Ubaidbhat/DatabaseTuned
0
1{2 "cells": [3 {4 "cell_type": "code",5 "execution_count": 2,6 "id": "108ad76c-9502-40d3-86b2-7309affc926e",7 "metadata": {},8 "outputs": [],9 "source": [10 "import shutil\n",11 "import requests\n",12 "import sys\n",13 "from typing import Optional, List, Tuple\n",14 "import json\n",15 "from langchain_community.llms import HuggingFaceHub"16 ]17 },18 {19 "cell_type": "code",20 "execution_count": 3,21 "id": "27579e99-9637-4fe5-902c-05c4969ea3aa",22 "metadata": {},23 "outputs": [24 {25 "data": {26 "application/vnd.jupyter.widget-view+json": {27 "model_id": "21ac52ca52fa4910a0e06e3286813c57",28 "version_major": 2,29 "version_minor": 030 },31 "text/plain": [32 "adapter_config.json: 0%| | 0.00/701 [00:00<?, ?B/s]"33 ]34 },35 "metadata": {},36 "output_type": "display_data"37 },38 {39 "name": "stdout",40 "output_type": "stream",41 "text": [42 "HuggingFaceH4/zephyr-7b-beta\n"43 ]44 },45 {46 "data": {47 "application/vnd.jupyter.widget-view+json": {48 "model_id": "d5c705f48d0d40fbb1dd3d409b7f4d7f",49 "version_major": 2,50 "version_minor": 051 },52 "text/plain": [53 "config.json: 0%| | 0.00/638 [00:00<?, ?B/s]"54 ]55 },56 "metadata": {},57 "output_type": "display_data"58 },59 {60 "data": {61 "application/vnd.jupyter.widget-view+json": {62 "model_id": "f209655a6cc743a99cdfa9e0fe40b9f3",63 "version_major": 2,64 "version_minor": 065 },66 "text/plain": [67 "model.safetensors.index.json: 0%| | 0.00/23.9k [00:00<?, ?B/s]"68 ]69 },70 "metadata": {},71 "output_type": "display_data"72 },73 {74 "data": {75 "application/vnd.jupyter.widget-view+json": {76 "model_id": "4b5d3521692d4dee9a6989a37b97d7bf",77 "version_major": 2,78 "version_minor": 079 },80 "text/plain": [81 "Downloading shards: 0%| | 0/8 [00:00<?, ?it/s]"82 ]83 },84 "metadata": {},85 "output_type": "display_data"86 },87 {88 "data": {89 "application/vnd.jupyter.widget-view+json": {90 "model_id": "f66b9cd6a8b04bddb99bbc34fafd6e2e",91 "version_major": 2,92 "version_minor": 093 },94 "text/plain": [95 "model-00001-of-00008.safetensors: 0%| | 0.00/1.89G [00:00<?, ?B/s]"96 ]97 },98 "metadata": {},99 "output_type": "display_data"100 },101 {102 "data": {103 "application/vnd.jupyter.widget-view+json": {104 "model_id": "3469f59063564f2f87022dbb69bc893e",105 "version_major": 2,106 "version_minor": 0107 },108 "text/plain": [109 "model-00002-of-00008.safetensors: 0%| | 0.00/1.95G [00:00<?, ?B/s]"110 ]111 },112 "metadata": {},113 "output_type": "display_data"114 },115 {116 "data": {117 "application/vnd.jupyter.widget-view+json": {118 "model_id": "2b9a7fd2ce904339b08fe41c4f912237",119 "version_major": 2,120 "version_minor": 0121 },122 "text/plain": [123 "model-00003-of-00008.safetensors: 0%| | 0.00/1.98G [00:00<?, ?B/s]"124 ]125 },126 "metadata": {},127 "output_type": "display_data"128 },129 {130 "data": {131 "application/vnd.jupyter.widget-view+json": {132 "model_id": "04318a351ff3418685c823f9cf77f2cd",133 "version_major": 2,134 "version_minor": 0135 },136 "text/plain": [137 "model-00004-of-00008.safetensors: 0%| | 0.00/1.95G [00:00<?, ?B/s]"138 ]139 },140 "metadata": {},141 "output_type": "display_data"142 },143 {144 "data": {145 "application/vnd.jupyter.widget-view+json": {146 "model_id": "87d710cb385e4db2ba8ffaf670841d40",147 "version_major": 2,148 "version_minor": 0149 },150 "text/plain": [151 "model-00005-of-00008.safetensors: 0%| | 0.00/1.98G [00:00<?, ?B/s]"152 ]153 },154 "metadata": {},155 "output_type": "display_data"156 },157 {158 "data": {159 "application/vnd.jupyter.widget-view+json": {160 "model_id": "dcdf133396924c2b84c4a86e5badfabe",161 "version_major": 2,162 "version_minor": 0163 },164 "text/plain": [165 "model-00006-of-00008.safetensors: 0%| | 0.00/1.95G [00:00<?, ?B/s]"166 ]167 },168 "metadata": {},169 "output_type": "display_data"170 },171 {172 "data": {173 "application/vnd.jupyter.widget-view+json": {174 "model_id": "1256b467e1a04d75a4bd969e081dff88",175 "version_major": 2,176 "version_minor": 0177 },178 "text/plain": [179 "model-00007-of-00008.safetensors: 0%| | 0.00/1.98G [00:00<?, ?B/s]"180 ]181 },182 "metadata": {},183 "output_type": "display_data"184 },185 {186 "data": {187 "application/vnd.jupyter.widget-view+json": {188 "model_id": "40e036179dcf43bbac1e2c783638e086",189 "version_major": 2,190 "version_minor": 0191 },192 "text/plain": [193 "model-00008-of-00008.safetensors: 0%| | 0.00/816M [00:00<?, ?B/s]"194 ]195 },196 "metadata": {},197 "output_type": "display_data"198 },199 {200 "data": {201 "application/vnd.jupyter.widget-view+json": {202 "model_id": "ff862b62cc0c4b3fa6c2d49e617d8dd2",203 "version_major": 2,204 "version_minor": 0205 },206 "text/plain": [207 "Loading checkpoint shards: 0%| | 0/8 [00:00<?, ?it/s]"208 ]209 },210 "metadata": {},211 "output_type": "display_data"212 },213 {214 "data": {215 "application/vnd.jupyter.widget-view+json": {216 "model_id": "09733da86a4c429db42a5d99d34f7035",217 "version_major": 2,218 "version_minor": 0219 },220 "text/plain": [221 "generation_config.json: 0%| | 0.00/111 [00:00<?, ?B/s]"222 ]223 },224 "metadata": {},225 "output_type": "display_data"226 },227 {228 "data": {229 "application/vnd.jupyter.widget-view+json": {230 "model_id": "a4951f876a47494aade4c74786c51b5b",231 "version_major": 2,232 "version_minor": 0233 },234 "text/plain": [235 "tokenizer_config.json: 0%| | 0.00/1.43k [00:00<?, ?B/s]"236 ]237 },238 "metadata": {},239 "output_type": "display_data"240 },241 {242 "data": {243 "application/vnd.jupyter.widget-view+json": {244 "model_id": "267a73bd9a494d43899a2ffc75f19953",245 "version_major": 2,246 "version_minor": 0247 },248 "text/plain": [249 "tokenizer.model: 0%| | 0.00/493k [00:00<?, ?B/s]"250 ]251 },252 "metadata": {},253 "output_type": "display_data"254 },255 {256 "data": {257 "application/vnd.jupyter.widget-view+json": {258 "model_id": "3536d4a8071041ec9137bc931366e77b",259 "version_major": 2,260 "version_minor": 0261 },262 "text/plain": [263 "tokenizer.json: 0%| | 0.00/1.80M [00:00<?, ?B/s]"264 ]265 },266 "metadata": {},267 "output_type": "display_data"268 },269 {270 "data": {271 "application/vnd.jupyter.widget-view+json": {272 "model_id": "a66e862cfb374441b2caa7ac5b232435",273 "version_major": 2,274 "version_minor": 0275 },276 "text/plain": [277 "added_tokens.json: 0%| | 0.00/42.0 [00:00<?, ?B/s]"278 ]279 },280 "metadata": {},281 "output_type": "display_data"282 },283 {284 "data": {285 "application/vnd.jupyter.widget-view+json": {286 "model_id": "ca714bb5a9c743219223f2163715e396",287 "version_major": 2,288 "version_minor": 0289 },290 "text/plain": [291 "special_tokens_map.json: 0%| | 0.00/168 [00:00<?, ?B/s]"292 ]293 },294 "metadata": {},295 "output_type": "display_data"296 },297 {298 "data": {299 "application/vnd.jupyter.widget-view+json": {300 "model_id": "a1b80a7393944f06a43db4ba4419c293",301 "version_major": 2,302 "version_minor": 0303 },304 "text/plain": [305 "adapter_model.safetensors: 0%| | 0.00/83.9M [00:00<?, ?B/s]"306 ]307 },308 "metadata": {},309 "output_type": "display_data"310 },311 {312 "name": "stderr",313 "output_type": "stream",314 "text": [315 "/usr/local/lib/python3.10/dist-packages/peft/tuners/lora/bnb.py:272: UserWarning: Merge lora module to 4-bit linear may get different generations due to rounding errors.\n",316 " warnings.warn(\n"317 ]318 }319 ],320 "source": [321 "##Loading the Model to answer questions\n",322 "import torch\n",323 "from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig\n",324 "from peft import PeftModel, PeftConfig\n",325 "\n",326 "\n",327 "peft_model_id = \"Ubaidbhat/zephr_database_finetuned\"\n",328 "config = PeftConfig.from_pretrained(peft_model_id)\n",329 "print(config.base_model_name_or_path)\n",330 "bnb_config = BitsAndBytesConfig(\n",331 " load_in_4bit = True,\n",332 " bnb_4bit_use_double_quant=True,\n",333 " bnb_4bit_quant_type=\"nf4\",\n",334 " bnb_4bit_compute_dtype=torch.bfloat16\n",335 ")\n",336 "\n",337 "d_map = {\"\": torch.cuda.current_device()} if torch.cuda.is_available() else None\n",338 "\n",339 "model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, quantization_config=bnb_config, device_map=d_map)\n",340 "tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)\n",341 "model = PeftModel.from_pretrained(model, peft_model_id)\n",342 "model = model.merge_and_unload()"343 ]344 },345 {346 "cell_type": "code",347 "execution_count": 4,348 "id": "be4a09d2-6cff-4937-a6fb-4f58e4f895ff",349 "metadata": {},350 "outputs": [],351 "source": [352 "##Creating base Model Chain\n",353 "from langchain.llms import HuggingFacePipeline\n",354 "from langchain.prompts import PromptTemplate\n",355 "from transformers import pipeline\n",356 "from langchain_core.output_parsers import StrOutputParser\n",357 "from langchain.chains import LLMChain\n",358 "\n",359 "text_generation_pipeline = pipeline(\n",360 " model=model,\n",361 " tokenizer=tokenizer,\n",362 " task=\"text-generation\",\n",363 " temperature=0.2,\n",364 " do_sample=True,\n",365 " repetition_penalty=1.1,\n",366 " return_full_text=True,\n",367 " max_new_tokens=400,\n",368 " pad_token_id=tokenizer.eos_token_id,\n",369 ")\n",370 "\n",371 "llm = HuggingFacePipeline(pipeline=text_generation_pipeline)\n",372 "\n",373 "prompt_template = \"\"\"\n",374 "<|system|>\n",375 "Answer the question based on your knowledge.\n",376 "</s>\n",377 "<|user|>\n",378 "{question}\n",379 "</s>\n",380 "<|assistant|>\n",381 "\"\"\"\n",382 "\n",383 "prompt = PromptTemplate(\n",384 " input_variables=[\"question\"],\n",385 " template=prompt_template,\n",386 ")\n",387 "\n",388 "llm_chain = prompt | llm | StrOutputParser()\n",389 "\n",390 "def inference(question):\n",391 " llmAnswer = llm_chain.invoke({\"question\": question})\n",392 " llmAnswer = llmAnswer.rstrip()\n",393 " return llmAnswer"394 ]395 },396 {397 "cell_type": "code",398 "execution_count": 8,399 "id": "9e4c410a-5fdf-4b52-96e2-6745b874cb16",400 "metadata": {},401 "outputs": [402 {403 "name": "stdout",404 "output_type": "stream",405 "text": [406 "Running on local URL: http://127.0.0.1:7864\n",407 "Running on public URL: https://98d1369a3563cae95e.gradio.live\n",408 "\n",409 "This share link expires in 72 hours. For free permanent hosting and GPU upgrades, run `gradio deploy` from Terminal to deploy to Spaces (https://huggingface.co/spaces)\n"410 ]411 },412 {413 "data": {414 "text/html": [415 "<div><iframe src=\"https://98d1369a3563cae95e.gradio.live\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"416 ],417 "text/plain": [418 "<IPython.core.display.HTML object>"419 ]420 },421 "metadata": {},422 "output_type": "display_data"423 },424 {425 "data": {426 "text/plain": []427 },428 "execution_count": 8,429 "metadata": {},430 "output_type": "execute_result"431 }432 ],433 "source": [434 "import gradio as gr\n",435 "from langchain_core.runnables import RunnablePassthrough\n",436 "\n",437 "def predict(question):\n",438 " return question\n",439 " \n",440 "pred = gr.Interface(\n",441 " fn=predict,\n",442 " inputs=[\n",443 " gr.Textbox(label=\"Question\", value = \"Your Question here......\"),\n",444 " ],\n",445 " outputs=\"text\",\n",446 " title=\"Finetuned Zephr Model in the Database Management Domaain\"\n",447 ")\n",448 "\n",449 "pred.launch(share=True)"450 ]451 },452 {453 "cell_type": "code",454 "execution_count": null,455 "id": "c7be3ade-8cd6-448b-b608-0182a9743315",456 "metadata": {},457 "outputs": [],458 "source": []459 }460 ],461 "metadata": {462 "kernelspec": {463 "display_name": "Python 3 (ipykernel)",464 "language": "python",465 "name": "python3"466 },467 "language_info": {468 "codemirror_mode": {469 "name": "ipython",470 "version": 3471 },472 "file_extension": ".py",473 "mimetype": "text/x-python",474 "name": "python",475 "nbconvert_exporter": "python",476 "pygments_lexer": "ipython3",477 "version": "3.10.12"478 }479 },480 "nbformat": 4,481 "nbformat_minor": 5482}483