
+
]
|
149 |
+
},
|
150 |
+
{
|
151 |
+
"cell_type": "code",
|
152 |
+
"execution_count": 8,
|
153 |
+
"id": "0e9f1058",
|
154 |
+
"metadata": {},
|
155 |
+
"outputs": [],
|
156 |
+
"source": [
|
157 |
+
"def generate_question(question):\n",
|
158 |
+
"\n",
|
159 |
+
" inputs_encoding = tokenizer(\n",
|
160 |
+
" question,\n",
|
161 |
+
" add_special_tokens=True,\n",
|
162 |
+
" max_length= INPUT_MAX_LEN,\n",
|
163 |
+
" padding = 'max_length',\n",
|
164 |
+
" truncation='only_first',\n",
|
165 |
+
" return_attention_mask=True,\n",
|
166 |
+
" return_tensors=\"pt\"\n",
|
167 |
+
" )\n",
|
168 |
+
"\n",
|
169 |
+
" \n",
|
170 |
+
" generate_ids = train_model.model.generate(\n",
|
171 |
+
" input_ids = inputs_encoding[\"input_ids\"],\n",
|
172 |
+
" attention_mask = inputs_encoding[\"attention_mask\"],\n",
|
173 |
+
" max_length = INPUT_MAX_LEN,\n",
|
174 |
+
" num_beams = 4,\n",
|
175 |
+
" num_return_sequences = 1,\n",
|
176 |
+
" no_repeat_ngram_size=2,\n",
|
177 |
+
" early_stopping=True,\n",
|
178 |
+
" )\n",
|
179 |
+
"\n",
|
180 |
+
" preds = [\n",
|
181 |
+
" tokenizer.decode(gen_id,\n",
|
182 |
+
" skip_special_tokens=True, \n",
|
183 |
+
" clean_up_tokenization_spaces=True)\n",
|
184 |
+
" for gen_id in generate_ids\n",
|
185 |
+
" ]\n",
|
186 |
+
"\n",
|
187 |
+
" return \"\".join(preds)\n"
|
188 |
+
]
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"cell_type": "code",
|
192 |
+
"execution_count": 9,
|
193 |
+
"id": "ee38a88c",
|
194 |
+
"metadata": {},
|
195 |
+
"outputs": [
|
196 |
+
{
|
197 |
+
"name": "stdout",
|
198 |
+
"output_type": "stream",
|
199 |
+
"text": [
|
200 |
+
"Ques: hi, how are you doing?\n",
|
201 |
+
"BOT: i'm so glad you're doing well.\n"
|
202 |
+
]
|
203 |
+
}
|
204 |
+
],
|
205 |
+
"source": [
|
206 |
+
"ques = \"hi, how are you doing?\"\n",
|
207 |
+
"print(\"Ques: \",ques)\n",
|
208 |
+
"print(\"BOT: \",generate_question(ques))"
|
209 |
+
]
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"cell_type": "code",
|
213 |
+
"execution_count": 11,
|
214 |
+
"id": "22aa4414",
|
215 |
+
"metadata": {},
|
216 |
+
"outputs": [
|
217 |
+
{
|
218 |
+
"name": "stdout",
|
219 |
+
"output_type": "stream",
|
220 |
+
"text": [
|
221 |
+
"Running on local URL: http://127.0.0.1:7861\n",
|
222 |
+
"\n",
|
223 |
+
"To create a public link, set `share=True` in `launch()`.\n"
|
224 |
+
]
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"data": {
|
228 |
+
"text/html": [
|
229 |
+
"<div><iframe src=\"http://127.0.0.1:7861/\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
|
230 |
+
],
|
231 |
+
"text/plain": [
|
232 |
+
"<IPython.core.display.HTML object>"
|
233 |
+
]
|
234 |
+
},
|
235 |
+
"metadata": {},
|
236 |
+
"output_type": "display_data"
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"data": {
|
240 |
+
"text/plain": []
|
241 |
+
},
|
242 |
+
"execution_count": 11,
|
243 |
+
"metadata": {},
|
244 |
+
"output_type": "execute_result"
|
245 |
+
}
|
246 |
+
],
|
247 |
+
"source": [
|
248 |
+
"import gradio as gr\n",
|
249 |
+
"import random\n",
|
250 |
+
"import time\n",
|
251 |
+
"\n",
|
252 |
+
"with gr.Blocks() as demo:\n",
|
253 |
+
" chatbot = gr.Chatbot()\n",
|
254 |
+
" gr.Chatbot.style(chatbot,height=400)\n",
|
255 |
+
" msg = gr.Textbox(info=\"Press \\'Enter\\' to send\")\n",
|
256 |
+
" clear = gr.Button(\"Clear\")\n",
|
257 |
+
"\n",
|
258 |
+
" def user(user_message, history):\n",
|
259 |
+
" return \"\", history + [[user_message, None]]\n",
|
260 |
+
"\n",
|
261 |
+
" def bot(history):\n",
|
262 |
+
" bot_message = generate_question(history[-1][0])\n",
|
263 |
+
" history[-1][1] = \"\"\n",
|
264 |
+
" for character in bot_message:\n",
|
265 |
+
" history[-1][1] += character\n",
|
266 |
+
" time.sleep(0.05)\n",
|
267 |
+
" yield history\n",
|
268 |
+
"\n",
|
269 |
+
" msg.submit(user, [msg, chatbot], [msg, chatbot], queue=True).then(\n",
|
270 |
+
" bot, chatbot, chatbot\n",
|
271 |
+
" )\n",
|
272 |
+
" clear.click(lambda: None, None, chatbot, queue=True)\n",
|
273 |
+
"\n",
|
274 |
+
"demo.queue(concurrency_count=2)\n",
|
275 |
+
"demo.launch()"
|
276 |
+
]
|
277 |
+
},
|
278 |
+
{
|
279 |
+
"cell_type": "code",
|
280 |
+
"execution_count": null,
|
281 |
+
"id": "fef38bdc",
|
282 |
+
"metadata": {
|
283 |
+
"scrolled": true
|
284 |
+
},
|
285 |
+
"outputs": [],
|
286 |
+
"source": [
|
287 |
+
"import gradio as gr\n",
|
288 |
+
"import random\n",
|
289 |
+
"import time\n",
|
290 |
+
"\n",
|
291 |
+
"with gr.Blocks() as demo:\n",
|
292 |
+
" chatbot = gr.Chatbot()\n",
|
293 |
+
" msg = gr.Textbox(placeholder='Got any spare time...let\\'s chat!!!')\n",
|
294 |
+
" gr.Textbox.style(msg,show_copy_button=True)\n",
|
295 |
+
" clear = gr.Button(\"Clear\")\n",
|
296 |
+
"\n",
|
297 |
+
" def respond(message, chat_history):\n",
|
298 |
+
" bot_message = generate_question(message)\n",
|
299 |
+
" bot_message = \"**\"+bot_message+\"**\"\n",
|
300 |
+
" chat_history.append((message, bot_message))\n",
|
301 |
+
" time.sleep(1)\n",
|
302 |
+
" return \"\", chat_history\n",
|
303 |
+
"\n",
|
304 |
+
" msg.submit(respond, [msg, chatbot], [msg, chatbot])\n",
|
305 |
+
" clear.click(lambda: None, None, chatbot, queue=False)\n",
|
306 |
+
"\n",
|
307 |
+
"demo.launch()"
|
308 |
+
]
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"cell_type": "code",
|
312 |
+
"execution_count": null,
|
313 |
+
"id": "a86d446a",
|
314 |
+
"metadata": {},
|
315 |
+
"outputs": [],
|
316 |
+
"source": []
|
317 |
+
}
|
318 |
+
],
|
319 |
+
"metadata": {
|
320 |
+
"kernelspec": {
|
321 |
+
"display_name": "Python 3 (ipykernel)",
|
322 |
+
"language": "python",
|
323 |
+
"name": "python3"
|
324 |
+
},
|
325 |
+
"language_info": {
|
326 |
+
"codemirror_mode": {
|
327 |
+
"name": "ipython",
|
328 |
+
"version": 3
|
329 |
+
},
|
330 |
+
"file_extension": ".py",
|
331 |
+
"mimetype": "text/x-python",
|
332 |
+
"name": "python",
|
333 |
+
"nbconvert_exporter": "python",
|
334 |
+
"pygments_lexer": "ipython3",
|
335 |
+
"version": "3.10.1"
|
336 |
+
}
|
337 |
+
},
|
338 |
+
"nbformat": 4,
|
339 |
+
"nbformat_minor": 5
|
340 |
+
}
|
__pycache__/gradio.cpython-310.pyc
ADDED
Binary file (790 Bytes). View file
|
|
app.py
ADDED
@@ -0,0 +1,123 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import T5Tokenizer, T5ForConditionalGeneration
|
2 |
+
|
3 |
+
from transformers import AdamW
|
4 |
+
import pandas as pd
|
5 |
+
import torch
|
6 |
+
import pytorch_lightning as pl
|
7 |
+
from pytorch_lightning.callbacks import ModelCheckpoint
|
8 |
+
from torch.nn.utils.rnn import pad_sequence
|
9 |
+
# from torch.utils.data import Dataset, DataLoader, random_split, RandomSampler, SequentialSampler
|
10 |
+
|
11 |
+
pl.seed_everything(100)
|
12 |
+
|
13 |
+
MODEL_NAME='t5-base'
|
14 |
+
DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
15 |
+
INPUT_MAX_LEN = 128
|
16 |
+
OUTPUT_MAX_LEN = 128
|
17 |
+
|
18 |
+
tokenizer = T5Tokenizer.from_pretrained(MODEL_NAME, model_max_length=512)
|
19 |
+
|
20 |
+
class T5Model(pl.LightningModule):
|
21 |
+
|
22 |
+
def __init__(self):
|
23 |
+
super().__init__()
|
24 |
+
self.model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME, return_dict = True)
|
25 |
+
|
26 |
+
|
27 |
+
def forward(self, input_ids, attention_mask, labels=None):
|
28 |
+
|
29 |
+
output = self.model(
|
30 |
+
input_ids=input_ids,
|
31 |
+
attention_mask=attention_mask,
|
32 |
+
labels=labels
|
33 |
+
)
|
34 |
+
return output.loss, output.logits
|
35 |
+
|
36 |
+
def training_step(self, batch, batch_idx):
|
37 |
+
|
38 |
+
input_ids = batch["input_ids"]
|
39 |
+
attention_mask = batch["attention_mask"]
|
40 |
+
labels= batch["target"]
|
41 |
+
loss, logits = self(input_ids , attention_mask, labels)
|
42 |
+
|
43 |
+
|
44 |
+
self.log("train_loss", loss, prog_bar=True, logger=True)
|
45 |
+
|
46 |
+
return {'loss': loss}
|
47 |
+
|
48 |
+
def validation_step(self, batch, batch_idx):
|
49 |
+
input_ids = batch["input_ids"]
|
50 |
+
attention_mask = batch["attention_mask"]
|
51 |
+
labels= batch["target"]
|
52 |
+
loss, logits = self(input_ids, attention_mask, labels)
|
53 |
+
|
54 |
+
self.log("val_loss", loss, prog_bar=True, logger=True)
|
55 |
+
|
56 |
+
return {'val_loss': loss}
|
57 |
+
|
58 |
+
def configure_optimizers(self):
|
59 |
+
return AdamW(self.parameters(), lr=0.0001)
|
60 |
+
|
61 |
+
train_model = T5Model.load_from_checkpoint('best-model.ckpt',map_location=DEVICE)
|
62 |
+
train_model.freeze()
|
63 |
+
|
64 |
+
def generate_question(question):
|
65 |
+
|
66 |
+
inputs_encoding = tokenizer(
|
67 |
+
question,
|
68 |
+
add_special_tokens=True,
|
69 |
+
max_length= INPUT_MAX_LEN,
|
70 |
+
padding = 'max_length',
|
71 |
+
truncation='only_first',
|
72 |
+
return_attention_mask=True,
|
73 |
+
return_tensors="pt"
|
74 |
+
)
|
75 |
+
|
76 |
+
|
77 |
+
generate_ids = train_model.model.generate(
|
78 |
+
input_ids = inputs_encoding["input_ids"],
|
79 |
+
attention_mask = inputs_encoding["attention_mask"],
|
80 |
+
max_length = INPUT_MAX_LEN,
|
81 |
+
num_beams = 4,
|
82 |
+
num_return_sequences = 1,
|
83 |
+
no_repeat_ngram_size=2,
|
84 |
+
early_stopping=True,
|
85 |
+
)
|
86 |
+
|
87 |
+
preds = [
|
88 |
+
tokenizer.decode(gen_id,
|
89 |
+
skip_special_tokens=True,
|
90 |
+
clean_up_tokenization_spaces=True)
|
91 |
+
for gen_id in generate_ids
|
92 |
+
]
|
93 |
+
|
94 |
+
return "".join(preds)
|
95 |
+
|
96 |
+
import gradio as gr
|
97 |
+
import random
|
98 |
+
import time
|
99 |
+
|
100 |
+
with gr.Blocks() as demo:
|
101 |
+
chatbot = gr.Chatbot()
|
102 |
+
gr.Chatbot.style(chatbot,height=400)
|
103 |
+
msg = gr.Textbox(info="Press \'Enter\' to send")
|
104 |
+
clear = gr.Button("Clear")
|
105 |
+
|
106 |
+
def user(user_message, history):
|
107 |
+
return "", history + [[user_message, None]]
|
108 |
+
|
109 |
+
def bot(history):
|
110 |
+
bot_message = generate_question(history[-1][0])
|
111 |
+
history[-1][1] = ""
|
112 |
+
for character in bot_message:
|
113 |
+
history[-1][1] += character
|
114 |
+
time.sleep(0.05)
|
115 |
+
yield history
|
116 |
+
|
117 |
+
msg.submit(user, [msg, chatbot], [msg, chatbot], queue=True).then(
|
118 |
+
bot, chatbot, chatbot
|
119 |
+
)
|
120 |
+
clear.click(lambda: None, None, chatbot, queue=True)
|
121 |
+
|
122 |
+
demo.queue(concurrency_count=2)
|
123 |
+
demo.launch()
|
best-model.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4d21c48743863e3b7355f0a432cf82b794091fa3ff2ad94c630bb3e9e2975b13
|
3 |
+
size 2675123255
|
requirements.txt
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
transformers==4.27.4
|
2 |
+
pandas==1.5.3
|
3 |
+
torch==2.0.0
|
4 |
+
pytorch-lightning==2.0.2
|
5 |
+
gradio==3.24.1
|