notebook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cloning into 'rwkv-v5-state-tune'...\n",
      "remote: Enumerating objects: 54, done.\u001b[K\n",
      "remote: Counting objects: 100% (54/54), done.\u001b[K\n",
      "remote: Compressing objects: 100% (37/37), done.\u001b[K\n",
      "remote: Total 54 (delta 27), reused 41 (delta 14), pack-reused 0\u001b[K\n",
      "Receiving objects: 100% (54/54), 387.79 KiB | 2.24 MiB/s, done.\n",
      "Resolving deltas: 100% (27/27), done.\n",
      "/home/harrison/finetune-service/rwkv-v5-state-tune/rwkv-v5-state-tune\n"
     ]
    }
   ],
   "source": [
    "!git clone https://github.com/harrisonvanderbyl/rwkv-v5-state-tune.git\n",
    "%cd rwkv-v5-state-tune"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: b2sdk in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from -r requirements.txt (line 1)) (2.1.0)\n",
      "Requirement already satisfied: torch in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from -r requirements.txt (line 2)) (2.3.0)\n",
      "Requirement already satisfied: datasets in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from -r requirements.txt (line 3)) (2.15.0)\n",
      "Requirement already satisfied: matplotlib in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from -r requirements.txt (line 4)) (3.8.2)\n",
      "Requirement already satisfied: logfury<2.0.0,>=1.0.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from b2sdk->-r requirements.txt (line 1)) (1.0.1)\n",
      "Requirement already satisfied: requests<3.0.0,>=2.9.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from b2sdk->-r requirements.txt (line 1)) (2.28.1)\n",
      "Requirement already satisfied: typing-extensions>=4.7.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from b2sdk->-r requirements.txt (line 1)) (4.8.0)\n",
      "Requirement already satisfied: filelock in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (3.9.0)\n",
      "Requirement already satisfied: sympy in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (1.11.1)\n",
      "Requirement already satisfied: networkx in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (3.0rc1)\n",
      "Requirement already satisfied: jinja2 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (3.1.2)\n",
      "Requirement already satisfied: fsspec in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (2023.10.0)\n",
      "Requirement already satisfied: nvidia-cuda-nvrtc-cu12==12.1.105 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.105)\n",
      "Requirement already satisfied: nvidia-cuda-runtime-cu12==12.1.105 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.105)\n",
      "Requirement already satisfied: nvidia-cuda-cupti-cu12==12.1.105 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.105)\n",
      "Requirement already satisfied: nvidia-cudnn-cu12==8.9.2.26 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (8.9.2.26)\n",
      "Requirement already satisfied: nvidia-cublas-cu12==12.1.3.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.3.1)\n",
      "Requirement already satisfied: nvidia-cufft-cu12==11.0.2.54 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (11.0.2.54)\n",
      "Requirement already satisfied: nvidia-curand-cu12==10.3.2.106 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (10.3.2.106)\n",
      "Requirement already satisfied: nvidia-cusolver-cu12==11.4.5.107 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (11.4.5.107)\n",
      "Requirement already satisfied: nvidia-cusparse-cu12==12.1.0.106 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.0.106)\n",
      "Requirement already satisfied: nvidia-nccl-cu12==2.20.5 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (2.20.5)\n",
      "Requirement already satisfied: nvidia-nvtx-cu12==12.1.105 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (12.1.105)\n",
      "Requirement already satisfied: triton==2.3.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from torch->-r requirements.txt (line 2)) (2.3.0)\n",
      "Requirement already satisfied: nvidia-nvjitlink-cu12 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from nvidia-cusolver-cu12==11.4.5.107->torch->-r requirements.txt (line 2)) (12.3.101)\n",
      "Requirement already satisfied: numpy>=1.17 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (1.26.4)\n",
      "Requirement already satisfied: pyarrow>=8.0.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (14.0.1)\n",
      "Requirement already satisfied: pyarrow-hotfix in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (0.6)\n",
      "Requirement already satisfied: dill<0.3.8,>=0.3.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (0.3.7)\n",
      "Requirement already satisfied: pandas in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (2.1.3)\n",
      "Requirement already satisfied: tqdm>=4.62.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (4.66.1)\n",
      "Requirement already satisfied: xxhash in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (3.4.1)\n",
      "Requirement already satisfied: multiprocess in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (0.70.15)\n",
      "Requirement already satisfied: aiohttp in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (3.9.1)\n",
      "Requirement already satisfied: huggingface-hub>=0.18.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (0.19.4)\n",
      "Requirement already satisfied: packaging in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (23.1)\n",
      "Requirement already satisfied: pyyaml>=5.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from datasets->-r requirements.txt (line 3)) (6.0.1)\n",
      "Requirement already satisfied: contourpy>=1.0.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (1.2.0)\n",
      "Requirement already satisfied: cycler>=0.10 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (0.12.1)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (4.45.1)\n",
      "Requirement already satisfied: kiwisolver>=1.3.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (1.4.5)\n",
      "Requirement already satisfied: pillow>=8 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (9.3.0)\n",
      "Requirement already satisfied: pyparsing>=2.3.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (3.1.1)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from matplotlib->-r requirements.txt (line 4)) (2.8.2)\n",
      "Requirement already satisfied: attrs>=17.3.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (23.1.0)\n",
      "Requirement already satisfied: multidict<7.0,>=4.5 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (6.0.4)\n",
      "Requirement already satisfied: yarl<2.0,>=1.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (1.9.3)\n",
      "Requirement already satisfied: frozenlist>=1.1.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (1.4.0)\n",
      "Requirement already satisfied: aiosignal>=1.1.2 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (1.3.1)\n",
      "Requirement already satisfied: async-timeout<5.0,>=4.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from aiohttp->datasets->-r requirements.txt (line 3)) (4.0.3)\n",
      "Requirement already satisfied: six>=1.5 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from python-dateutil>=2.7->matplotlib->-r requirements.txt (line 4)) (1.16.0)\n",
      "Requirement already satisfied: charset-normalizer<3,>=2 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from requests<3.0.0,>=2.9.1->b2sdk->-r requirements.txt (line 1)) (2.0.4)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from requests<3.0.0,>=2.9.1->b2sdk->-r requirements.txt (line 1)) (3.4)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from requests<3.0.0,>=2.9.1->b2sdk->-r requirements.txt (line 1)) (1.26.13)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from requests<3.0.0,>=2.9.1->b2sdk->-r requirements.txt (line 1)) (2024.2.2)\n",
      "Requirement already satisfied: MarkupSafe>=2.0 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from jinja2->torch->-r requirements.txt (line 2)) (2.1.3)\n",
      "Requirement already satisfied: pytz>=2020.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from pandas->datasets->-r requirements.txt (line 3)) (2023.3.post1)\n",
      "Requirement already satisfied: tzdata>=2022.1 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from pandas->datasets->-r requirements.txt (line 3)) (2023.3)\n",
      "Requirement already satisfied: mpmath>=0.19 in /home/harrison/.conda/envs/cuda/lib/python3.10/site-packages (from sympy->torch->-r requirements.txt (line 2)) (1.2.1)\n"
     ]
    }
   ],
   "source": [
    "!pip install -r requirements.txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ipywidgets import interact, Textarea\n",
    "import datasets\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "725b229088e94e628ba8d3b550060daa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='Model', options=('0.4B', '1.5B', '3.0B', '7.0B'), value='0.4B'), D…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<function __main__.chooseModel(Model, Dataset, prompt, response, **kwargs)>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "choices = {\n",
    "    \n",
    "}\n",
    "\n",
    "models = {\n",
    "    '0.4B': 'https://huggingface.co/BlinkDL/rwkv-5-world/resolve/main/RWKV-5-World-0.4B-v2-20231113-ctx4096.pth?download=true',\n",
    "    '1.5B': 'https://huggingface.co/BlinkDL/rwkv-5-world/resolve/main/RWKV-5-World-1B5-v2-20231025-ctx4096.pth?download=true',\n",
    "    '3.0B': 'https://huggingface.co/BlinkDL/rwkv-5-world/resolve/main/RWKV-5-World-3B-v2-20231113-ctx4096.pth?download=true',\n",
    "    '7.0B': 'https://huggingface.co/BlinkDL/rwkv-5-world/resolve/main/RWKV-5-World-7B-v2-20240128-ctx4096.pth?download=true'\n",
    "}\n",
    "\n",
    "datasetOptions = {\n",
    "    \"ysharma/short_jokes\": [\"user: Tell me a joke\\nassistant:\", \" {Joke}\"],\n",
    "    \"Hazzzardous/synthetic-translations-6k-unvalidated\": [\"user: {Response_fr}\\n\\nuser: Translate the above to English\\nassistant:\", \" {Input_ger}\"],\n",
    "    \"PocketDoc/RUCAIBox-Story-Generation-Alpaca\": [\"system: {instruction}\\n\\nuser: {input}\\nassistant:\", \" {output}\"],\n",
    "    \"lonestar108/naughty-chat\": [\"user: {input}\\nassistant:\", \" {output}\"],\n",
    "    \"custom\": [\"user: {prompt}\\nassistant:\", \" {response}\"]\n",
    "}\n",
    "pp0 = Textarea(value=datasetOptions['custom'][0])\n",
    "pp1 = Textarea(value=datasetOptions['custom'][1])\n",
    "exampleout = Textarea(value=\"\")\n",
    "    \n",
    "def chooseModel(Model, Dataset, prompt, response, **kwargs): \n",
    "    choices['model'] = Model,models[Model]\n",
    "    \n",
    "    choices['dataset'] = Dataset if Dataset != 'custom' else None\n",
    "    # if Dataset == 'custom':\n",
    "    #     choices['custom'] = input('Enter the URL to the dataset: ')\n",
    "    \n",
    "    # get first row of dataset\n",
    "    # dataset = datasets.load_dataset(Dataset, split='train', streaming=True)\n",
    "    # firstitem = next(iter(dataset))\n",
    "    \n",
    "    promptf, responsef = datasetOptions[Dataset]\n",
    "    if pp0.value != promptf or pp1.value != responsef:\n",
    "        pp0.value = promptf\n",
    "        pp1.value = responsef\n",
    "    \n",
    "    choices['prompt'] = prompt\n",
    "    choices['response'] = response\n",
    "    \n",
    "\n",
    "interact(chooseModel, **{\"Model\":models.keys(), \"Dataset\":datasetOptions.keys()},  prompt=pp0, response=pp1)\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No ROCm runtime is found, using ROCM_HOME='/opt/rocm'\n",
      "Using /home/harrison/.cache/torch_extensions/py310_cu121 as PyTorch extensions root...\n",
      "Detected CUDA files, patching ldflags\n",
      "Emitting ninja build file /home/harrison/.cache/torch_extensions/py310_cu121/wkv5/build.ninja...\n",
      "/home/harrison/.conda/envs/cuda/lib/python3.10/site-packages/torch/utils/cpp_extension.py:1967: UserWarning: TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation. \n",
      "If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'].\n",
      "  warnings.warn(\n",
      "Building extension module wkv5...\n",
      "Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)\n",
      "Loading extension module wkv5...\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ninja: no work to do.\n",
      "{'': '', 'f=/home/harrison/.local/share/jupyter/runtime/kernel-v2-4317F9KgETpr7xka.json': ''}\n",
      "Masked input:\n",
      " user: I sent you a $25.00 tip\n",
      "assistant:\n",
      "Completion:\n",
      "  oh that great thank you\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/harrison/finetune-service/rwkv-v5-state-tune/train.py:333: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  return torch.tensor(tensor, device=self.emb.weight.device)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.96484\n",
      "3.94473\n",
      "3.88343\n",
      "3.87832\n",
      "3.86253\n",
      "3.90221\n",
      "3.88054\n",
      "3.81855\n",
      "3.79871\n",
      "3.76971\n",
      "3.74763\n",
      "3.75635\n",
      "3.75212\n",
      "3.69967\n",
      "3.68266\n",
      "3.68446\n",
      "3.70336\n",
      "3.64866\n",
      "3.61389\n",
      "3.56679\n",
      "3.53532\n",
      "3.52731\n",
      "3.52360\n",
      "3.46539\n",
      "3.43274\n",
      "3.46111\n",
      "3.43024\n",
      "3.41029\n",
      "3.37337\n",
      "3.42189\n",
      "3.41798\n",
      "3.36583\n",
      "3.37020\n",
      "3.34700\n",
      "3.36090\n",
      "3.35379\n",
      "3.30563\n",
      "3.31301\n",
      "3.25673\n",
      "3.25483\n",
      "3.23818\n",
      "3.24346\n",
      "3.26020\n",
      "3.21125\n",
      "3.19834\n",
      "3.20952\n",
      "3.18029\n",
      "3.16425\n",
      "3.13338\n",
      "3.13530\n",
      "3.14573\n",
      "3.11500\n",
      "3.08972\n",
      "3.04305\n",
      "3.02058\n",
      "3.01955\n",
      "2.95920\n",
      "2.99015\n",
      "2.99455\n",
      "3.03622\n",
      "3.02035\n",
      "3.05293\n",
      "3.01591\n",
      "2.99089\n",
      "2.95853\n",
      "2.95045\n",
      "2.90918\n",
      "2.90981\n",
      "2.89167\n",
      "2.85099\n",
      "2.89985\n",
      "2.91345\n",
      "2.89199\n",
      "2.85443\n",
      "2.85155\n",
      "2.80350\n",
      "2.79770\n",
      "2.76485\n",
      "2.76020\n",
      "2.80891\n",
      "2.79737\n",
      "2.81922\n",
      "2.86341\n",
      "2.81790\n",
      "2.79419\n",
      "2.80917\n",
      "2.79293\n",
      "2.82985\n",
      "2.83132\n",
      "2.83194\n",
      "2.81769\n",
      "2.75610\n",
      "2.74955\n",
      "2.72223\n",
      "2.81893\n",
      "2.81783\n",
      "2.79100\n",
      "2.77957\n",
      "2.77653\n",
      "2.75802\n",
      "2.75137\n",
      "2.75989\n",
      "2.73362\n",
      "2.71022\n",
      "2.72080\n",
      "2.71367\n",
      "2.71001\n",
      "2.67920\n",
      "2.66165\n",
      "2.63716\n",
      "2.62639\n",
      "2.58882\n",
      "2.59766\n",
      "2.59669\n",
      "2.59732\n",
      "2.55964\n",
      "2.53166\n",
      "2.52461\n",
      "2.56713\n",
      "2.55830\n",
      "2.52765\n",
      "2.51924\n",
      "2.54797\n",
      "2.57057\n",
      "2.55063\n",
      "2.59107\n",
      "2.59042\n",
      "2.59371\n",
      "2.63825\n",
      "2.64930\n",
      "2.61762\n",
      "2.68518\n",
      "2.74154\n",
      "2.74353\n",
      "2.74854\n",
      "2.77127\n",
      "2.76708\n",
      "2.77951\n",
      "2.74288\n",
      "2.77448\n",
      "2.77560\n",
      "2.75557\n",
      "2.78967\n",
      "2.77909\n",
      "2.76045\n",
      "2.74118\n",
      "2.71193\n",
      "2.64938\n",
      "Saving state\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from train import train_model\n",
    "\n",
    "model = train_model(\n",
    "    model_location=choices['model'][0]+\".pth\",\n",
    "    huggingface_dataset=choices['dataset'],\n",
    "    data_path=choices['dataset'].replace(\"/\",\"_\")+\".jsonl\",\n",
    "    model_url=choices['model'][1],\n",
    "    prompt_formatter=pp0.value,\n",
    "    response_formatter=pp1.value,\n",
    "    save_filename=\"state_\"+choices['model'][0]+\".pth\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b776184a330142948864fadd88519071",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Textarea(value='', description='message'), FloatSlider(value=1.0, description='statemix'…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<function __main__.newMessage(message, statemix, temp, top_p)>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "conversation = []\n",
    "import torch\n",
    "from math import e\n",
    "def dart_sample(out, temperature=1.0, top_p=0.8):\n",
    "    out = torch.softmax(out, dim=-1)    \n",
    "    min = torch.min(out)\n",
    "    max = torch.max(out)\n",
    "    topp = pow(e,-pow(e,e-(top_p*e*e)))\n",
    "    dart = torch.rand(1, device=out.device).pow(temperature)\n",
    "    dart = dart\n",
    "    min = max * (1 - topp) + min * topp\n",
    "    dart = min * (1 - dart) + max * dart\n",
    "    \n",
    "    out = torch.argmin(torch.abs(out - dart)).item()\n",
    "    \n",
    "    return out\n",
    "\n",
    "from ipywidgets import interact, Textarea, FloatSlider\n",
    "model = model.eval()\n",
    "from tokenizer import world\n",
    "state = torch.load(\"state_\"+choices['model'][0]+\".pth\")\n",
    "model.load_state(state)\n",
    "messagebox = Textarea(value=\"\")\n",
    "\n",
    "howMuchSlider = FloatSlider(value=1,min=0,max=1,step=0.1)\n",
    "temp = FloatSlider(value=0.0,min=0,max=1,step=0.1)\n",
    "top_p = FloatSlider(value=0.5,min=0,max=1,step=0.1)\n",
    "lasttop_p = 0.5\n",
    "last_temp = 0.0\n",
    "currentstatemix = 1.0\n",
    "def newMessage(message, statemix, temp, top_p):\n",
    "    global conversation, currentstatemix, lasttop_p, last_temp\n",
    "    if \"\\n\" not in message:\n",
    "        currentstatemix = statemix\n",
    "        \n",
    "        state = torch.load(\"state_\"+choices['model'][0]+\".pth\")\n",
    "\n",
    "        mixstate = model.new_state(1)\n",
    "        newstate = {\n",
    "            key: torch.lerp(mixstate[key], state[key],statemix) for key in mixstate.keys()\n",
    "        }\n",
    "        model.load_state(newstate)\n",
    "        convo = \"\\n\\n\".join(conversation)\n",
    "        context = world.encode(convo)\n",
    "        model([context]) if convo else \"\"\n",
    "        \n",
    "    saveit = False\n",
    "    \n",
    "    if \"\\n\" in message:\n",
    "        \n",
    "        conversation += [\"user: \"+ message[:-1]]\n",
    "        \n",
    "        messagebox.value = \"\"\n",
    "        nextinps = world.encode(\"user: \"+message[:-1]+\"\\n\\nassistant:\")\n",
    "        logits = model([nextinps])[0,-1]\n",
    "        print(\"\\n\")\n",
    "        print(\"\\n\".join(conversation)+\"\\n\\nassistant:\", end=\"\", flush=True)\n",
    "        saveit = True\n",
    "    else:\n",
    "        \n",
    "        nextinps = world.encode(\"user: \"+message+\"\\n\\nassistant:\")\n",
    "        logits = model([nextinps])[0,-1]\n",
    "        \n",
    "        print(\"\\n\".join(conversation)+\"\\n\\nuser: \"+message + \"\\n\\nassistant:\", end=\"\", flush=True)\n",
    "    \n",
    "    \n",
    "    outputs = \"\"\n",
    "    buffer = []\n",
    "    \n",
    "    \n",
    "    for i in range(50):\n",
    "        nextobj = dart_sample(logits,temp,top_p)#logits.argmax(-1).cpu().item()\n",
    "        \n",
    "        buffer += [nextobj]\n",
    "        try:\n",
    "            nextmessage = world.decode(buffer)\n",
    "            buffer = []\n",
    "            if \"\\n\" in nextmessage:\n",
    "                break\n",
    "            outputs += nextmessage\n",
    "            \n",
    "            print(nextmessage, end=\"\", flush=True)\n",
    "            \n",
    "            logits = model([[nextobj]])\n",
    "        except:\n",
    "            pass\n",
    "    if saveit:\n",
    "        conversation += [\"assistant: \" + outputs]\n",
    "        model([[261]])\n",
    "\n",
    "\n",
    "interact(newMessage, message=messagebox, statemix=howMuchSlider, temp=temp, top_p=top_p)\n",
    "\n",
    "# newMessage(\"\\n\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cuda",
   "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
}