{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "For text TextFileToDocument\n", "for pdf PyPDFToDocument" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "/unity/f2/asugandhi/Downloads/LLM_Playground\n", "\n" ] }, { "ename": "ValueError", "evalue": "Input batch_size not found in component PdfFileConverter.", "output_type": "error", "traceback": [ "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", "\u001b[0;31mValueError\u001b[0m Traceback (most recent call last)", "Cell \u001b[0;32mIn[21], line 27\u001b[0m\n\u001b[1;32m 25\u001b[0m pipeline\u001b[38;5;241m.\u001b[39mconnect(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mPdfFileConverter\u001b[39m\u001b[38;5;124m\"\u001b[39m,\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mPdfwriter_chroma\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m 26\u001b[0m pipeline\u001b[38;5;241m.\u001b[39mconnect(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mTextFileConverter\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mwriter_chroma\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[0;32m---> 27\u001b[0m \u001b[43mpipeline\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m 28\u001b[0m \u001b[43m \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mPdfFileConverter\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43msources\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mfile_paths\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mbatch_size\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m1\u001b[39;49m\u001b[43m}\u001b[49m\u001b[43m}\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 29\u001b[0m \u001b[43m \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mTextFileConverter\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43m{\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43msources\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[43mfile_paths\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mbatch_size\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m:\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m1\u001b[39;49m\u001b[43m}\u001b[49m\u001b[43m}\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 30\u001b[0m \u001b[43m)\u001b[49m\n\u001b[1;32m 33\u001b[0m querying \u001b[38;5;241m=\u001b[39m Pipeline()\n\u001b[1;32m 34\u001b[0m reader \u001b[38;5;241m=\u001b[39m ExtractiveReader(model\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mdeepset/roberta-base-squad2-distilled\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n", "File \u001b[0;32m/conda/asugandhi/miniconda3/envs/RAGAPP/lib/python3.10/site-packages/haystack/core/pipeline/pipeline.py:688\u001b[0m, in \u001b[0;36mPipeline.run\u001b[0;34m(self, data, debug)\u001b[0m\n\u001b[1;32m 682\u001b[0m logger\u001b[38;5;241m.\u001b[39mwarning(\n\u001b[1;32m 683\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mInputs \u001b[39m\u001b[38;5;132;01m%s\u001b[39;00m\u001b[38;5;124m were not matched to any component inputs, please check your run parameters.\u001b[39m\u001b[38;5;124m\"\u001b[39m,\n\u001b[1;32m 684\u001b[0m \u001b[38;5;28mlist\u001b[39m(unresolved_inputs\u001b[38;5;241m.\u001b[39mkeys()),\n\u001b[1;32m 685\u001b[0m )\n\u001b[1;32m 687\u001b[0m \u001b[38;5;66;03m# Raise if input is malformed in some way\u001b[39;00m\n\u001b[0;32m--> 688\u001b[0m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_validate_input\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 689\u001b[0m \u001b[38;5;66;03m# NOTE: The above NOTE and TODO are technically not true.\u001b[39;00m\n\u001b[1;32m 690\u001b[0m \u001b[38;5;66;03m# This implementation of run supports only the first format, but the second format is actually\u001b[39;00m\n\u001b[1;32m 691\u001b[0m \u001b[38;5;66;03m# never received by this method. It's handled by the `run()` method of the `Pipeline` class\u001b[39;00m\n\u001b[0;32m (...)\u001b[0m\n\u001b[1;32m 695\u001b[0m \u001b[38;5;66;03m# deepcopying the inputs prevents the Pipeline run logic from being altered unexpectedly\u001b[39;00m\n\u001b[1;32m 696\u001b[0m \u001b[38;5;66;03m# when the same input reference is passed to multiple components.\u001b[39;00m\n\u001b[1;32m 697\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m component_name, component_inputs \u001b[38;5;129;01min\u001b[39;00m data\u001b[38;5;241m.\u001b[39mitems():\n", "File \u001b[0;32m/conda/asugandhi/miniconda3/envs/RAGAPP/lib/python3.10/site-packages/haystack/core/pipeline/pipeline.py:594\u001b[0m, in \u001b[0;36mPipeline._validate_input\u001b[0;34m(self, data)\u001b[0m\n\u001b[1;32m 592\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m input_name \u001b[38;5;129;01min\u001b[39;00m component_inputs\u001b[38;5;241m.\u001b[39mkeys():\n\u001b[1;32m 593\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m input_name \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;129;01min\u001b[39;00m instance\u001b[38;5;241m.\u001b[39m__haystack_input__\u001b[38;5;241m.\u001b[39m_sockets_dict:\n\u001b[0;32m--> 594\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mInput \u001b[39m\u001b[38;5;132;01m{\u001b[39;00minput_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m not found in component \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mcomponent_name\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m.\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m 596\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m component_name \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mgraph\u001b[38;5;241m.\u001b[39mnodes:\n\u001b[1;32m 597\u001b[0m instance \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mgraph\u001b[38;5;241m.\u001b[39mnodes[component_name][\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124minstance\u001b[39m\u001b[38;5;124m\"\u001b[39m]\n", "\u001b[0;31mValueError\u001b[0m: Input batch_size not found in component PdfFileConverter." ] } ], "source": [ "import os\n", "from haystack import Pipeline, Document\n", "from haystack.components.converters import TextFileToDocument, PyPDFToDocument\n", "from haystack.components.writers import DocumentWriter\n", "from haystack.components.readers import ExtractiveReader\n", "from haystack_integrations.document_stores.chroma import ChromaDocumentStore\n", "from haystack_integrations.components.retrievers.chroma import ChromaQueryTextRetriever\n", "from pathlib import Path\n", "HERE = Path(os.getcwd())\n", "print(HERE)\n", "\n", "data_path = HERE / \"data\"\n", "file_paths = [data_path / Path(name) for name in os.listdir(\"data\")]\n", "print()\n", "chroma_store = ChromaDocumentStore()\n", "# Resolve the absolute path\n", "# absolute_file_path = file_path.resolve()\n", "# print(absolute_file_path)\n", "pipeline = Pipeline()\n", "pipeline.add_component(\"PdfFileConverter\", PyPDFToDocument())\n", "pipeline.add_component(\"TextFileConverter\", TextFileToDocument())\n", "pipeline.add_component(\"Pdfwriter_chroma\", DocumentWriter(document_store=chroma_store))\n", "pipeline.add_component(\"writer_chroma\", DocumentWriter(document_store=chroma_store))\n", "\n", "pipeline.connect(\"PdfFileConverter\",\"Pdfwriter_chroma\")\n", "pipeline.connect(\"TextFileConverter\", \"writer_chroma\")\n", "pipeline.run(\n", " {\"PdfFileConverter\": {\"sources\": file_paths, \"batch_size\": 1}},\n", " {\"TextFileConverter\": {\"sources\": file_paths, \"batch_size\": 1}},\n", ")\n", " \n", " \n", "querying = Pipeline()\n", "reader = ExtractiveReader(model=\"deepset/roberta-base-squad2-distilled\")\n", "querying.add_component(\"retriever\", ChromaQueryTextRetriever(chroma_store))\n", "querying.add_component(\"reader\",reader)\n", "results = querying.run({\"retriever\": {\"query\": \"Vishwam\", \"top_k\": 3}})\n" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "/unity/f2/asugandhi/Downloads/LLM_Playground\n", "{'reader': {'answers': [ExtractedAnswer(query='Who is Aditya?', score=0.6858945488929749, data='Software Engineer', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=31, end=48), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.627069890499115, data='Sugandhi', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=7, end=15), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.5672385096549988, data='Software Engineer', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=4616, end=4633), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.5219605565071106, data='software engineer', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=4961, end=4978), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.5016087889671326, data='Sugandhi', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=4592, end=4600), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.44805991649627686, data='Web Developer Intern', document=Document(id=ce02ebe3fa97972f0c76b2c175f658873b2d0e19987e9cbc38dcacb25b8ebdba, content: 'Aditya Sugandhi's journey as a Software Engineer is characterized by a deep commitment to excellence...', meta: {'file_path': '/unity/f2/asugandhi/Downloads/LLM_Playground/data/Aditya_train.txt', 'source_id': '228fb178549cb032d67e0b2da301131f48d7c88c814b6d6920c92727b1c8f5fd'}, score: 1.191292405128479, embedding: vector of size 384), context=None, document_offset=ExtractedAnswer.Span(start=3343, end=3363), context_offset=None, meta={}), ExtractedAnswer(query='Who is Aditya?', score=0.0066661882226549205, data=None, document=None, context=None, document_offset=None, context_offset=None, meta={})]}}\n" ] } ], "source": [ "from pathlib import Path\n", "import os\n", "from haystack import Pipeline\n", "from haystack.components.embedders import SentenceTransformersDocumentEmbedder\n", "from haystack.components.converters import PyPDFToDocument, TextFileToDocument\n", "from haystack.components.preprocessors import DocumentCleaner, DocumentSplitter\n", "from haystack.components.readers import ExtractiveReader\n", "from haystack.components.routers import FileTypeRouter\n", "from haystack.components.joiners import DocumentJoiner\n", "from haystack.components.writers import DocumentWriter\n", "from haystack_integrations.document_stores.chroma import ChromaDocumentStore\n", "from haystack_integrations.components.retrievers.chroma import ChromaQueryTextRetriever\n", "\n", "HERE = Path(os.getcwd())\n", "print(HERE)\n", "\n", "data_path = HERE / \"data\"\n", "file_paths = [str(data_path / name) for name in os.listdir(data_path)]\n", "\n", "chroma_store = ChromaDocumentStore()\n", "\n", "pipeline = Pipeline()\n", "pipeline.add_component(\"FileTypeRouter\", FileTypeRouter(mime_types=[\"text/plain\", \"application/pdf\"]))\n", "pipeline.add_component(\"TextFileConverter\", TextFileToDocument())\n", "pipeline.add_component(\"PdfFileConverter\", PyPDFToDocument())\n", "pipeline.add_component(\"Joiner\", DocumentJoiner())\n", "pipeline.add_component(\"Cleaner\", DocumentCleaner())\n", "pipeline.add_component(\"Splitter\", DocumentSplitter(split_by=\"sentence\", split_length=250, split_overlap=30))\n", "# pipeline.add_component(\"Embedder\", SentenceTransformersDocumentEmbedder(model=\"sentence-transformers/all-MiniLM-L6-v2\"))\n", "pipeline.add_component(\"Writer\", DocumentWriter(document_store=chroma_store))\n", "\n", "pipeline.connect(\"FileTypeRouter.text/plain\", \"TextFileConverter.sources\")\n", "pipeline.connect(\"FileTypeRouter.application/pdf\", \"PdfFileConverter.sources\")\n", "pipeline.connect(\"TextFileConverter.documents\", \"Joiner.documents\")\n", "pipeline.connect(\"PdfFileConverter.documents\", \"Joiner.documents\")\n", "pipeline.connect(\"Joiner.documents\", \"Cleaner.documents\")\n", "pipeline.connect(\"Cleaner.documents\", \"Splitter.documents\")\n", "pipeline.connect(\"Splitter.documents\", \"Writer.documents\")\n", "# pipeline.connect(\"Embedder.documents\", \"Writer.documents\")\n", "\n", "pipeline.run(\n", " {\"FileTypeRouter\": {\"sources\": file_paths}},\n", ")\n", "\n", "# Querying pipeline\n", "querying = Pipeline()\n", "reader = ExtractiveReader(model=\"deepset/roberta-base-squad2-distilled\")\n", "querying.add_component(\"retriever\", ChromaQueryTextRetriever(chroma_store))\n", "querying.add_component(\"reader\", reader)\n", "querying.connect(\"retriever\", \"reader\")\n", "query = \"Who is Aditya?\"\n", "input_data = {\n", " \"retriever\": {\"query\": query, \"top_k\": 1},\n", " \"reader\": {\"query\": query},\n", " # Use 'max_tokens' instead of 'max_new_tokens'\n", " }\n", "results = querying.run(input_data)\n", "print(results)\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#DON'T RUN" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "who is Aditya?\n", "{'llm': {'replies': ['Aditya Sugandhi is a Software Engineer with a strong foundation in both theoretical knowledge and practical application, known for his commitment to excellence, passion for technological advancements, and dedication to pushing boundaries in software development. He has experience in various roles such as a Research Assistant, Full Stack Developer, Customer Service Executive, and Web Developer Intern. Aditya is currently pursuing a Master’s of Science in Computer Science at Florida State University and holds a Bachelor of Technology in Computer Science Engineering from SRM University. He is characterized by technical excellence, innovation, and a holistic understanding of software development. Aditya enjoys spending time with his friends SAS, Hunterr, MF, and Rocco.'], 'meta': [{'model': 'gpt-3.5-turbo-0125', 'index': 0, 'finish_reason': 'stop', 'usage': {'completion_tokens': 138, 'prompt_tokens': 917, 'total_tokens': 1055}}]}}\n" ] } ], "source": [ "from haystack import Pipeline\n", "from haystack.utils import Secret\n", "from haystack_integrations.components.retrievers.chroma import ChromaQueryTextRetriever\n", "from haystack.components.readers import ExtractiveReader\n", "from haystack.components.generators import GPTGenerator\n", "from haystack.components.builders.prompt_builder import PromptBuilder\n", "from haystack.components.generators import OpenAIGenerator\n", "\n", "template = \"\"\"\n", "Answer all the questions in the following format and based on Aditya.\n", "\n", "Context:\n", "{% for doc in documents %}\n", " {{ doc.content }}\n", "{% endfor %}\n", "Question: {{question}}\n", "Answer:\n", "\"\"\"\n", "\n", "prompt_builder = PromptBuilder(template=template)\n", "retriever = ChromaQueryTextRetriever(document_store = chroma_store)\n", "#ExtractiveReader to extract answers from the relevant context\n", "api_key = Secret.from_token(\"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\")\n", "llm = OpenAIGenerator(model=\"gpt-3.5-turbo-0125\",api_key=api_key)\n", "reader = ExtractiveReader(model=\"deepset/roberta-base-squad2-distilled\")\n", "\n", "extractive_qa_pipeline = Pipeline()\n", "extractive_qa_pipeline.add_component(\"retriever\", retriever)\n", "# extractive_qa_pipeline.add_component(\"reader\",reader)\n", "extractive_qa_pipeline.add_component(instance=prompt_builder, name=\"prompt_builder\")\n", "extractive_qa_pipeline.add_component(\"llm\", llm)\n", "\n", "\n", "# extractive_qa_pipeline.connect(\"retriever\", \"reader\")\n", "extractive_qa_pipeline.connect(\"retriever\", \"prompt_builder.documents\")\n", "extractive_qa_pipeline.connect(\"prompt_builder\", \"llm\")\n", "\n", "\n", "query = \"who is Aditya?\"\n", "print(query)\n", "# Define the input data for the pipeline components\n", "input_data = {\n", " \"retriever\": {\"query\": query, \"top_k\": 1},\n", " \"prompt_builder\": {\"question\": query},\n", " # Use 'max_tokens' instead of 'max_new_tokens'\n", "}\n", "\n", "# Run the pipeline with the updated input data\n", "results = extractive_qa_pipeline.run(input_data)\n", "print(results)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "llama_model_loader: loaded meta data with 23 key-value pairs and 291 tensors from openchat-3.5-1210.Q3_K_S.ggml (version GGUF V3 (latest))\n", "llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.\n", "llama_model_loader: - kv 0: general.architecture str = llama\n", "llama_model_loader: - kv 1: general.name str = openchat_openchat-3.5-1210\n", "llama_model_loader: - kv 2: llama.context_length u32 = 8192\n", "llama_model_loader: - kv 3: llama.embedding_length u32 = 4096\n", "llama_model_loader: - kv 4: llama.block_count u32 = 32\n", "llama_model_loader: - kv 5: llama.feed_forward_length u32 = 14336\n", "llama_model_loader: - kv 6: llama.rope.dimension_count u32 = 128\n", "llama_model_loader: - kv 7: llama.attention.head_count u32 = 32\n", "llama_model_loader: - kv 8: llama.attention.head_count_kv u32 = 8\n", "llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010\n", "llama_model_loader: - kv 10: llama.rope.freq_base f32 = 10000.000000\n", "llama_model_loader: - kv 11: general.file_type u32 = 11\n", "llama_model_loader: - kv 12: tokenizer.ggml.model str = llama\n", "llama_model_loader: - kv 13: tokenizer.ggml.tokens arr[str,32002] = [\"\", \"\", \"\", \"<0x00>\", \"<...\n", "llama_model_loader: - kv 14: tokenizer.ggml.scores arr[f32,32002] = [0.000000, 0.000000, 0.000000, 0.0000...\n", "llama_model_loader: - kv 15: tokenizer.ggml.token_type arr[i32,32002] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...\n", "llama_model_loader: - kv 16: tokenizer.ggml.bos_token_id u32 = 1\n", "llama_model_loader: - kv 17: tokenizer.ggml.eos_token_id u32 = 32000\n", "llama_model_loader: - kv 18: tokenizer.ggml.padding_token_id u32 = 0\n", "llama_model_loader: - kv 19: tokenizer.ggml.add_bos_token bool = true\n", "llama_model_loader: - kv 20: tokenizer.ggml.add_eos_token bool = false\n", "llama_model_loader: - kv 21: tokenizer.chat_template str = {{ bos_token }}{% for message in mess...\n", "llama_model_loader: - kv 22: general.quantization_version u32 = 2\n", "llama_model_loader: - type f32: 65 tensors\n", "llama_model_loader: - type q3_K: 225 tensors\n", "llama_model_loader: - type q6_K: 1 tensors\n", "llm_load_vocab: special tokens definition check successful ( 261/32002 ).\n", "llm_load_print_meta: format = GGUF V3 (latest)\n", "llm_load_print_meta: arch = llama\n", "llm_load_print_meta: vocab type = SPM\n", "llm_load_print_meta: n_vocab = 32002\n", "llm_load_print_meta: n_merges = 0\n", "llm_load_print_meta: n_ctx_train = 8192\n", "llm_load_print_meta: n_embd = 4096\n", "llm_load_print_meta: n_head = 32\n", "llm_load_print_meta: n_head_kv = 8\n", "llm_load_print_meta: n_layer = 32\n", "llm_load_print_meta: n_rot = 128\n", "llm_load_print_meta: n_embd_head_k = 128\n", "llm_load_print_meta: n_embd_head_v = 128\n", "llm_load_print_meta: n_gqa = 4\n", "llm_load_print_meta: n_embd_k_gqa = 1024\n", "llm_load_print_meta: n_embd_v_gqa = 1024\n", "llm_load_print_meta: f_norm_eps = 0.0e+00\n", "llm_load_print_meta: f_norm_rms_eps = 1.0e-05\n", "llm_load_print_meta: f_clamp_kqv = 0.0e+00\n", "llm_load_print_meta: f_max_alibi_bias = 0.0e+00\n", "llm_load_print_meta: n_ff = 14336\n", "llm_load_print_meta: n_expert = 0\n", "llm_load_print_meta: n_expert_used = 0\n", "llm_load_print_meta: rope scaling = linear\n", "llm_load_print_meta: freq_base_train = 10000.0\n", "llm_load_print_meta: freq_scale_train = 1\n", "llm_load_print_meta: n_yarn_orig_ctx = 8192\n", "llm_load_print_meta: rope_finetuned = unknown\n", "llm_load_print_meta: model type = 7B\n", "llm_load_print_meta: model ftype = Q3_K - Small\n", "llm_load_print_meta: model params = 7.24 B\n", "llm_load_print_meta: model size = 2.95 GiB (3.50 BPW) \n", "llm_load_print_meta: general.name = openchat_openchat-3.5-1210\n", "llm_load_print_meta: BOS token = 1 ''\n", "llm_load_print_meta: EOS token = 32000 '<|end_of_turn|>'\n", "llm_load_print_meta: UNK token = 0 ''\n", "llm_load_print_meta: PAD token = 0 ''\n", "llm_load_print_meta: LF token = 13 '<0x0A>'\n", "llm_load_tensors: ggml ctx size = 0.56 MiB\n", "llm_load_tensors: offloading 32 repeating layers to GPU\n", "llm_load_tensors: offloading non-repeating layers to GPU\n", "llm_load_tensors: offloaded 33/33 layers to GPU\n", "llm_load_tensors: CPU buffer size = 53.71 MiB\n", "llm_load_tensors: CUDA0 buffer size = 804.66 MiB\n", "llm_load_tensors: CUDA1 buffer size = 715.25 MiB\n", "llm_load_tensors: CUDA2 buffer size = 715.25 MiB\n", "llm_load_tensors: CUDA3 buffer size = 728.40 MiB\n", ".................................................................................................\n", "llama_new_context_with_model: n_ctx = 10000\n", "llama_new_context_with_model: freq_base = 10000.0\n", "llama_new_context_with_model: freq_scale = 1\n", "llama_kv_cache_init: CUDA0 KV buffer size = 351.56 MiB\n", "llama_kv_cache_init: CUDA1 KV buffer size = 312.50 MiB\n", "llama_kv_cache_init: CUDA2 KV buffer size = 312.50 MiB\n", "llama_kv_cache_init: CUDA3 KV buffer size = 273.44 MiB\n", "llama_new_context_with_model: KV self size = 1250.00 MiB, K (f16): 625.00 MiB, V (f16): 625.00 MiB\n" ] }, { "ename": "", "evalue": "", "output_type": "error", "traceback": [ "\u001b[1;31mThe Kernel crashed while executing code in the current cell or a previous cell. \n", "\u001b[1;31mPlease review the code in the cell(s) to identify a possible cause of the failure. \n", "\u001b[1;31mClick here for more info. \n", "\u001b[1;31mView Jupyter log for further details." ] } ], "source": [ "from haystack import Pipeline\n", "from haystack.utils import Secret\n", "from haystack_integrations.components.retrievers.chroma import ChromaQueryTextRetriever\n", "from haystack_integrations.components.generators.llama_cpp import LlamaCppGenerator\n", "from haystack.components.readers import ExtractiveReader\n", "from haystack.components.generators import GPTGenerator\n", "from haystack.components.builders.prompt_builder import PromptBuilder\n", "from haystack.components.builders.answer_builder import AnswerBuilder\n", "from haystack.components.generators import OpenAIGenerator\n", "\n", "\n", "\n", "\n", "template = \"\"\"\n", "Answer all the questions in the following format and based on Aditya \n", "and if not found generate answer accordingly using the given information.\n", "\n", "Context:\n", "{% for doc in documents %}\n", "{{ doc.content }}\n", "{% endfor %}\n", "Question: {{question}}\n", "Answer:\n", "\"\"\"\n", "\n", "prompt_builder = PromptBuilder(template=template)\n", "retriever = ChromaQueryTextRetriever(document_store = chroma_store)\n", "#ExtractiveReader to extract answers from the relevant context\n", "\n", "llm = LlamaCppGenerator(\n", "model_path=\"openchat-3.5-1210.Q3_K_S.ggml\", \n", "n_ctx=10000,\n", "n_batch=256,\n", "model_kwargs={\"n_gpu_layers\": -1},\n", "generation_kwargs={\"max_tokens\": 250, \"temperature\": 0.9},\n", ")\n", "\n", "reader = ExtractiveReader(model=\"deepset/roberta-base-squad2-distilled\",)\n", "\n", "extractive_qa_pipeline = Pipeline()\n", "extractive_qa_pipeline.add_component(\"retriever\", ChromaQueryTextRetriever(chroma_store))\n", "# extractive_qa_pipeline.add_component(\"reader\",reader)\n", "extractive_qa_pipeline.add_component(instance=prompt_builder, name=\"prompt_builder\")\n", "extractive_qa_pipeline.add_component(\"llm\", llm)\n", "extractive_qa_pipeline.add_component(instance=AnswerBuilder(), name=\"answer_builder\")\n", "\n", "# extractive_qa_pipeline.connect(\"retriever.documents\", \"reader\")\n", "extractive_qa_pipeline.connect(\"retriever\", \"prompt_builder.documents\") \n", "extractive_qa_pipeline.connect(\"prompt_builder\", \"llm\")\n", "extractive_qa_pipeline.connect(\"llm.replies\", \"answer_builder.replies\")\n", "extractive_qa_pipeline.connect(\"retriever\", \"answer_builder.documents\")\n", "\n", "query = \"who is Aditya did Aditya Pursued his Masters from?\"\n", "\n", "# Define the input data for the pipeline components\n", "input_data = {\n", " \"retriever\": {\"query\": query, \"top_k\": 3},\n", " # \"reader\": {\"query\": query},\n", " \"prompt_builder\": {\"question\": query},\n", " \"answer_builder\": {\"query\": query},\n", " # Use 'max_tokens' instead of 'max_new_tokens'\n", "}\n", "\n", "# Run the pipeline with the updated input data\n", "results = extractive_qa_pipeline.run(input_data)\n", "\n", " \n", " " ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " Aditya pursued his Masters from Florida State University.\n" ] } ], "source": [ "# Assuming results is the dictionary containing the output\n", "generated_content = results['llm']['meta'][0]['choices'][0]['text']\n", "#print(results)\n", "# Print the generated content\n", "print(generated_content)\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "RAGAPP", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.13" } }, "nbformat": 4, "nbformat_minor": 2 }