413 lines
18 KiB
Text
413 lines
18 KiB
Text
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 21,
|
|
"id": "7018890b-b220-48a3-a1b8-d8f6c5483f6c",
|
|
"metadata": {
|
|
"editable": true,
|
|
"slideshow": {
|
|
"slide_type": ""
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
|
|
"To disable this warning, you can either:\n",
|
|
"\t- Avoid using `tokenizers` before the fork if possible\n",
|
|
"\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Cloning into 'gpt-neo-125M'...\n",
|
|
"remote: Enumerating objects: 65, done.\u001b[K\n",
|
|
"remote: Counting objects: 100% (5/5), done.\u001b[K\n",
|
|
"remote: Compressing objects: 100% (5/5), done.\u001b[K\n",
|
|
"remote: Total 65 (delta 1), reused 0 (delta 0), pack-reused 60 (from 1)\u001b[K\n",
|
|
"Unpacking objects: 100% (65/65), 1.11 MiB | 9.13 MiB/s, done.\n",
|
|
"Filtering content: 100% (4/4), 1.93 GiB | 54.88 MiB/s, done.\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"!git clone https://huggingface.co/EleutherAI/gpt-neo-125M"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "36078c20-4185-4f45-84f2-d3e95e1dcac9",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/nvfuser-0.2.13a0+0d33366-py3.12-linux-x86_64.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/dill-0.3.9-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/lightning_utilities-0.11.8-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/opt_einsum-3.4.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/igraph-0.11.8-py3.12-linux-x86_64.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/lightning_thunder-0.2.0.dev0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/texttable-1.7.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/looseversion-1.3.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\n",
|
|
"\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m24.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m25.1.1\u001b[0m\n",
|
|
"\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython -m pip install --upgrade pip\u001b[0m\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"!pip install accelerate>=0.26.0"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 6,
|
|
"id": "11d0a83e-391e-4590-9db6-6a2a3c140ae2",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
|
|
"To disable this warning, you can either:\n",
|
|
"\t- Avoid using `tokenizers` before the fork if possible\n",
|
|
"\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/nvfuser-0.2.13a0+0d33366-py3.12-linux-x86_64.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/dill-0.3.9-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/lightning_utilities-0.11.8-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/opt_einsum-3.4.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/igraph-0.11.8-py3.12-linux-x86_64.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/lightning_thunder-0.2.0.dev0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/texttable-1.7.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\u001b[33mDEPRECATION: Loading egg at /usr/local/lib/python3.12/dist-packages/looseversion-1.3.0-py3.12.egg is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to use pip for package installation. Discussion can be found at https://github.com/pypa/pip/issues/12330\u001b[0m\u001b[33m\n",
|
|
"\u001b[0m\n",
|
|
"\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m24.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m25.1.1\u001b[0m\n",
|
|
"\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython -m pip install --upgrade pip\u001b[0m\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"!pip install datasets transformers torch gradio --quiet"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "7057ed8b-c98c-446e-b40b-2de8a800cb7c",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"/usr/local/lib/python3.12/dist-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
|
|
" from .autonotebook import tqdm as notebook_tqdm\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from transformers import AutoTokenizer, AutoModelForCausalLM\n",
|
|
"import torch\n",
|
|
"\n",
|
|
"# Ścieżka do lokalnego modelu\n",
|
|
"local_model_path = \"./gpt-neo-125M\"\n",
|
|
"\n",
|
|
"# Załaduj tokenizer i model w trybie offline\n",
|
|
"tokenizer = AutoTokenizer.from_pretrained(local_model_path, local_files_only=True)\n",
|
|
"tokenizer.pad_token = tokenizer.eos_token\n",
|
|
"model = AutoModelForCausalLM.from_pretrained(local_model_path, local_files_only=True)\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 17,
|
|
"id": "3575956e-aa32-4173-8678-9a34b838ce3c",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Map: 100%|██████████| 20/20 [00:00<00:00, 1442.80 examples/s]"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Dataset({\n",
|
|
" features: ['input_ids', 'attention_mask'],\n",
|
|
" num_rows: 20\n",
|
|
"})\n",
|
|
"Liczba próbek po tokenizacji: 20\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from datasets import Dataset\n",
|
|
"from transformers import AutoTokenizer, AutoModelForCausalLM, TrainingArguments, Trainer, DataCollatorForLanguageModeling\n",
|
|
"\n",
|
|
"# Ścieżka do lokalnego modelu\n",
|
|
"local_model_path = \"./gpt-neo-125M\"\n",
|
|
"\n",
|
|
"# Tokenizer\n",
|
|
"tokenizer = AutoTokenizer.from_pretrained(local_model_path, local_files_only=True)\n",
|
|
"tokenizer.pad_token = tokenizer.eos_token # <-- wymagane do paddingu\n",
|
|
"\n",
|
|
"# Wczytanie danych z pliku\n",
|
|
"with open(\"data.txt\", encoding=\"utf-8\") as f:\n",
|
|
" lines = [line.strip() for line in f if line.strip()] # Usuwamy puste linie\n",
|
|
"\n",
|
|
"# Budujemy Dataset z listy słowników\n",
|
|
"data = [{\"text\": line} for line in lines]\n",
|
|
"raw_dataset = Dataset.from_list(data)\n",
|
|
"\n",
|
|
"# Funkcja tokenizująca\n",
|
|
"def tokenize_function(example):\n",
|
|
" return tokenizer(\n",
|
|
" example[\"text\"],\n",
|
|
" truncation=True,\n",
|
|
" max_length=128,\n",
|
|
" padding=\"max_length\"\n",
|
|
" )\n",
|
|
"\n",
|
|
"# Tokenizacja datasetu\n",
|
|
"tokenized_dataset = raw_dataset.map(tokenize_function, batched=True, remove_columns=[\"text\"])\n",
|
|
"\n",
|
|
"# Sprawdź czy dane są OK\n",
|
|
"print(tokenized_dataset)\n",
|
|
"print(f\"Liczba próbek po tokenizacji: {len(tokenized_dataset)}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 18,
|
|
"id": "c5eedf83-e003-434e-af7c-a57d6efdd120",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"/tmp/ipykernel_6455/2971221750.py:23: FutureWarning: `tokenizer` is deprecated and will be removed in version 5.0.0 for `Trainer.__init__`. Use `processing_class` instead.\n",
|
|
" trainer = Trainer(\n"
|
|
]
|
|
},
|
|
{
|
|
"data": {
|
|
"text/html": [
|
|
"\n",
|
|
" <div>\n",
|
|
" \n",
|
|
" <progress value='60' max='60' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
|
|
" [60/60 00:12, Epoch 3/3]\n",
|
|
" </div>\n",
|
|
" <table border=\"1\" class=\"dataframe\">\n",
|
|
" <thead>\n",
|
|
" <tr style=\"text-align: left;\">\n",
|
|
" <th>Step</th>\n",
|
|
" <th>Training Loss</th>\n",
|
|
" </tr>\n",
|
|
" </thead>\n",
|
|
" <tbody>\n",
|
|
" <tr>\n",
|
|
" <td>50</td>\n",
|
|
" <td>2.103000</td>\n",
|
|
" </tr>\n",
|
|
" </tbody>\n",
|
|
"</table><p>"
|
|
],
|
|
"text/plain": [
|
|
"<IPython.core.display.HTML object>"
|
|
]
|
|
},
|
|
"metadata": {},
|
|
"output_type": "display_data"
|
|
},
|
|
{
|
|
"data": {
|
|
"text/plain": [
|
|
"('./gpt-neo-finetuned/tokenizer_config.json',\n",
|
|
" './gpt-neo-finetuned/special_tokens_map.json',\n",
|
|
" './gpt-neo-finetuned/vocab.json',\n",
|
|
" './gpt-neo-finetuned/merges.txt',\n",
|
|
" './gpt-neo-finetuned/added_tokens.json',\n",
|
|
" './gpt-neo-finetuned/tokenizer.json')"
|
|
]
|
|
},
|
|
"execution_count": 18,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"# Wczytanie modelu\n",
|
|
"model = AutoModelForCausalLM.from_pretrained(local_model_path, local_files_only=True)\n",
|
|
"\n",
|
|
"# Collator (bez maskowania)\n",
|
|
"data_collator = DataCollatorForLanguageModeling(\n",
|
|
" tokenizer=tokenizer,\n",
|
|
" mlm=False\n",
|
|
")\n",
|
|
"\n",
|
|
"# Argumenty treningowe\n",
|
|
"training_args = TrainingArguments(\n",
|
|
" output_dir=\"./gpt-neo-finetuned\",\n",
|
|
" overwrite_output_dir=True,\n",
|
|
" per_device_train_batch_size=1,\n",
|
|
" num_train_epochs=3,\n",
|
|
" save_steps=500,\n",
|
|
" logging_steps=50,\n",
|
|
" prediction_loss_only=True,\n",
|
|
" fp16=True\n",
|
|
")\n",
|
|
"\n",
|
|
"# Tworzymy Trainer\n",
|
|
"trainer = Trainer(\n",
|
|
" model=model,\n",
|
|
" args=training_args,\n",
|
|
" train_dataset=tokenized_dataset, # <-- to była literówka, nie \"train_dataset\"\n",
|
|
" tokenizer=tokenizer,\n",
|
|
" data_collator=data_collator\n",
|
|
")\n",
|
|
"\n",
|
|
"# Start treningu\n",
|
|
"trainer.train()\n",
|
|
"\n",
|
|
"# Zapis modelu i tokenizer\n",
|
|
"trainer.save_model(\"./gpt-neo-finetuned\")\n",
|
|
"tokenizer.save_pretrained(\"./gpt-neo-finetuned\")\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 19,
|
|
"id": "0d5c46d3-e115-4a68-9337-1f20b62a21a4",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"model_path = \"./gpt-neo-finetuned\"\n",
|
|
"\n",
|
|
"tokenizer = AutoTokenizer.from_pretrained(model_path)\n",
|
|
"model = AutoModelForCausalLM.from_pretrained(model_path)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 20,
|
|
"id": "e656cc27-3748-4530-9ae5-05b2f8e9e105",
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"* Running on local URL: http://127.0.0.1:7863\n",
|
|
"* Running on public URL: https://596aa820b4401f3637.gradio.live\n",
|
|
"\n",
|
|
"This share link expires in 1 week. For free permanent hosting and GPU upgrades, run `gradio deploy` from the terminal in the working directory to deploy to Hugging Face Spaces (https://huggingface.co/spaces)\n"
|
|
]
|
|
},
|
|
{
|
|
"data": {
|
|
"text/html": [
|
|
"<div><iframe src=\"https://596aa820b4401f3637.gradio.live\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
|
|
],
|
|
"text/plain": [
|
|
"<IPython.core.display.HTML object>"
|
|
]
|
|
},
|
|
"metadata": {},
|
|
"output_type": "display_data"
|
|
},
|
|
{
|
|
"data": {
|
|
"text/plain": []
|
|
},
|
|
"execution_count": 20,
|
|
"metadata": {},
|
|
"output_type": "execute_result"
|
|
}
|
|
],
|
|
"source": [
|
|
"import gradio as gr\n",
|
|
"\n",
|
|
"def chat(message, chat_history):\n",
|
|
" if chat_history:\n",
|
|
" prompt = chat_history + f\"\\nUser: {message}\\nAI:\"\n",
|
|
" else:\n",
|
|
" prompt = f\"User: {message}\\nAI:\"\n",
|
|
" \n",
|
|
" inputs = tokenizer(prompt, return_tensors=\"pt\")\n",
|
|
" outputs = model.generate(\n",
|
|
" **inputs,\n",
|
|
" max_length=len(inputs[\"input_ids\"][0]) + 200,\n",
|
|
" temperature=0.7,\n",
|
|
" pad_token_id=tokenizer.eos_token_id,\n",
|
|
" do_sample=True,\n",
|
|
" top_p=0.9\n",
|
|
" )\n",
|
|
" full_response = tokenizer.decode(outputs[0], skip_special_tokens=True)\n",
|
|
" response = full_response.split(\"AI:\")[-1].strip()\n",
|
|
" chat_history += f\"\\nUser: {message}\\nAI: {response}\"\n",
|
|
" return chat_history, chat_history\n",
|
|
"\n",
|
|
"# Gradio UI w trybie notebookowym\n",
|
|
"with gr.Blocks() as demo:\n",
|
|
" gr.Markdown(\"### Lokalny Czat z GPT\")\n",
|
|
" chatbot_output = gr.Textbox(label=\"Historia rozmowy\", lines=20, interactive=False)\n",
|
|
" user_input = gr.Textbox(label=\"Twoje pytanie\", placeholder=\"Zadaj pytanie i naciśnij Enter\")\n",
|
|
" state = gr.State(\"\")\n",
|
|
"\n",
|
|
" user_input.submit(chat, [user_input, state], [chatbot_output, state])\n",
|
|
" user_input.submit(lambda: \"\", None, user_input) # Czyści input po wysłaniu\n",
|
|
"\n",
|
|
"demo.launch(inline=True,share=True)"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"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": "3.12.3"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|