{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1f19ebe8",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = transformers.AutoModelForCausalLM.from_pretrained(\n",
    "        script_args.model_name_or_path,\n",
    "        device_map=\"auto\",\n",
    "    )\n",
    "\n",
    "rotation_config = RotationConfig(**rotation_adapter_config)\n",
    "\n",
    "# hra_config = OFTConfig(\n",
    "#     r=script_args.hrft_r,\n",
    "#     eps=script_args.eps,\n",
    "#     init_weights=script_args.init_weights,\n",
    "#     target_modules=[\"q_proj\", \"v_proj\"],\n",
    "#     task_type=\"CAUSAL_LM\",\n",
    "# )\n",
    "for adapter_name in adapter_names:\n",
    "    RotationTuner(model,\n",
    "                rotation_config,\n",
    "                adapter_name=adapter_name,\n",
    "                    )\n",
    "    model.set_adapter(adapter_name)\n",
    "    \n",
    "\n",
    "\n",
    "rotation_layers = filter(\n",
    "                lambda p: p.requires_grad, model.parameters()\n",
    "            )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e701e0d3",
   "metadata": {},
   "outputs": [
    {
     "ename": "SystemExit",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "An exception has occurred, use %tb to see the full traceback.\n",
      "\u001b[31mSystemExit\u001b[39m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/cosmos13/Documents/VS_code/General_code/.venv3/lib/python3.11/site-packages/IPython/core/interactiveshell.py:3707: UserWarning: To exit: use 'exit', 'quit', or Ctrl-D.\n",
      "  warn(\"To exit: use 'exit', 'quit', or Ctrl-D.\", stacklevel=1)\n"
     ]
    }
   ],
   "source": [
    "from dataclasses import dataclass, field, fields\n",
    "from typing import Optional, List, Literal, Dict, Any\n",
    "\n",
    "@dataclass\n",
    "class TrainingConfig:\n",
    "    model_name_or_path: Optional[str] = field(default=\"huggyllama/llama-7b\")\n",
    "    adapter_name_or_path: Optional[str] = field(default=None)\n",
    "    data_path: str = field(default=None, metadata={\"help\": \"Path to the training data.\"})\n",
    "    dataset_split: str = field(\n",
    "        default=\"train[:100000]\", metadata={\"help\": \"(`['train', 'test', 'eval']`):\"}\n",
    "    )\n",
    "    dataset_field: List[str] = field(\n",
    "        default=None, metadata={\"help\": \"Fields of dataset input and output.\"}\n",
    "    )\n",
    "    optim: str = field(default=\"adamw_torch\")\n",
    "    model_max_length: int = field(default=512, metadata={\n",
    "        \"help\": \"Maximum sequence length. Sequences will be right padded (and possibly truncated).\"}, )\n",
    "    hrft_r: int = field(default=8, metadata={\n",
    "        \"help\": \"The rank of the adapter. When passing `None` and `adapter_name_or_path` is also `None`, full fine-tuning is used.\"})\n",
    "    init_a: float = field(default=1e-4, metadata={\"help\": \"The initial weights\"})\n",
    "    eps: float = field(default=1e-4, metadata={\"help\": \"The control strength of COFT. The freedom of rotation.\"})\n",
    "    lamda: float = field(default=1e-4, metadata={\"help\": \"The control strength of regularity\"})\n",
    "    add_orth: str = field(default='none', metadata={\"help\": \"\"})\n",
    "    init_weights: Literal[True, \"pissa\"] = field(\n",
    "        default=True,\n",
    "        metadata={\n",
    "            \"help\": (\n",
    "                \"Passing True (default) results in the LoRA initialization.\"\n",
    "                \"Passing `pissa` results in PiSSA initialization.\"\n",
    "            ),\n",
    "        },\n",
    "    )\n",
    "    extension: Optional[Dict[str, Any]] = field(\n",
    "        default=None, \n",
    "        metadata={\"help\": \"Serialized MainConfig excluding training args\"}\n",
    "    )\n",
    "\n",
    "\n",
    "\n",
    "from omegaconf import OmegaConf, DictConfig\n",
    "import torch\n",
    "import yaml\n",
    "from dataclasses import asdict\n",
    "\n",
    "import os\n",
    "import transformers\n",
    "from transformers import (AutoModelForCausalLM, AutoTokenizer, \n",
    "                          LlamaTokenizer, AutoModel, AutoConfig, \n",
    "                          TrainingArguments)\n",
    "import inspect\n",
    "\n",
    "valid_hf_arg_names = set(inspect.signature(TrainingArguments).parameters.keys())\n",
    "# training_config_dict = OmegaConf.to_container(\n",
    "#     training_cfg, resolve=True\n",
    "# )\n",
    "train = TrainingConfig()\n",
    "training_config_dict = asdict(train)\n",
    "filtered_training_args_dict = {\n",
    "    key: value for key, value in training_config_dict.items()\n",
    "    if key in valid_hf_arg_names\n",
    "}\n",
    "import sys\n",
    "sys.exit()\n",
    "print(filtered_training_args_dict.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1fd93799",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--- Starting Environment Check ---\n",
      "Python Executable: /Users/cosmos13/Documents/VS_code/General_code/.venv3/bin/python\n",
      "Python Version: 3.11.3 (main, Jul 29 2025, 12:05:44) [Clang 16.0.0 (clang-1600.0.26.4)]\n",
      "Success: transformers version 4.57.2 is imported.\n",
      "Success: datasets version 4.4.1 is imported.\n",
      "--- End of Environment Check ---\n"
     ]
    }
   ],
   "source": [
    "# debug_env.py\n",
    "\n",
    "import sys\n",
    "\n",
    "# Function to check imports and print versions\n",
    "def check_environment():\n",
    "    print(\"--- Starting Environment Check ---\")\n",
    "    \n",
    "    # 1. Check Python version\n",
    "    # It is important to know which python executable is running\n",
    "    print(f\"Python Executable: {sys.executable}\")\n",
    "    print(f\"Python Version: {sys.version}\")\n",
    "\n",
    "    try:\n",
    "        # 2. Try importing transformers\n",
    "        # This usually triggers the import of file_utils, imports, etc.\n",
    "        import transformers\n",
    "        print(f\"Success: transformers version {transformers.__version__} is imported.\")\n",
    "    except ImportError as e:\n",
    "        print(f\"Error: Failed to import transformers. Details: {e}\")\n",
    "    except Exception as e:\n",
    "        print(f\"Critical Error during transformers import: {e}\")\n",
    "\n",
    "    try:\n",
    "        # 3. Try importing datasets\n",
    "        # This is where your error originated (aiohttp -> aiohappyeyeballs chain)\n",
    "        import datasets\n",
    "        print(f\"Success: datasets version {datasets.__version__} is imported.\")\n",
    "    except SyntaxError as e:\n",
    "        # Catching the specific SyntaxError from your traceback\n",
    "        print(f\"SyntaxError detected in dependencies: {e}\")\n",
    "        print(\"Suggestion: Please reinstall 'aiohttp' and 'datasets'.\")\n",
    "    except Exception as e:\n",
    "        print(f\"Error: Failed to import datasets. Details: {e}\")\n",
    "\n",
    "    print(\"--- End of Environment Check ---\")\n",
    "\n",
    "if __name__ == \"__main__\":\n",
    "    check_environment()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a98dac91",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Transformers: 4.57.2\n",
      "Datasets: 4.4.1\n",
      "HuggingFace Hub: 0.36.0\n",
      "Success! All core libraries imported correctly.\n"
     ]
    }
   ],
   "source": [
    "# Create a quick test file: check_imports.py\n",
    "import transformers\n",
    "import datasets\n",
    "import huggingface_hub\n",
    "\n",
    "print(f\"Transformers: {transformers.__version__}\")\n",
    "print(f\"Datasets: {datasets.__version__}\")\n",
    "print(f\"HuggingFace Hub: {huggingface_hub.__version__}\")\n",
    "print(\"Success! All core libraries imported correctly.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "791432eb",
   "metadata": {},
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "/home/work/miniconda3/envs/llmr/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so: undefined symbol: iJIT_NotifyEvent",
     "output_type": "error",
     "traceback": [
      "\u001b[31m---------------------------------------------------------------------------\u001b[39m",
      "\u001b[31mImportError\u001b[39m                               Traceback (most recent call last)",
      "\u001b[36mCell\u001b[39m\u001b[36m \u001b[39m\u001b[32mIn[2]\u001b[39m\u001b[32m, line 1\u001b[39m\n\u001b[32m----> \u001b[39m\u001b[32m1\u001b[39m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorch\u001b[39;00m\n\u001b[32m      3\u001b[39m \u001b[38;5;66;03m# 1. Check if CUDA is available\u001b[39;00m\n\u001b[32m      4\u001b[39m \u001b[38;5;28mprint\u001b[39m(\u001b[33mf\u001b[39m\u001b[33m\"\u001b[39m\u001b[33mCUDA Available: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mtorch.cuda.is_available()\u001b[38;5;132;01m}\u001b[39;00m\u001b[33m\"\u001b[39m)\n",
      "\u001b[36mFile \u001b[39m\u001b[32m~/miniconda3/envs/llmr/lib/python3.12/site-packages/torch/__init__.py:367\u001b[39m\n\u001b[32m    365\u001b[39m     \u001b[38;5;28;01mif\u001b[39;00m USE_GLOBAL_DEPS:\n\u001b[32m    366\u001b[39m         _load_global_deps()\n\u001b[32m--> \u001b[39m\u001b[32m367\u001b[39m     \u001b[38;5;28;01mfrom\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mtorch\u001b[39;00m\u001b[34;01m.\u001b[39;00m\u001b[34;01m_C\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mimport\u001b[39;00m *  \u001b[38;5;66;03m# noqa: F403\u001b[39;00m\n\u001b[32m    370\u001b[39m \u001b[38;5;28;01mclass\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[34;01mSymInt\u001b[39;00m:\n\u001b[32m    371\u001b[39m \u001b[38;5;250m    \u001b[39m\u001b[33;03m\"\"\"\u001b[39;00m\n\u001b[32m    372\u001b[39m \u001b[33;03m    Like an int (including magic methods), but redirects all operations on the\u001b[39;00m\n\u001b[32m    373\u001b[39m \u001b[33;03m    wrapped node. This is used in particular to symbolically record operations\u001b[39;00m\n\u001b[32m    374\u001b[39m \u001b[33;03m    in the symbolic shape workflow.\u001b[39;00m\n\u001b[32m    375\u001b[39m \u001b[33;03m    \"\"\"\u001b[39;00m\n",
      "\u001b[31mImportError\u001b[39m: /home/work/miniconda3/envs/llmr/lib/python3.12/site-packages/torch/lib/libtorch_cpu.so: undefined symbol: iJIT_NotifyEvent"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "# 1. Check if CUDA is available\n",
    "print(f\"CUDA Available: {torch.cuda.is_available()}\")\n",
    "\n",
    "# 2. Get the number of available GPUs (this is the value for --num_processes)\n",
    "num_gpus = torch.cuda.device_count()\n",
    "print(f\"Number of GPUs: {num_gpus}\")\n",
    "\n",
    "# 3. Print information for each GPU\n",
    "for i in range(num_gpus):\n",
    "    print(f\"--- GPU {i} ---\")\n",
    "    print(f\"Name: {torch.cuda.get_device_name(i)}\")\n",
    "    print(f\"Memory Total: {torch.cuda.get_device_properties(i).total_memory / (1024**3):.2f} GB\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "51a3a678",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/vllm/lib/python3.11/site-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",
      "None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "xaaa ['', '/home/work/.local/lib/python3.12/site-packages', '/home/work/an_nguyen/OminiControlRotation/nl_tasks/src', '/home/work/miniconda3/envs/vllm/lib/python311.zip', '/home/work/miniconda3/envs/vllm/lib/python3.11', '/home/work/miniconda3/envs/vllm/lib/python3.11/lib-dynload', '/home/work/miniconda3/envs/vllm/lib/python3.11/site-packages']\n",
      "--- DIAGNOSTIC START ---\n",
      "Current Python Executable: /home/work/miniconda3/envs/vllm/bin/python\n",
      "Omegaconf in sys.path: False\n",
      "--- DIAGNOSTIC END ---\n"
     ]
    }
   ],
   "source": [
    "import transformers\n",
    "import sys\n",
    "print('xaaa', sys.path)\n",
    "print(\"--- DIAGNOSTIC START ---\")\n",
    "print(f\"Current Python Executable: {sys.executable}\")\n",
    "print(f\"Omegaconf in sys.path: {'omegaconf' in sys.modules}\")\n",
    "print(\"--- DIAGNOSTIC END ---\")\n",
    "from transformers import Trainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "68e1eaf7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time per batch = 0.0027077293395996096\n"
     ]
    }
   ],
   "source": [
    "\n",
    "import time\n",
    "from torch.utils.data import DataLoader, TensorDataset\n",
    "import torch\n",
    "\n",
    "ds = TensorDataset(torch.randn(40000,512))\n",
    "dl = DataLoader(ds, batch_size=8, num_workers=8, pin_memory=True)\n",
    "\n",
    "# warmup\n",
    "next(iter(dl))\n",
    "\n",
    "t0=time.time()\n",
    "for i, batch in enumerate(dl):\n",
    "    if i==50: break\n",
    "print(\"time per batch =\", (time.time()-t0)/50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1ccffba0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8 → sec per batch = 0.0014403629302978515\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:627: UserWarning: This DataLoader will create 16 worker processes in total. Our suggested max number of worker in current system is 12, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
      "  warnings.warn(\n",
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:627: UserWarning: This DataLoader will create 32 worker processes in total. Our suggested max number of worker in current system is 12, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16 → sec per batch = 0.0015110158920288086\n",
      "32 → sec per batch = 0.002375731468200684\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:627: UserWarning: This DataLoader will create 48 worker processes in total. Our suggested max number of worker in current system is 12, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "48 → sec per batch = 0.00206387996673584\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:627: UserWarning: This DataLoader will create 64 worker processes in total. Our suggested max number of worker in current system is 12, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "64 → sec per batch = 0.003051443099975586\n"
     ]
    }
   ],
   "source": [
    "dataset = TensorDataset(torch.randn(40000,512))\n",
    "for w in [8, 16, 32, 48, 64]:\n",
    "    dl = DataLoader(dataset, batch_size=8, num_workers=w, pin_memory=True, prefetch_factor=4, persistent_workers=True)\n",
    "    import time\n",
    "    next(iter(dl))   # warmup\n",
    "    t0=time.time()\n",
    "    for i,b in enumerate(dl):\n",
    "        if i==50: break\n",
    "    print(w, \"→ sec per batch =\", (time.time()-t0)/50)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "796e037a",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = PeftModel.from_pretrained(model, mainCfg.model.adapter_path, is_trainable = True)\n",
    "trainer = MyTrainer(model=model, processing_class=tokenizer,\n",
    "                        lamda=mainCfg.model.lambda_reg,\n",
    "                        optimizers=(optimizer, None),\n",
    "                        args=training_args, **data_module)\n",
    "results = trainer.evaluate()\n",
    "print(results)\n",
    "# 'eval_loss': 0.208\n",
    "model2 = model.merge_and_unload()\n",
    "\n",
    "trainer2 = MyTrainer(model=model2, processing_class=tokenizer,\n",
    "                        lamda=mainCfg.model.lambda_reg,\n",
    "                        optimizers=(optimizer, None),\n",
    "                        args=training_args, **data_module)\n",
    "# model2.config.use_cache = False\n",
    "results2 = trainer2.evaluate()\n",
    "#'eval_loss': 0.885\n",
    "print(results2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4ef1a714",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['base_model.model.model.layers.0.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.0.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.0.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.0.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.1.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.1.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.1.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.1.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.2.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.2.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.2.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.2.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.3.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.3.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.3.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.3.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.4.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.4.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.4.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.4.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.5.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.5.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.5.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.5.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.6.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.6.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.6.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.6.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.7.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.7.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.7.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.7.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.8.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.8.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.8.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.8.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.9.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.9.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.9.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.9.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.10.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.10.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.10.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.10.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.11.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.11.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.11.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.11.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.12.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.12.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.12.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.12.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.13.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.13.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.13.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.13.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.14.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.14.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.14.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.14.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.15.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.15.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.15.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.15.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.16.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.16.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.16.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.16.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.17.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.17.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.17.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.17.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.18.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.18.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.18.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.18.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.19.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.19.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.19.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.19.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.20.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.20.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.20.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.20.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.21.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.21.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.21.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.21.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.22.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.22.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.22.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.22.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.23.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.23.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.23.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.23.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.24.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.24.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.24.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.24.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.25.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.25.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.25.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.25.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.26.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.26.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.26.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.26.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.27.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.27.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.27.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.27.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.28.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.28.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.28.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.28.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.29.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.29.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.29.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.29.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.30.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.30.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.30.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.30.self_attn.v_proj.rotation.default.V', 'base_model.model.model.layers.31.self_attn.q_proj.rotation.default.U', 'base_model.model.model.layers.31.self_attn.q_proj.rotation.default.V', 'base_model.model.model.layers.31.self_attn.v_proj.rotation.default.U', 'base_model.model.model.layers.31.self_attn.v_proj.rotation.default.V'])\n",
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0007\n",
      "Standard deviation: 0.2664\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0008\n",
      "Standard deviation: 0.2623\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0005\n",
      "Standard deviation: 0.2762\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0010\n",
      "Standard deviation: 0.2755\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0002\n",
      "Standard deviation: 0.2651\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0012\n",
      "Standard deviation: 0.2635\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0000\n",
      "Standard deviation: 0.2699\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0016\n",
      "Standard deviation: 0.2680\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: -0.0009\n",
      "Standard deviation: 0.2653\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0008\n",
      "Standard deviation: 0.2658\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0018\n",
      "Standard deviation: 0.2757\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import torch\n",
    "\n",
    "def analyze_value_distribution(V: torch.Tensor):\n",
    "    \"\"\"\n",
    "    Analyzes the distribution of all individual elements in the matrix V.\n",
    "\n",
    "    Args:\n",
    "        V: The input matrix (can be any shape, e.g., N x R x D).\n",
    "    \n",
    "    Returns:\n",
    "        torch.Tensor: A 1D tensor containing all elements of V.\n",
    "    \"\"\"\n",
    "    # Flatten the tensor into a 1D vector containing all elements.\n",
    "    flat_values = V.flatten()\n",
    "    \n",
    "    # Calculate basic statistics\n",
    "    mean = flat_values.mean().item()\n",
    "    std = flat_values.std().item()\n",
    "    print('-'*100)\n",
    "    print(f\"Matrix V Shape: {V.shape}\")\n",
    "    print(f\"Number of elements: {flat_values.numel()}\")\n",
    "    print(f\"Mean value: {mean:.4f}\")\n",
    "    print(f\"Standard deviation: {std:.4f}\")\n",
    "\n",
    "    # return flat_values\n",
    "\n",
    "def visualize_value_distribution(values: torch.Tensor, text: str = 'UV', bins_mode: str = 'auto', bins_count: int = 50):\n",
    "    \"\"\"\n",
    "    Visualizes the distribution of tensor elements using a histogram.\n",
    "\n",
    "    Args:\n",
    "        values: 1D tensor of all element values.\n",
    "        bins_mode: 'auto' (default) or 'manual'.\n",
    "        bins_count: Number of bins if bins_mode is 'manual'.\n",
    "    \"\"\"\n",
    "    # Move tensor to CPU and convert to numpy array for plotting\n",
    "    data = values.cpu().numpy()\n",
    "    \n",
    "    plt.figure(figsize=(8, 5))\n",
    "    \n",
    "    if bins_mode == 'auto':\n",
    "        # Use 'auto' to let Matplotlib choose the best bin size based on data\n",
    "        plt.hist(data, bins='auto', edgecolor='black', alpha=0.7)\n",
    "        plt.title(f'Value Distribution of {text[-40:]} (Auto Binning)')\n",
    "    elif bins_mode == 'manual':\n",
    "        # Manually set the number of bins\n",
    "        plt.hist(data, bins=bins_count, edgecolor='black', alpha=0.7)\n",
    "        plt.title(f'Value Distribution of Matrix V (Manual Binning: {bins_count} bins)')\n",
    "    else:\n",
    "        raise ValueError(\"bins_mode must be 'auto' or 'manual'.\")\n",
    "\n",
    "    plt.xlabel('Element Value')\n",
    "    plt.ylabel('Frequency')\n",
    "    plt.grid(axis='y', alpha=0.5)\n",
    "    plt.show()\n",
    "# Đọc file .bin\n",
    "state_dict = torch.load(\"../exp395/run_ex02/ft2/adapter_model.bin\", map_location=\"cpu\")\n",
    "\n",
    "# Kiểm tra các key trong state_dict\n",
    "print(state_dict.keys())\n",
    "cnt = 0\n",
    "for k,v in state_dict.items():\n",
    "    v = v.flatten()\n",
    "    analyze_value_distribution(v)\n",
    "    visualize_value_distribution(v, k, bins_count=100)\n",
    "    if cnt >= 10:\n",
    "        break \n",
    "    else:\n",
    "        cnt += 1\n",
    "\n",
    "# v = state_dict['base_model.model.model.layers.0.self_attn.q_proj.rotation.default.U']\n",
    "# v = v.squeeze().flatten()\n",
    "# print(v.shape)\n",
    "# visualize_value_distribution(v,bins_count=100)\n",
    "\n",
    "# analyze_value_distribution(v)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4e729a5a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0000\n",
      "Standard deviation: 0.2651\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsAAAAHUCAYAAAA0gJ7/AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAVqhJREFUeJzt3Ql4VOX1+PETEhJISCAbAQpEQUQr1CIpmygggqC4AHVDFBVcqoUiUKv4s4JVcamg1YqoLCpWsFWp1haFiriwCYoCKiCEsAWIhCWBkEyS+T/n5X+nM5OZbMxkZnK/n+e5kHvnnZk7984kZ9573vNGOZ1OpwAAAAA20SDUOwAAAADUJQJgAAAA2AoBMAAAAGyFABgAAAC2QgAMAAAAWyEABgAAgK0QAAMAAMBWCIABAABgKwTAAAAAsBUCYCBEhg4dKo0bN5bDhw/7bXPDDTdIw4YNZf/+/dV+3KioKJkyZYrUtU8++cQ8t7XExsZKenq6nH/++fLAAw9ITk5OhfvMmzfPtN2xY0eNnuuxxx6TRYsW1eg+vp6rb9++0qlTJwmkf//7336P/2mnnSY333yzhLOvv/5a+vTpI02bNjXH65lnnvHZTo+j3v7nP//Z5+263TreeXl55v1w3XXX+X3eo0ePSnx8vFxxxRVV7uNnn30mcXFxPt9T6rzzzqt034L9XqsOfS+4f14aNWokZ5xxhkyYMEF++uknj7b6ftI2waafB12C6dChQ9KsWbOgHFOgRnQqZAB17/3339dpyJ1//etffd5++PBhZ+PGjZ1XXXVVjR5XH/Ohhx5y1rVly5aZ537sscecK1eudH7++efOf/7zn87Jkyc7W7RoYV7L/PnzPe5z4MAB0/bEiRM1eq6EhATnqFGjanQfX8/Vp08f5znnnOMMpLvvvtscB1+++uor548//ugMZ7/85S+dHTp0cP773/82xys3N9dnu+zsbPM6n3rqKZ+363a9Xdup4cOHO+Pi4pz5+fk+28+aNcu0X7RoUaX7V15e7jzvvPPMcfbl66+/No+jy1lnneU8VbV5r1VHZmam8/zzzzfHWJePP/7Y+eSTTzrj4+OdXbt29Wi7a9cu0ybYNm3aZJZgmzJlivOMM85wFhcXB/25AH/oAQZCZPDgwdKqVSuZM2eOz9vffPNNKSoqktGjR0sk6dChg/To0cP0/Gpv3qOPPiqbNm2Ss846y/R+btiwwdVWe4i1rfbmBYseQ/1eUBfPVZUuXbpI+/btJZxt3LhRLr74YvP+1OPVokWLgDyuvo+Li4vljTfe8Hm7fg4yMjLksssuq/RxFi9eLF999ZWMHTvW5+2vvPKK+V8f54cffpAVK1ZIuNKeUD3GuvTr109+//vfy8SJE2XdunWyZcsWV7vWrVubNsH285//3CzBduedd5orA//4xz+C/lyAPwTAQIhER0fLqFGjzB8796DQMnfuXGnZsqUJRPQS8l133WX+ODVp0kSaN28uF110kbkUXBV/l0/9pR8sXLhQevbsKQkJCea5LrnkEnNZ/FSkpKTIrFmzpLS0VGbMmFHpPuhzDRkyxLxGDVb1S4IGM7t37za3a/tjx47Jq6++6rp8bF22tR7vo48+kltvvdUEvXpZXQOvytIt9DhqgKEpKT/72c/kwQcflLKysgrpHfq/rzQAfWylAf5f//pX135ai/WcvlIgdu7cKSNHjnS93rPPPluefvppKS8vr/A8ekl/+vTpcvrpp5tzo+dp1apV1Q5sr7zySklOTjaX23/5y1+aY+h9LvQczZw507XvgaLvIw3k9H3t7fvvv5fVq1fLTTfdJDExMZU+ju7br371K+nYsWOF206cOCF/+9vfpGvXrq73ma8vmHoO9FxU9Vmp7L1WnWNaG5p6ojT1yd9+Kd1//ZzoFwJN+dD3rn7J9H691nldtmyZ/OY3v5G0tDRJTU2VYcOGyd69eytNgajp++7ll1+WM88807yP9XeVngtfx1q/6AwYMEBefPHFUzpWwKkgAAZCSIM0/QPj/Ufru+++kzVr1pgAWQPl/Px8s/2hhx6SDz74wAQR7dq1M3+svIOyU813vP76680fr7feektef/11KSgokAsuuMDs06nQoEUD+k8//dRvGw029A+j5jxrILlkyRKTg9q2bVuzH2rlypXmj/2ll15qftblhRdeqHBcNYDQ/ddeJvdgwtu+fftMbqrmW//zn/+UX//61/LII4/I7373uxq/Rg2c9f7WflqLvm5f9ItNr169TMD+pz/9Sd577z3T+zpp0iT57W9/W6G9+zHRnlQ9Xnocjhw5Uul+bd682TyP9sT/5S9/kXfeececYw1OnnzySdNGv2Tovip9Dda+B0qDBg3M82nv7TfffONxmxUU63mrTElJiSxdutT0lvqir0tzTPVx9EpE7969zRe6wsLCWu1zZe+16hzTquiVCf3CoYvuowapem716okGm1XR46g9xvfcc4957/7iF78wPe2+PmNjxowxnwMNSnX/9PeGfvGqjuq871566SW5/fbbzT7osfi///s/mTp1qt/fT/q764svvqh0DAQQVH6TIwDUCc1DTUtLc5aUlLi2TZw40eQwbtmyxed9SktLnQ6Hw9m/f3/n0KFDK80B1p99fdTnzp3rkaO5c+dOZ0xMjHPs2LEe7QoKCkwO7zXXXFOtHOC///3vftt0797d5AL724e1a9dWKw/UX16m9Xg33XRTla/XOva6TXOV3d12223OBg0aOHNycjxem/7vKw9WH7s6OcCa9+m+3/fdd59pu3r1ao92v/nNb5xRUVHOzZs3ezxP586dzbm3rFmzxmx/8803KzlaTud1111n8m/1HLsbPHiwyTnVfHOLPp6//NpTyQFW27dvN69r3Lhxrm36Ptb3l+bDVkWPkz7mggULfN5+0UUXORs1auQ8dOiQxzmfPXu2Rzs9B3ouvPn6rPh7r9XkmPqiz2/lKrsv3bp1q5B37Wu/9P76Wq33qCoqKnKmpKQ477jjDtc26xjcddddHvfXfGPd7v5c+nnQxVLd911ZWZk5h/r5dqf71rBhQ5/HesmSJeYx/vOf/1R6nIBgoQcYCDHtsdFR39r7p7Q3aP78+abXVXuxLHq5UC916qVWvUysvTn//e9/zeXjQPjwww/Nc+tlaKtXShd9Pq0KEIie5pPxlX86Cl4vJ//hD38wr7e2vc7Dhw+vdtvExMQKlQdGjBhhUhAq660OhI8//tj0Gnbr1s1ju/Yi6rHS291pL61eEbBob5vyVw3B/Xn69+8vbdq0qfA8x48fD2hPb2W0V1N7b7UXUXtz1X/+8x/TC19V76+yLtlruoi37Oxs04Oql/Y1t1ZdffXV5vz6y7M/FYE4ptpD/eWXX5pFe0Nnz55trgpoepN3JQhfNOVCr45Y9LOqKQi+3g/e7/Hqvneq877T3nA9h9dcc43H/XTftDfbF+sc7tmzp8rnB4KBABgIMb3crHl/1mVgLaOlKQDug980/07z97p37y5vv/22yb/TP5qDBg0yg7wCwSq1pqkKGly7L3oZuTp/kKui+a6a0+uPHofly5ebP+yTJ0+Wc845x7TX1A+Hw1Ht5/GXcuCL5iN6swZ+HTx4UIJJH9/XvlrHyPv5NXfTnTWgr6r3QE2fpzqsXF33XGl3+uVJeaef6Ptan8/6wqfve80r9Q6efLFepwZ63jTI1S8N+nnSy+q66HtGAz8NLnVAXCAF4pjq+z0rK8ssmk6hXwI0RUG/1GoeeFW83w/We8LX+6G2753q3Nd6rb4+S762uZ/DQP3+Amqq8tEGAIJOcww171YHkOTm5po/5Nprpb1XFu0R1pw5HQDkzsqLrYz1h0YHgrlXQPAOaHVwjNKc2czMTAk0zWnWXqKqqlp07txZFixYYIKZb7/91gziefjhh81xuu+++6r1XDUZvOWrxrLup/sffvdj6O5UvxTo4+s599fTaZ2TUxWM59H7aK+gvx483a63ewdP2kOrvfz6PtcrC//617/MVQcNgqvznMrKibdob701EFEf3xd9Pis3V8+n97ms6fkM1rmzele986TDmXWOK/ssebPOYaDe40BN0QMMhAENCrUn7amnnjI9wDooS6sXuAd03uW7NDiszmVWawS2tnf3/vvvVxilr71627Ztc/VKeS+1pX/stPSR9gbqgJ3q0Nd87rnnmtH8eklbB09V1ctVG/olwuqNtGgvnA7auvDCCys9ht73s/ZNVWf/9BK6pnm4vzb12muvmdfvb7BXTenz6CV771H/+jz6PqtNiS0NIvXyth4Drb7gTtd1u17i9+6t1XVNMdGBf0888YTppa1O+oPSChlK36Pe6TtaJeTuu+82aRDei15J0Ndq9Urr+Txw4IBHwKYpGfo43vy914JxTNX69ev9pnmEK63IoVdNdOCs9xUff2Xotm/fbv6vi7JrgC/0AANhQINL7fnRUdba8+ndS6rljrRKgKYCaK+Z5txpr6jmVFp/1P3R0dpahkwfU++jQa72lu3atcujnQYFervO2qZ/nDS9QnvqNEjQ3lsti6ajuquydetWk6KhvXJ6aVTLW2luo870pcGBBiP+aG+gjrK/6qqrTJULPRY6olwvZ2t1CPdeYs1J1iBeL0Nrj7mvsljV7b3S9BL9Y635k/oFRHvjdZuVX6l/3LU6w7Rp08wx0R5yzb/WffOm+6Y0uNMSdtoLqudWZ0Lzpl8G9JhojqUee31crfKhx0CfX/cnEPR9o8dWA+o//vGP5v2gebj6XNorapXeqqnHH3/cPKaWxRo/frw5Xnoc9X2s7xvtyfdF34taWUBTe7R0l17+rw4to6bvC31/jRs3zrVd31/6vta0GV8pNnfccYdpr69Xy5Zde+215jjoF02tvasBu1Zy8JXO4e+9Fohjqu9rq5yYfhHQ1AetxKJBtwbzkUK/LOrvBj3OmoKiX2j0tek2PWZ6uzd93frZsz4vQJ0L2vA6ADXy7LPPmlHRP//5zyvcpjMmTZo0yfmzn/3MjPzWmbC0UoKv0ey+ZoLTUdu9evUyI9r1MfT2V155pcIofaWP269fP2dSUpIZ5a6P/+tf/9q5dOnSSvffqpRgLVpRIjU11dmzZ08zG9yOHTuqrMzwww8/OK+//npn+/btTbWIpk2bmlHx8+bN87jf+vXrTdUAHW2v97dGrluP9+WXX1b5XO4zwX3yySfOrKws83pbtmxp9lerE7jT0fJ6HHSUve7XyJEjXVUr3KtA6LkaM2aMMz093VQ8cH9O7yoQ1kj5ESNGmGOlI+Y7duxoKijoyPrqVFyo7sx/GzZscF5++eVm32NjY53nnnuux37XtAqERY+BViLRSibR0dHmf11ft25dpffr0qWLeS6tRlATDz74oDM5Odk1o19eXp55PZXNmKhVIfT9pK/fojPd6ax3ur1du3bO559/3me1BX/vtZoc0+pUgdBj17ZtW/Me09nsqlMF4rLLLqvwuN6VHPx9JnxVNvFXBaK677uXXnrJzPCmx+LMM890zpkzx3nllVeac+09m5/uv3fFGaAuRek/dR92AwBQc5pyoFc+tOdce3IRvrQXWK9i6BUdrRNs0asnAwcOdM0QCYQCATAAIKJomTwtn6b5sr4ur6Pu6WA3nfZcU0I0tUFLpGn+vlbfWLt2rUfqk7bRkoeaagSECjnAAICIorOM6UAzrTThXYcXoaF5yzp1sk7ZroNerYGAWs/bPfjVmfp0HIO2A0KJHmAAAADYCteOAAAAYCsEwAAAALAVAmAAAADYCoPgqkmL+mv5HS2CXpNpVgEAAFA3tLqvzvCpk+JUViWGALiaNPhltDEAAED409lOdfZIfwiAq0l7fq0DmpSUFJizAwAAgIA5evSo6bC04jZ/CICryUp70OCXABgAACB8VZWuyiA4AAAA2AoBMAAAAGyFABgAAAC2QgAMAAAAWyEABgAAgK0QAAMAAMBWCIABAABgKwTAAAAAsBUCYAAAANgKATAAAABsJaQB8KeffiqXX365tGrVykxZt2jRIo/bdZuv5amnnnK16du3b4Xbr7vuOo/HOXTokNx4443StGlTs+jPhw8frrPXCQAAgPAR0gD42LFjcu6558rzzz/v8/bc3FyPZc6cOSbAHT58uEe72267zaPdrFmzPG4fMWKErF+/XhYvXmwW/VmDYAAAANhPTCiffPDgwWbxp0WLFh7r//znP6Vfv37Srl07j+3x8fEV2lq+//57E/SuWrVKunfvbra9/PLL0rNnT9m8ebN07NgxIK8FAAAAkSGkAXBN7N+/Xz744AN59dVXK9z2xhtvyPz58yUjI8ME1A899JAkJiaa21auXGnSHqzgV/Xo0cNsW7Fihd8AuLi42CyWo0ePmv/LysrMAgAAgPBS3RgtYgJgDXw1qB02bJjH9htuuEFOP/100wO8ceNGuf/+++Wbb76RJUuWmNv37dsnzZs3r/B4uk1v82fatGkyderUCtu3bdsmTZo0CchrAoBIoOlq7h0CcXFxkpCQENJ9AgBfCgsLpV4FwJr/q8Fuo0aNKuT/Wjp16iQdOnSQrKws+eqrr+S8884z2zVv2JvT6fS53aKB9IQJEzx6gNu0aSPt27eXpKSkAL0qAAitn376yXWFS+nvt7S0NI/b77n3fjlUeNy1LblJvMx7+UWPdgAQDtx/n0V8APzZZ5+ZfN2FCxdW2VaD3oYNG8rWrVvNz9ozrOkT3vLy8kzKhD/aw6GLt+joaLMAQKTR33vufxwOHjwof/i/KVJY7HBtS0mMl/lzX5H09HRXb8pPRwolvedwSUjJkGP5+yVv5dtme2W/QwEgFKobo0VEADx79mzp2rWrqRhRlU2bNonD4ZCWLVuadR3sduTIEVmzZo1069bNbFu9erXZ1qtXr6DvOwCES/A78pYxkl/wv57cE0XHZfeeXMm67h5pltHaFdxqkGwFwBYNfpOatz75WNV8Pu+eGO1d9n5cAAiFkAbA2oPw448/utazs7NNibKUlBRp27at2aa/QP/+97/L008/7TMfVwfAXXrppeZS3HfffScTJ06ULl26yPnnn2/anH322TJo0CCTKmGVR7v99ttlyJAhVIAAYBv6u1SDX6snVx3YtlFyds2RuKSUGgW3tQm2ffUuA4AtA+C1a9easmYWK+d21KhRMm/ePPPzggULTL7u9ddfX+H+sbGx8t///leeffZZE0xrju5ll11mqkC4d4FrkDxu3DgZOHCgWb/iiiv81h4GgPrMvSe38KD/gcCBDrYr610GAFsFwDqLmwa3ldHeWl180YB3+fLlVT6P9ihrmTQAQGiC7UD1LgNAxM8EBwAAANS1iBgEBwCoG46SEsnJyXGt68+ljlIOP4B6hQAYAGAUFx6RHdnbZfzkKa4ykFaliLaO/5VKA4BIRwAMAPWUeymy6vTkOoqLpDwqRtJ6DJPUVpkelSLKSgmAAdQfBMAAYINJLmrSkxufnB70ShEAEEoEwABgg0ku6MkFgP8hAAYAG0xyQU8uAPwPATAA1BN1MckFANQH1AEGAACArRAAAwAAwFYIgAEAAGAr5AADQISVOFNJSUmSnp4eFrPFhXp/AKCmCIABIMJKnKmUxHiZP/eVOg86fc0WF8r9AYDaIAAGgAgrcXYsf7/sXf6mbNiwQTIzM6s1y1ug+JotTvcnb+XbZl8JgAFEAgJgAIiwEmfevbA1meUtUNxni1N5dfbMAHDqCIABIMJ498KGwyxv7nnB/nqkvXOHyRsGECoEwAAQxoPeKktvsHphQz3pRXV6pH3lDpM3DCBUCIABIIwHvYUivSEYPdLebcgbBhBKBMAAEMaD3sIhvaG6qtMj7Z47TN4wgFBhIgwACONBb/HN0kK9KwBQ7xAAAwAAwFZIgQAAhO2sd1SKABAMBMAAgLCd9Y5KEQCCgQAYABCWAwCpFAEgWAiAAQBhO+sdlSIABAOD4AAAAGArBMAAAACwFQJgAAAA2AoBMAAAAGyFQXAAgJBwlJRITk6Oa11/LnWUcjYABB0BMACgzhUXHpEd2dtl/OQpEhcXZ7adKDouu/fkSluHgzMCIKgIgAEAdc5RXCTlUTGS1mOYpLbKNNsObNsoObvmSFkpATCA4CIABgCETHxyuqvmb+HBfZwJAHWCQXAAAACwFQJgAAAA2AopEAAQQnl5eXL06FHXOpUQACD4CIABIITB78hbxkh+wXHXNiohAEDwEQADQIhoz68Gv+k9h0tCSobZRiUEAAg+AmAACDENfqmEAAB1h0FwAAAAsBUCYAAAANhKSAPgTz/9VC6//HJp1aqVREVFyaJFizxuv/nmm81296VHjx4ebYqLi2Xs2LGSlpYmCQkJcsUVV8ju3bs92hw6dEhuvPFGadq0qVn058OHD9fJawQAAEB4CWkAfOzYMTn33HPl+eef99tm0KBBkpub61r+/e9/e9w+fvx4effdd2XBggXy+eefS2FhoQwZMkTKyspcbUaMGCHr16+XxYsXm0V/1iAYABDeHCUlpjTctm3bXItWzwCAiB0EN3jwYLNUJi4uTlq0aOHztiNHjsjs2bPl9ddfl4svvthsmz9/vrRp00aWLl0ql1xyiXz//fcm6F21apV0797dtHn55ZelZ8+esnnzZunYsWMQXhkA4FQVFx6RHdnbZfzkKeZvgSUlMV7mz31F0tPTOcgA6mcViE8++USaN28uzZo1kz59+sijjz5q1tW6devE4XDIwIEDXe01naJTp06yYsUKEwCvXLnSpD1Ywa/SNArdpm38BcCaWqGLxSpUrz3L7r3LAFBb5eXlEh3dQBpE6eU4p9mmP8fERLu2ea/bqU1ZSZE0aBgnGT2HSUqrTNPmWP5++Wn1uyaNLSUlhTcfAA/VjdHCOgDW3uGrr75aMjMzJTs7Wx588EG56KKLTOCrvQH79u2T2NhYSU5O9rhfRkaGuU3p/1bA7E63WW18mTZtmkydOrXCdr381qRJk4C8PgD2lp+fL+d3y5JmKVES16jAbGvROlHSBg2QzIxYadKooMK6Lduc9TNp0jTJtClOiZLDziyTBkFnBABvmgob8QHwtdde6/pZe3WzsrJMMPzBBx/IsGHD/N7P6XSaAXMW95/9tfF2//33y4QJEzx6gDW1on379pKUdPIXMQCciu3bt8sXa9ZKZmqWJEUnmm17d2+RlYuXSO82vSUjLrHCut3bHM0/Kjlr1sqdo2+Wdu3a8QYE4MF9avmIDYC9tWzZ0gTAW7duNeuaG1xSUmKqPLj3Ah84cEB69erlarN///4Kj6W9B9pT7I/2MLvnnFmio6PNAgCnqkGDBlJWVi7lTpFyOfmFXH8uLS1zbfNep41e4iw3x47fxQC8Vff3QkTVAT548KDs2rXLBMKqa9eu0rBhQ1myZImrjVaK2LhxoysA1sFuOlhuzZo1rjarV68226w2AAAAsI+YUOdp/Pjjj651zfPVEmU6sEGXKVOmyPDhw03Au2PHDpk8ebKp9zt06FDTXgeyjR49WiZOnCipqanmPpMmTZLOnTu7qkKcffbZppTabbfdJrNmzTLbbr/9dlMqjQoQAAAA9hPSAHjt2rXSr18/17qVcztq1CiZOXOmbNiwQV577TUz2leDYG27cOFCSUw8mQumZsyYITExMXLNNddIUVGR9O/fX+bNm+fRBf7GG2/IuHHjXNUidLKMymoPAwAAoP4KaQDct29fMxjNnw8//LDKx2jUqJE899xzZvFHe4a1PjAAAAAQUTnAAAAAwKkiAAYAAICtEAADAADAVgiAAQAAYCsEwAAAALCViJoJDgAinc5CaU3VmZOTI6WO0lDvEgDYDgEwANRh8DvyljGSX3DcrJ8oOi679+RKW4eDcwAAdYgAGADqiPb8avCb3nO4JKRkyIFtGyVn1xwpKyUABoC6RA4wANQxDX6TmreW+GZpHHsACAECYAAAANgKATAAAABshQAYAAAAtkIADAAAAFshAAYAAICtEAADAADAVgiAAQAAYCsEwAAAALAVAmAAAADYClMhAwAiiqOkRHJyclzrJSUlEhsb61pPSkqS9PT0EO0dgEhAAAwAiBjFhUdkR/Z2GT95isTFxZlgeM/OHGmdebrENDz5Jy0lMV7mz32FIBiAXwTAABBEeXl5cvToUfOz9lqWOko53qfAUVwk5VExktZjmKS2ypQD2zbK9h1zJLnblWb9WP5+yVv5tjnm9AID8IcAGACCEOyqgwcPyh/+b4oUFjvM+omi47J7T660dZxcR+3FJ6dLUvPWUnhwn8e6OQ8cWABVIAAGgAAFvyNvGSP5Bcdd26yAN+u6e6RZRmvTW5mza46UlRIA12WOsCIvGIA7AmAACADt+dXgN73ncElIyTDbrIA3LinFo7cSdZcjbCEvGIA7AmAACCANfq1L8QS8oc8RVuQFA/BGAAwAqHfcc4IVecEA3DERBgAAAGyFABgAAAC2QgAMAAAAWyEABgAAgK0QAAMAAMBWCIABAABgKwTAAAAAsBUCYAAAANgKATAAAABshQAYAAAAtkIADAAAAFshAAYAAICtxIR6BwAACDZHSYnk5OS41pOSkiQ9PZ0DD9gUATAAoF4rLjwiO7K3y/jJUyQuLs5sS0mMl/lzXyEIBmyKABgAUK85ioukPCpG0noMk9RWmXIsf7/krXxbjh49SgAM2BQBMADAFuKT0yWpeWvzc16odwaAfQfBffrpp3L55ZdLq1atJCoqShYtWuS6zeFwyB/+8Afp3LmzJCQkmDY33XST7N271+Mx+vbta+7rvlx33XUebQ4dOiQ33nijNG3a1Cz68+HDh+vsdQIAACB8hDQAPnbsmJx77rny/PPPV7jt+PHj8tVXX8mDDz5o/n/nnXdky5YtcsUVV1Roe9ttt0lubq5rmTVrlsftI0aMkPXr18vixYvNoj9rEAwAAAD7CWkKxODBg83ii/bULlmyxGPbc889J926dZOdO3dK27ZtXdvj4+OlRYsWPh/n+++/N0HvqlWrpHv37mbbyy+/LD179pTNmzdLx44dA/qaAAAAEN4iKgf4yJEjJsWhWbNmHtvfeOMNmT9/vmRkZJiA+qGHHpLExERz28qVK00wbQW/qkePHmbbihUr/AbAxcXFZrHoYAlVVlZmFgBwV15eLtHRDaRBlF5ac5pt+nNMTLRrW1Xr1bkPbQJzTPVc6Tnj9zlQv1T3Mx0xAfCJEyfkvvvuM+kMWr/RcsMNN8jpp59ueoA3btwo999/v3zzzTeu3uN9+/ZJ8+bNKzyebtPb/Jk2bZpMnTq1wvZt27ZJkyZNAva6ANQP+fn5cn63LGmWEiVxjQrMthatEyVt0ADJzIiVJo0Kqlyvzn1oc+rHtDglSlp3y5K8vDwCYKCeKSwsrD8BsA6I04Ft+m39hRdeqJD/a+nUqZN06NBBsrKyTN7weeedZ7Zrr7E3p9Ppc7tFA+kJEyZ49AC3adNG2rdv7xGAA7Cvn376yXV1qKioSJZ/sVrap2ZJUvTJK1B7d2+RlYuXSO82vSUjLrHK9erchzanfkyP5h+VnDVr5c7RN0u7du1C9v4BEHjW7+SID4A1+L3mmmskOztbPv744yqDTw16GzZsKFu3bjU/a8/w/v37K7TTb/6aMuGPFku3Cqa7i46ONgsAe9PfITeNvl3yC46b9RNFx2X3nlz5WYlDmsjJL9flTpHS0jLzf7lEVblenfvQJjDHtKysXBo0aMDvc6CeqW6MFtIqENUNfjWYXbp0qaSmplZ5n02bNpn7tWzZ0qzrYDfNHV6zZo2rzerVq822Xr16BXX/AdTvXgYNftN7DpfTLrtLUroMkrJyp5SVOkK9awCAcO4B1jyNH3/80bWuvbxaoiwlJcXU/f31r39tUhn+9a9/mTwtK2dXb4+NjTX5uDoA7tJLL5W0tDT57rvvZOLEidKlSxc5//zzTduzzz5bBg0aZFIlrPJot99+uwwZMoQKEABOWUJKhplcofCg/zEFAIDwEtIAeO3atdKvXz/XupVzO2rUKJkyZYq89957Zv2Xv/ylx/2WLVtmJsDQIPi///2vPPvssyaY1hzdyy67zFSBcO8C1yB53LhxMnDgQLOutYR91R4GAABA/RfSAFiDWB2M5k9ltykNeJcvX17l82iPsZZJAwAAAMI6BxgAAAAINAJgAAAA2AoBMAAAAGyFABgAAAC2QgAMAAAAWyEABgAAgK2E/VTIABAuUx+7zzGfk5MjpY7SkO4TAKB2CIABoBrB78hbxpipjy0nio7L7j250tbB1McAEGkIgAGgCtrzq8Fves/hZupjdWDbRsnZNUfKSgmAASDSEAADQDVp8JvUvLX5ufDgPo5bBHOUlJg0FndJSUmSnp4esn0CUHcIgAEAtlJceER2ZG+X8ZOnSFxcnGt7SmK8zJ/7CkEwYAMEwAAAW3EUF0l5VIyk9Rgmqa0yzbZj+fslb+XbJt2FXmCg/iMABoAqqj5Q8aF+ik9Od6W0qLyQ7g2AukQADABVVH2g4gMA1C8EwABQRdUHKj4AQP3CTHAAUEXVh/hmaRwjAKhHCIABAABgKwTAAAAAsBUCYAAAANgKATAAAABshQAYAAAAtkIADAAAAFshAAYAAICtEAADAADAVgiAAQAAYCsEwAAAALAVAmAAAADYCgEwAAAAbCUm1DsAAKGWl5cnR48eda3n5ORIqaM0pPsEAAgeAmAAYvfgd+QtYyS/4Lhr24mi47J7T660dThCum8AgOAgAAZga9rzq8Fves/hkpCSYbYd2LZRcnbNkbJSAmAAqI9qlQOcnZ0d+D0BgBDS4DepeWuzxDdL41wAQD1WqwD4jDPOkH79+sn8+fPlxIkTgd8rAADqmKOkxOR/b9u2zSyaHgOgfqpVAPzNN99Ily5dZOLEidKiRQu54447ZM2aNYHfOwAA6kBx4RHZkb1dxk+eItfdeqdZNDecIBion2oVAHfq1EmmT58ue/bskblz58q+ffukd+/ecs4555jt/MIAAEQSR3GRlEfFSFqPYXLaZXeZnHDNDXevDgKg/jilOsAxMTEydOhQeeutt+SJJ54wl4wmTZokrVu3lptuuklyc3MDt6cAAARZfHK6yQO3BkQCqJ9OqQrE2rVrZc6cObJgwQJJSEgwwe/o0aNl79698sc//lGuvPJKUiMAhHXdX2r+oqqcYHdJSUmSnp7OQQPsGABrmoOmPmzevFkuvfRSee2118z/DRqc7FA+/fTTZdasWXLWWWcFen8BIKB1f6n5i6pyguPi4lzbUxLjZf7cVwiCATsGwDNnzpRbb71VbrnlFjMIzpe2bdvK7NmzT3X/ACCodX+p+YuqcoJTW2Wabcfy90veyrfNe4heYMCGAfDWrVurbBMbGyujRo2qzcMDQJ3V/S08uI+jjSpzgi0URgNsPAhO0x/+/ve/V9iu21599dVA7BcAAAAQPgHw448/LmlpFWdKat68uTz22GPVfpxPP/1ULr/8cmnVqpVERUXJokWLPG53Op0yZcoUc3vjxo2lb9++smnTJo82xcXFMnbsWLM/OhDviiuukN27d3u0OXTokNx4443StGlTs+jPhw8frvHrBgAAgE0DYB0VqwPdvGVmZsrOnTur/TjHjh2Tc889V55//nmftz/55JNmwJ3e/uWXX5p84wEDBkhBQYGrzfjx4+Xdd981lSg+//xzKSwslCFDhkhZWZmrzYgRI2T9+vWyePFis+jPGgQDAADAfmqVA6w9vd9++62cdtppFWaIS01NrfbjDB482Cy+aO/vM888Iw888IAMGzbMbNP0ioyMDPnb3/5mZp87cuSIGWj3+uuvy8UXX2za6PTMbdq0kaVLl8oll1wi33//vQl6V61aJd27dzdtXn75ZenZs6epYtGxY8faHAIAAADYKQC+7rrrZNy4cZKYmCgXXnih2bZ8+XL53e9+Z24LhOzsbDPD3MCBA13btBRNnz59ZMWKFSYAXrdunTgcDo82mi6hM9VpGw2AV65cadIerOBX9ejRw2zTNv4CYE2t0MVi1QzVnmX33mUAkaW8vFyioxtIgyi9BOY0/8fERLvWlfe2YLUJ5XPX1zbBfm597+h7iL8DQHiq7mezVgHwI488YtIg+vfvb2aDU/oLQWd/q0kOcGU0+FXa4+tO163C5NpGq00kJydXaGPdX//XHmtvus1q48u0adNk6tSpFbbrbHdNmjSp5asCEGr5+flyfrcsaZYSJXGNCqRF60RJGzRAMjNipUmjk+lV3tuC1SaUz11f2wTzuYtToqR1tyxTS5oAGAhPmgobtABYg86FCxfKn/70J5P2oAPUOnfubHKAA00Hx3mnRnhv8+bdxlf7qh7n/vvvlwkTJnj0AGtqRfv27c1MQAAi0/bt2+WLNWslMzVLkqITZe/uLbJy8RLp3aa3ZMQlmjbe24LVJpTPXV/bBPO5j+YflZw1a+XO0TdLu3bt6uT9CqBmrCv2QZ0K+cwzzzRLMFgTbGgvbcuWLV3bDxw44OoV1jYlJSWmyoN7L7C26dWrl6vN/v37Kzy+foP37l12p+kW7rP/WKKjo80CIDLpjJVlZeVS7hQplyjzf2lpmWtdeW8LVptQPnd9bRPs59b3jr6H+DsAhKfqfjZrVQVCL/3o4DOtrqCDzy666CKPJRC0yoQGr0uWLHFt02BXc42t4LZr167SsGFDjza5ubmyceNGVxsd7KaD5dasWeNqs3r1arPNagMAAAD7qFUPsA52mzdvnlx22WVmwFlVKQmV5Wn8+OOPHgPftERZSkqKmUpZS5xpTnGHDh3Moj/Hx8ebwFvpQLbRo0fLxIkTTfUJvd+kSZNMOoZVFeLss8+WQYMGyW233SazZs0y226//XZTKo0KEAAAAPZTqwBYa+6+9dZbcumll57Sk69du1b69evnWrdybnUKZQ2w7733XikqKpK77rrLpDloJYePPvrIVJ+wzJgxwwzEu+aaa0xbHZin93XvAn/jjTdM1QqrWoROluGv9jAAAP44SkpcA7GVjglJT0/ngAERptaD4M4444xTfnKd2U0Ho/mjPcs6E5wu/jRq1Eiee+45s/ijPcNaHxgAgNoqLjwiO7K3y/jJU1xjRFIS42X+3FcIgoEIU6scYE05ePbZZysNXgEAqE8cxUVSHhUjaT2GyWmX3SXpPYdLfsHxao86BxDhPcA65fCyZcvkP//5j5xzzjlmIJq7d955J1D7BwBAWIlPTpek5q3Nz3mh3hkAdRcAN2vWTIYOHVq7ZwQAAAAiLQCeO3du4PcEAAAACNccYFVaWipLly41pcUKCk5OE7l3795qT0EHAAAAREwPsJaA0dq6O3fulOLiYhkwYIApTfbkk0/KiRMn5MUXXwz8ngIAAACh6gHWiTCysrJMbd7GjRu7tmte8H//+99A7BcAAAAQXlUgvvjiC1MP2F1mZqbs2bMnUPsGAAAAhEcAXF5eLmVlZRW2796922OWNgAItby8PI86rZrCVeooDek+AQAiMADWnN9nnnlGXnrpJdeMbTr47aGHHjrl6ZEBIJDB78hbxpjJCiwnio7L7j250tbh4EADgE3VKgCeMWOG9OvXT37+85+bQW8jRoyQrVu3Slpamrz55puB30sAqAXt+dXgV2fsSkjJMNsObNsoObvmSFkpATAA2FWtAuBWrVrJ+vXrTbD71VdfmZSI0aNHyw033OAxKA4AwoEGv9bMXYUH94V6dwAAkRgAKw10b731VrMAAAAA9ToAfu211yq9/aabbqrt/gAAAADhFwBrHWB3DodDjh8/bsqixcfHEwADAACgfk2EoRNguC9aAWLz5s3Su3dvBsEBAACg/gXAvnTo0EEef/zxCr3DAAAAQL0YBOdLdHS07N27N5APCQC1nviCSS8AAAELgN977z2PdafTKbm5ufL888/L+eefX5uHBICAT3zBpBcAgIAFwFdddZXHus4El56eLhdddJE8/fTTtXlIAAj4xBdMegEACFgArBNfAEC4T3zBpBcAgKAOggMAAADqbQ/whAkTqt12+vTptXkKAAAAIHwC4K+//lq++uorKS0tlY4dO5ptW7ZsMVUgzjvvPI/cYAAAACDiA+DLL79cEhMT5dVXX5Xk5GSzTSfEuOWWW+SCCy6QiRMnBno/AQAAgNAFwFrp4aOPPnIFv0p/fuSRR2TgwIEEwAAAW3CUlJh60+6SkpJMZSQA9SwA1lJD+/fvl3POOcdj+4EDB6SgoCBQ+wYAQNgqLjwiO7K3y/jJUyQuLs61vUlstDzx6MOSmppq1gmIgXoSAA8dOtSkO2hPcI8ePcy2VatWye9//3sZNmxYoPcRAICw4ygukvKoGEnrMUxSW2Wabfm7f5R1b/1Fxoyb5AqKUxLjZf7cV+gVBiI9AH7xxRdl0qRJMnLkSHE4HCcfKCZGRo8eLU899VSg9xEAgLAVn5xu6k4rrT3tHhQfy98veSvfNldOSYsAIjwAjo+PlxdeeMEEu9u2bTNTIZ9xxhmSkJAQ+D0EACCCg+K8UO8MgMBOhJGbm2uWM8880wS/GggDAAAA9S4APnjwoPTv398EvpdeeqkJgtWYMWOoAAEAAID6FwDfc8890rBhQ9m5c6dJh7Bce+21snjx4kDuHwAAABD6HGCtAfzhhx9K69Yn85ssHTp0qFAPEQAAAIj4HuBjx4559PxafvrpJ49aiAAAAEC9CIAvvPBCee2111zrUVFRUl5ebqpC9OvXL5D7BwB+5eXlmUo01qJXoEodpRwxAEDgUyA00O3bt6+sXbtWSkpK5N5775VNmzZJfn6+fPHFF7V5SACocfA78pYxkl9w3LXtRNFx2b0nV9r+//rkQDhgumSgngTAP//5z+Xbb7+VmTNnSnR0tEmJ0Bng7r77bmnZsmXg9xIAvOjEAhr8pvccLgkpGWbbgW0bJWfXHCkrJQBGeE+XzOxwQIQFwDrz28CBA2XWrFkyderU4OwVAFSTBr/us3AB4T5dMrPDAREYAGv5s40bN5q8XwAAULOZ4RSzwwEROAjupptuktmzZwd+bwAAAIBwzAHWgW+vvPKKLFmyRLKyssw0yO6mT58eqP0DAAAAQtcDvH37dlPuTFMgzjvvPElKSpItW7bI119/7VrWr18f0B087bTTTLqF96ID7tTNN99c4bYePXp4PEZxcbGMHTtW0tLSTLB+xRVXyO7duwO6nwAAAKiHPcA601tubq4sW7bMNfXxX/7yF8nIODkCOxi+/PJLKSsrc61r8D1gwAC5+uqrXdsGDRokc+fOda3HxsZ6PMb48ePl/ffflwULFkhqaqpMnDhRhgwZIuvWrTNVLAAAAGAfNQqAnU6nx/p//vMfUwItmNLT0z3WH3/8cWnfvr306dPHtU1Ly7Ro0cLn/Y8cOWLylV9//XW5+OKLzbb58+dLmzZtZOnSpXLJJZcEdf8BAABQD3KA/QXEwaa5xxq8TpgwwaMKxSeffCLNmzeXZs2amcD40UcfNetKe3mt0m2WVq1aSadOnWTFihV+A2BNm9DFveao0t5o9x5pAKGh6VjR0Q2kQZTmcp38XaQ/x8REu7Z5r4dbm3Dfv0hsE+77Z7XR966+h/l7AgRWdT9TNQqArRxb7211ZdGiRXL48GGT92sZPHiwSYfIzMyU7OxsefDBB+Wiiy4yga/2DO/bt8+kRCQnJ3s8lqZt6G3+TJs2zWedY51utUmTJgF+ZQBqSmeePL9bljRLiZK4RgVmW4vWiZI2aIBkZsRKk0YFFdbDrU24718ktgn3/VPFKVHSuluWmc2QABgIrMLCwuCkQGjwac1mc+LECbnzzjsrVIF45513JBg0lUEDXu3BtWgeskV7dbUqhQbDH3zwgZmdrrLXUlnwfv/995ueZvceYE2b0PQLHfwHILR0UO4Xa9ZKZmqWJEUnmm17d2+RlYuXSO82vSUjLrHCeri1Cff9i8Q24b5/6mj+UclZs1buHH2ztGvXLqifE8Bujv7/K/YBDYBHjRrlsT5y5EipKzk5OSZnt6rgWqdi1gB469atZl1zgzV14tChQx69wAcOHJBevXr5fRwN8t2nrbTooDkGzgGh16BBAykrK5dyp0i5nPwyqz+Xlpa5tnmvh1ubcN+/SGwT7vtntdH3rr6H+XsCBFZ1P1M1CoDdKy3UNX1uzeu97LLLKm138OBB2bVrlwmEVdeuXc3sdVqz+JprrjHbtJKFVpN48skn62TfAQSGXjK2vt3rl+JSRymHFgBQt4Pg6ooOFNAAWHugY2JiPPI8pkyZIsOHDzcB744dO2Ty5Mmm3u/QoUNNm6ZNm8ro0aNN6TMtgZaSkiKTJk2Szp07u6pCAIiM4HfkLWMkv+C4WT9RdFx278mVtg5HqHcNABBhIiIA1tSHnTt3yq233lqhm3vDhg3y2muvmcFxGgT369dPFi5cKImJJ3Ot1IwZM0zgrD3ARUVF0r9/f5k3bx6XnoAIoj2/Gvym9xwuCSkZcmDbRsnZNUfKSgmAAQD1MADWEma+Sq41btxYPvzwwyrv36hRI3nuuefMAiCyafCb1Ly1FB70X8UFCHeOkhKTxmPRwdXede8B2DwABgCgviguPCI7srfL+MlTXIOtUxLjZf7cVwiCgTpCAAwAQB1yFBdJeVSMpPUYJqmtMuVY/n7JW/m2SfOhFxioGwTAAACEQHxyuknnUXmcAaBONajbpwMAAABCiwAYAAAAtkIADAAAAFshAAYAAICtEAADAADAVgiAAQAAYCsEwAAAALAVAmAAAADYCgEwAAAAbIUAGAAAALZCAAwAAABbIQAGAACArcSEegcAwJe8vDw5evSoaz0nJ0dKHaUcLADAKSMABhCWwe/IW8ZIfsFx17YTRcdl955caetwhHTfAACRjwAYQNjRnl8NftN7DpeElAyz7cC2jZKza46UlRIAAwBODQEwgLClwW9S89bm58KD+0K9O0DQOEpKTJqPu6SkJElPT+eoA0FAAAwAQAgVFx6RHdnbZfzkKRIXF+fanpIYL/PnvkIQDAQBATAAACHkKC6S8qgYSesxTFJbZZptx/L3S97Kt006EL3AQOARAAMAEAbik9NdKT8qL6R7A9RvBMAAwq7sGSXPAADBRAAMIOzKnlHyDAAQTATAAMKu7BklzwAAwcRUyADCruxZfLO0UO8KAKAeIwAGAACArRAAAwAAwFYIgAEAAGArDIIDACACpkdmamQgcAiAAQCIgOmRmRoZCBwCYAAAwnx6ZKZGBgKLABgAgAiYHnmvV0qEIi0CqB0CYAAAIjAlQpEWAdQOATAAABGWEqFIiwBqjwAYAIAITIlQeSHdGyByUQcYAAAAtkIADAAAAFshAAYAAICtEAADAADAVgiAAQAAYCsEwAAAALCVsA6Ap0yZIlFRUR5LixYtXLc7nU7TplWrVtK4cWPp27evbNq0yeMxiouLZezYsZKWliYJCQlyxRVXyO7du0PwagAAABAOwjoAVuecc47k5ua6lg0bNrhue/LJJ2X69Ony/PPPy5dffmmC4wEDBkhBQYGrzfjx4+Xdd9+VBQsWyOeffy6FhYUyZMgQKSsrC9ErApCXlyfbtm1zLTq9a6mjlAMDAKgTYT8RRkxMjEevr3vv7zPPPCMPPPCADBs2zGx79dVXJSMjQ/72t7/JHXfcIUeOHJHZs2fL66+/LhdffLFpM3/+fGnTpo0sXbpULrnkkjp/PYDdafA78pYxkl9w3LXtRNFx2b0nV9o6HCHdNwCAPYR9ALx161aT4qBzn3fv3l0ee+wxadeunWRnZ8u+fftk4MCBrrbapk+fPrJixQoTAK9bt04cDodHG32sTp06mTaVBcCaOqGL5ejRo+Z/7Tmm9xiovcOHD8uR4ycko9dwSUjJMNvytm+SPbnzxFnmkAbilAZR+uU32vyv68p7W31oE+77F4ltwn3/At0mOrqBlJeX83cJ+P+qG6OFdQCsAe9rr70mZ555puzfv18eeeQR6dWrl8nz1eBXaY+vO13Xy6lK28TGxkpycnKFNtb9/Zk2bZpMnTq1wna9XNukSZMAvDrAnvLz8+X8blnS7IwWEpeQaLYVND5N2pQOkMyMWGnSqEBatE6UtEH/W1fe2+pDm3Dfv0hsE+77F8g2xSlR0rpblrmqQscMcJKmukZ8ADx48GDXz507d5aePXtK+/btTapDjx49zHYdGOedGuG9zVt12tx///0yYcIEjx5gTZ3Q509KSqrlKwLqv59++sl1xcSinxkdiKq2b98uX6xZK5mpWZIUfTIA3rt7i6xcvER6t+ktGXGJFdbra5tw379IbBPu+xfINkfzj0rOmrVy5+ibzZVRAFLh709EBsDetIqDBsKaFnHVVVeZbdqT27JlS1ebAwcOuHqFNXe4pKREDh065NELrG20J7kymk6hi7fo6GizAKhIe6JuGn27R36vSkmMl/lzX5H09HRp0KCBlJWVS7lTpFxOfhHVn0tLy1zbvNfra5tw379IbBPu+xfoNvpZ0s8Uf5eAk6r7WQj7KhDuNCf3+++/NwHv6aefbgLcJUuWuG7XYHf58uWu4LZr167SsGFDjzZaSWLjxo1VBsAAavfNW4Pf9J7D5bTL7jKL/qzbqvutHACAYAvrHuBJkybJ5ZdfLm3btjW9tpoDrH9ER40aZVIYtMSZDorr0KGDWfTn+Ph4GTFihLl/06ZNZfTo0TJx4kRJTU2VlJQU85jai2xVhQAQeDq4Lal5a9f63pISV24+Jc8AAKEW1gGwTlhx/fXXm5xCvXSqeb+rVq2SzMxMc/u9994rRUVFctddd5k0Bx0099FHH0li4sn8KDVjxgxTSu2aa64xbfv37y/z5s3jchFQR4oLj8iO7O0yfvIUk1ZEyTMgcBxuXy6tfHv9ewkgggNgnbyiMtoLrDPB6eJPo0aN5LnnnjMLgLrnKC6S8qgYSesxTFJbZcqBbRslZ9ccKSul5i8QyC+X3vn2ACI0AAZQf8Qnp5u0iMKDlZcgBFC7L5fH8vdL3sq3TaogATBQOQJgAADqwZdLlRfqnQEiRERVgQAAAABOFT3AAE659q9V4owKDwCASEAADOCUgt+Rt4xxTXxBhQcAQCQgAAYQkIkvtPYvFR4AAJGAHGAAAZv4Ir5ZGkcTABD2CIABAABgKwTAAAAAsBUCYAAAANgKATAAAABshQAYAAAAtkIADAAAAFshAAYAAICtMBEGgFpNe6yY+hgIL46SEvO5dJeUlCTp6ekh2ycgHBEAA6jVtMeKqY+B8FFceER2ZG+X8ZOnSFxcnGt7SmK8zJ/7CkEw4IYAGECtpj1WTH0MhA9HcZGUR8VIWo9hktoq02w7lr9f8la+bT6/9AID/0MADKBW0x6rwoP7OHpAmIlPTnd9RlVeSPcGCE8EwAAA2CgvmJxggAAYAABb5QU3iY2WJx59WFJTU13tCIphN/QAAwBgk7zg/N0/yrq3/iJjxk1ioBxsjQAYAACb5AVr3j4D5QACYAAAbIeBcrA7ZoIDAACArRAAAwAAwFYIgAEAAGArBMAAAACwFQJgAAAA2AoBMAAAAGyFOsAA/MrLy5OjR4+an3Uq1VJHKUcLABDxCIAB+A1+R94yRvILjpv1E0XHZfeeXGnrcHDEAAARjQAYgE/a86vBb3rP4ZKQkiEHtm2UnF1zpKyUABgAENnIAQZQKQ1+dQrV+GZpHCkAQL1AAAwAAABbIQAGAACArRAAAwAAwFYYBAcAgM05SkpMqUNLUlKSpKenh3SfgGAiAAYAwMaKC4/IjuztMn7yFImLizPbmsRGyxOPPiypqalmnYAY9Q0BMAAXJr4A7MdRXCTlUTGS1mOYpLbKlPzdP8q6t/4iY8ZNcgXEKYnxMn/uK/QKo94gAAZgMPEFYG/xyemm5GHhwX0eAfGx/P2St/JtUxuctAjUFwTAAAwmvgDgKyBWeRwa1DNhXQVi2rRp8qtf/UoSExOlefPmctVVV8nmzZs92tx8880SFRXlsfTo0cOjTXFxsYwdO1bS0tIkISFBrrjiCtm9e3cdvxog/Hp8t23b5lp0AEypo5SJLwAA9V5Y9wAvX75c7r77bhMEl5aWygMPPCADBw6U7777zgSylkGDBsncuXNd67GxsR6PM378eHn//fdlwYIFJqF/4sSJMmTIEFm3bp1ER0fX6WsCwjHdQZ0oOi679+RKWwdTHQMA6rewDoAXL17ssa5BrvYEa+B64YUXurZrkn6LFi18PsaRI0dk9uzZ8vrrr8vFF19sts2fP1/atGkjS5culUsuuSTIrwII/3QHdWDbRsnZNUfKSgmAAQD1W1gHwL6CWZWSkuKx/ZNPPjGBcbNmzaRPnz7y6KOPmnWlwbLD4TA9x5ZWrVpJp06dZMWKFX4DYE2b0MU9YFBlZWVmASJZeXm5REc3kMTUDElK/5nZdjx/n8TEREuDKM2Ncpr/K1tXtDm148MxDfz7iWManGOqvy/09wZ//xDuqvsejZgA2Ol0yoQJE6R3794meLUMHjxYrr76asnMzJTs7Gx58MEH5aKLLjKBr/YM79u3z6REJCcnezxeRkaGua2y/OOpU6dW2K65kk2aNAnwqwPqVn5+vpzfLUuapURJXKMCs61F60RJGzRAMjNipUmjgirXq3Mf2nBM6/p9wHsw8Me0OCVKWnfLMqlTBMAId4WFhfUrAP7tb38r3377rXz++ece26+99lrXzxoYZ2VlmWD4gw8+kGHDhlUaUOuAOX/uv/9+E3C79wBr2kT79u1NQXAgkm3fvl2+WLNWMlOzJCk60Wzbu3uLrFy8RHq36S0ZcYlVrlfnPrThmNb1+4D3YOCP6dH8o5KzZq3cOfpmadeuXQh+YwHVZ12xrxcBsFZweO+99+TTTz+V1q1PlmTxp2XLliYA3rp1q1nX3OCSkhI5dOiQRy/wgQMHpFevXn4fR3uPrQLg7nTQHAPnEOkaNGggZWXlUu4UKZeTXwT159LSMte2qtarcx/acEzr+n3AezA4x/RE0QnZtWuX+d1hYXY4hKPqxmhhHQBrL60Gv++++67J8z399NOrvM/BgwfNh1QDYdW1a1dp2LChLFmyRK655hqzLTc3VzZu3ChPPvlk0F8DAAD1bapkxexwiGRhHQBrCbS//e1v8s9//tPUArZydps2bSqNGzc2eR5TpkyR4cOHm4B3x44dMnnyZFPvd+jQoa62o0ePNqXPtASaDqCbNGmSdO7c2VUVAgAAVG+qZMXscIh0YR0Az5w50/zft2/fCuXQdAIM7ebesGGDvPbaa3L48GETBPfr108WLlxoAmbLjBkzJCYmxvQAFxUVSf/+/WXevHmkMgAAUIuZ4RSzwyGShX0KRGW0F/jDDz+s8nEaNWokzz33nFkAAABgb2E9FTIAAAAQaATAAAAAsJWwToEAUDtasN67FqKWA9RJYVROTo6UOko5vAAAWyIABuph8DvyljGSX3Dctc1RUiJ7duZI68zTJaZhjJwoOi679+RKW4cjpPsKAEAoEAAD9Yz2/Grwm95zuCSkZJhtB7ZtlO075khytytNGSNdz9k1R8pKCYABAPZDAAzUUxr8WiWLCg/u8yhjZK0DAGBHBMAAAKDGNLVKxxNYmBoZkYQAGKhng94Y4AYgFNMjMzUyIgkBMFDPBr0xwA1AXU+PzNTIiDQEwEA9G/TGADcAoZgemamREUmYCAOoZ4Pe4pulhXpXAAAIawTAAAAAsBVSIIAIn+WNQW8AwrEqhKIyBMIVATAQ4bO8MegNQDhWhVBNYqPliUcfltTUVLNOQIxwQQAM1INZ3pjVDUA4VYVQ+bt/lHVv/UXGjJtEqTSEHQJgoJ7M8gYA4VQVQn83USoN4YoAGAAABA2l0hCOqAIBAAAAW6EHGAAA1AkqRSBcEAADAICQVYpISYyX+XNfkfT0dM4C6gwBMAAACEmliGP5+yVv5dumwg0BMOoSATAQQRNfMOkFgPo0KE7lhXRvYFcEwEAETXzBpBcA6nteMJNloC4QAAMRNPEFk14AqO95weQEoy4QAAMRNPEFk14AqM95weQEo64QAAMAgLDJC97rlRKhSkpKJDY21rVOmgROFQEwEILBbP5+qXuvM+gNgN1TIjRHeM/OHGmdebrENDwZtpAmgVNFAAyEYDCbr1/qvn7JM+gNgN1LpenYh+075khytytJk0DAEAADIRjM5uuXuve61SZn1xwpK3VwngDYMiXCGvvgvo3SaThVBMBACAaz+fql7uuXPIPeAAAIvAZBeEwAAAAgbNEDDAAAInryDEVlCNQEATAQpAoP/DIGgLqpFKGoDIGaIAAGglDhQTWJjZYnHn1YUlNTKWcGAEGsFMEEGqgpAmDgFHt89TLcgfyj0vLCa10VHvJ3/yjr3vqLjBk3yfRQUM4MAALLfcCw+b1cye9pC1fmYCEABk6xx9cV3CameFRvcO+hoJwZANRdXvDBgwflD/83RQqLPUtIkiYBCwEwcIo1fSsLbr1LnAEAgp8XbHVMZF13jzTLONkxQZoE3BEAA1XwvoxmTU9s1fQluAWA8MoLtjom4pL+d2VOMYEGLATAQA0HuJHPCwDhqaqrbt7l08gJti8CYKCGA9zI5wWA+lE+zb1aj7+g2NdgupKSEomNjfV7H4Q/AmDYmvcvNu+BE/4GuAEAIjtNwrtaj6+Bcr6uAmov8p6dOdI683SJaXgyjGJwXeQhAIZtA15fo4S9B07Q2wsA9TdNwlc94b3L35QNGzZIZmam36uA23fMkeRuV5r7MbguMtkqAH7hhRfkqaeektzcXDnnnHPkmWeekQsuuCDUu4Vq9Mz6urzk3cb7kpT3ur/eXfdRwt4DJ+jtBQD71BP2V03C11VA9/vtJbc44tgmAF64cKGMHz/eBMHnn3++zJo1SwYPHizfffedtG3bNtS7Bze+Ljl552l5B7Pel6R8XaLy17vrPkqYgBcA7MtfNQlfZS5rmltcVSdNdduQbxwYtgmAp0+fLqNHj5YxY8aYde39/fDDD2XmzJkybdo0idRe0Lp67kB+eKtq433JyVeelq9g1v2SlPe6oncXAFAdNanhXp3c4up00lSnja/gOlCBdF41Zs4LVJtwYIsAWE/8unXr5L777vPYPnDgQFmxYoXP+xQXF5vFcuTIEfP/oUOHpKysLMh7fLKH8zdjx8uhY0WubYmx0fLg5PskOTk5qM+tr/FP056QguJS17bSEofk7tklrVpnSnTD6ArrgWxTXFQke/ftl5RjR6VRQqI4jh2RqOiG0viM7pKUcvIDdDh3h8ieXCktOiZlxcfFWVos0Q2ixFlywue6srYd279LDjUQOfbTHo915b2NNhyLYL4PeH9xTCPhfRDu+xcubay/N/7+ZpXu3CWx7bLMNu/16rYpyMuVb5a/I2PGTpTYuNha/531jil8/d0PZJvkhMYy87lnPHrEg8UKvp1OZ+UNnTawZ88ePQrOL774wmP7o48+6jzzzDN93uehhx4y92HhGPAe4D3Ae4D3AO8B3gO8BySijsGuXbsqjQ1t0QNsiYqK8ljXbwfe2yz333+/TJgwwbVeXl4u+fn55tuLv/sgMN/c2rRpI7t27TKXTBBZOH+Rj3MY2Th/kY9zeGo0tisoKJBWrVpV2s4WAXBaWppER0fLvn2euTwHDhyQjIyTZU28ad6Oe11A1axZs6DuJ/5Hg18C4MjF+Yt8nMPIxvmLfJzD2mvatGmVbf5/Rkv9ponfXbt2lSVLlnhs1/VevXqFbL8AAABQ92zRA6w0neHGG2+UrKws6dmzp7z00kuyc+dOufPOO0O9awAAAKhDtgmAr732WlNZ4eGHHzYTYXTq1En+/e9/m5leED407eShhx6qkH6CyMD5i3ycw8jG+Yt8nMO6EaUj4erouQAAAICQs0UOMAAAAGAhAAYAAICtEAADAADAVgiAAQAAYCsEwAi5Rx991NRjjo+Pr/ZkIzp2c8qUKWaml8aNG0vfvn1l06ZNQd9XVKRzv2uJQS08rov+fPjw4UoP1c0332xmVHRfevToweGtIy+88IKcfvrp0qhRI1Mj/bPPPqu0/fLly007bd+uXTt58cUXOVcRcv4++eSTCp81XX744Yc63Wec9Omnn8rll19u/nbpeVi0aFGVh4bPX3AQACPkSkpK5Oqrr5bf/OY31b7Pk08+KdOnT5fnn39evvzyS2nRooUMGDDATH+IujVixAhZv369LF682Cz6swbBVRk0aJApSWgtWpYQwbdw4UIZP368PPDAA/L111/LBRdcIIMHDzZ10X3Jzs6WSy+91LTT9pMnT5Zx48bJ22+/zemKgPNn2bx5s8fnrUOHDnW2z/ifY8eOybnnnmv+dlUHn78g0jJoQDiYO3eus2nTplW2Ky8vd7Zo0cL5+OOPu7adOHHC3PfFF18M8l7C3XfffadlFJ2rVq1ybVu5cqXZ9sMPP/g9WKNGjXJeeeWVHMwQ6Natm/POO+/02HbWWWc577vvPp/t7733XnO7uzvuuMPZo0ePoO4nAnP+li1bZj6Phw4d4pCGGT0v7777bqVt+PwFDz3AiDj6jXjfvn0ycOBAj8Lhffr0kRUrVoR03+xm5cqVJu2he/furm2ayqDbqjoXemm2efPmcuaZZ8ptt90mBw4cqIM9tje92rJu3TqPz47SdX/nS8+xd/tLLrlE1q5dKw6HI6j7i1M/f5YuXbpIy5YtpX///rJs2TIObYTg8xc8BMCIOBr8qoyMDI/tum7dhro7FxrEetNtlZ0LvWT7xhtvyMcffyxPP/20SWO56KKLpLi4OMh7bG8//fSTlJWV1eizo9t9tS8tLTWPh/A+fxr0vvTSSyZl5Z133pGOHTuaIFhzURH++PwFj22mQkbd0gFqU6dOrbSNBj1ZWVm1fg4dQOBOryh5b0Nwz5+v81Cdc6FTk1t0WnJ9H+i05B988IEMGzaM0xZkNf3s+GrvazvC7/xpwKuLpWfPnrJr1y7585//LBdeeGHQ9xWnjs9fcBAAIyh++9vfynXXXVdpm9NOO61Wj60D3qxvxtq7YdFL6N49Iwju+fv2229l//79FW7Ly8ur0bnQ86gB8NatW2u1v6ietLQ0iY6OrtBbWNlnRz9vvtrHxMRIamoqhz7Mz58vmqY0f/78IOwhAo3PX/AQACNov6h1CQYt/6O/FJYsWWLy2qzcOC0V88QTTwTlOe2muudPe5OOHDkia9askW7dupltq1evNtu0tF11HTx40PRKuX+hQeDFxsaasln62Rk6dKhru65feeWVfs/x+++/77Hto48+Mr32DRs25DSF+fnzRatH8FmLDHz+giiIA+yAasnJyXF+/fXXzqlTpzqbNGliftaloKDA1aZjx47Od955x7WuFSC06oNu27Bhg/P66693tmzZ0nn06FGOeh0bNGiQ8xe/+IWp/qBL586dnUOGDPFo437+9LxOnDjRuWLFCmd2drYZpd6zZ0/nz372M85fHViwYIGzYcOGztmzZ5sqHuPHj3cmJCQ4d+zYYW7XagI33nijq/327dud8fHxznvuuce01/vp/f/xj3/Uxe7iFM/fjBkzTKWBLVu2ODdu3Ghu1z/9b7/9Nsc2BPT3n/U3Ts/D9OnTzc/6d9DX+ePzFzwEwAg5LYmlvwi8Fw2MLLquZdLcS6E99NBDphxaXFyc88ILLzSBMOrewYMHnTfccIMzMTHRLPqzd8kl9/N3/Phx58CBA53p6enmD3nbtm3Ne2Dnzp2cvjry17/+1ZmZmemMjY11nnfeec7ly5e7btNz0adPH4/2n3zyibNLly6m/WmnneacOXMm5ypCzt8TTzzhbN++vbNRo0bO5ORkZ+/evZ0ffPBBiPYcVlk670XPm6/zp/j8BUeU/hPMHmYAAAAgnFAGDQAAALZCAAwAAABbIQAGAACArRAAAwAAwFYIgAEAAGArBMAAAACwFQJgAAAA2AoBMAAAAGyFABgAgigqKkoWLVrEMea4AAgjBMAAUEs333yzCXC9l0GDBkXkMT3ttNPkmWee8Xt7SUmJpKWlySOPPOLz9mnTppnbtR0AhDMCYAA4BRrs5ubmeixvvvlmvTymsbGxMnLkSJk3b544nc4Kt8+dO1duvPFG0w4AwhkBMACcgri4OGnRooXHkpyc7Lf9nj175NprrzVtUlNT5corr5QdO3Z49CpfddVV8thjj0lGRoY0a9ZMpk6dKqWlpfL73/9eUlJSpHXr1jJnzpxaPe6f//xnadmypWlz9913i8PhMLf37dtXcnJy5J577nH1ZPsyevRo2bZtm3z66ace2z/77DPZunWruf3LL7+UAQMGmN7gpk2bSp8+feSrr77ye0w++eQT83yHDx92bVu/fr3Z5v4aVqxYIRdeeKE0btxY2rRpI+PGjZNjx475fVwA8IcAGADqyPHjx6Vfv37SpEkTE0B+/vnn5mftRXZPG/j4449l7969ps306dNlypQpMmTIEBPcrl69Wu68806z7Nq1q0aPu2zZMhO86v+vvvqq6cnVRb3zzjsmsH744YddPdm+dO7cWX71q1+Z3l53GpB369ZNOnXqJAUFBTJq1CgTFK9atUo6dOggl156qdleWxs2bJBLLrlEhg0bJt9++60sXLjQvM7f/va3tX5MADbmBADUyqhRo5zR0dHOhIQEj+Xhhx92tdFfs++++675efbs2c6OHTs6y8vLXbcXFxc7Gzdu7Pzwww9dj5mZmeksKytztdH7XHDBBa710tJS8zxvvvlmjR9X72u5+uqrnddee61rXW+fMWNGla975syZ5vkLCgrMuv6v67NmzfLZXp8zMTHR+f777/s8LsuWLTPrhw4dct3+9ddfm23Z2dlm/cYbb3TefvvtHo/72WefORs0aOAsKiqqcp8BwF1MqANwAIhk2vM6c+ZMj22apuDLunXr5Mcff5TExESP7SdOnDA9s5ZzzjlHGjT43wU6TYXQnlVLdHS0SWE4cOBAjR9X72vRVAjtWa2p66+/XiZMmGB6YTXlQf/XmPa6664zt+t+/fGPfzQ92fv375eysjLTS71z506pLes1vvHGG65t+pzl5eWSnZ0tZ599dq0fG4D9EAADwClISEiQM844o1ptNVjr2rWrRxBnSU9Pd/3csGFDj9s0F9bXNn28U31c6zFqQvN6f/3rX5s0CA2A9X9dT0pKcuUb5+XlmYoSmZmZJk+6Z8+efqtDWMG++8A6KzfZovt5xx13mLxfb23btq3xawBgbwTAAFBHzjvvPNNb2rx5c1ewGE6Pq9UbtLe2OjTw1YFz//rXv+SLL74wg/Ysmvv7wgsvmLxfpbnKP/30k9/HsoJ0zTu2BhDqIDjv17hp06Zqf9kAgMowCA4ATkFxcbHs27fPY/EX7N1www2mMoJWaNAgUS/dL1++XH73u9/J7t27a70PgXpcrQOsg+i0okRlAavSyg4ajN50003mf63OYNH1119/Xb7//nszaE/3Tys3+KPttaqDDvbbsmWLfPDBB/L00097tPnDH/4gK1euNJUrNDjWihPvvfeejB07ttqvDwAsBMAAcAoWL15scmndl969e/tsGx8fbwJMvWSv1Qw0b/XWW2+VoqKiU+q5DdTjagUILTvWvn17j9QJf/Q5Dh06ZP73rgih27t06WLqAmvagvZO+6OpGVo7+YcffpBzzz1XnnjiiQqTbfziF78wQb0GvhdccIF57AcffNAcbwCoqSgdCVfjewEAAAARih5gAAAA2AoBMAAAAGyFABgAAAC2QgAMAAAAWyEABgAAgK0QAAMAAMBWCIABAABgKwTAAAAAsBUCYAAAANgKATAAAABshQAYAAAAYif/D5YKTIiBV++9AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------------------------------------------\n",
      "Matrix V Shape: torch.Size([65536])\n",
      "Number of elements: 65536\n",
      "Mean value: 0.0019\n",
      "Standard deviation: 0.2997\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn.init as init\n",
    "import math\n",
    "\n",
    "num_rotations = 1\n",
    "r = 16\n",
    "dim = 4096\n",
    "U = torch.empty(r, dim)\n",
    "V = torch.empty(r, dim)\n",
    "init.kaiming_normal_(U, a=math.sqrt(0.8), mode='fan_out')\n",
    "init.normal_(V, 0, 0.30)\n",
    "U = U.flatten()\n",
    "V = V.flatten()\n",
    "visualize_value_distribution(U,bins_count=100)\n",
    "\n",
    "analyze_value_distribution(U)\n",
    "visualize_value_distribution(V,bins_count=100)\n",
    "\n",
    "\n",
    "analyze_value_distribution(V)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8a3ccc6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import time\n",
    "\n",
    "def khatri_rao_col(A1, A2):\n",
    "    \"\"\"\n",
    "    Computes the Column-wise Khatri-Rao product of two matrices.\n",
    "    \n",
    "    Logic:\n",
    "        For each column k, compute the Kronecker product of column k of A1 \n",
    "        and column k of A2.\n",
    "        \n",
    "    Args:\n",
    "        A1 (torch.Tensor): Shape (r, d)\n",
    "        A2 (torch.Tensor): Shape (r, d)\n",
    "        \n",
    "    Returns:\n",
    "        torch.Tensor: Shape (r*r, d)\n",
    "    \"\"\"\n",
    "    # Check if dimensions match (number of columns must be equal)\n",
    "    assert A1.shape[1] == A2.shape[1], f\"Columns mismatch: {A1.shape[1]} vs {A2.shape[1]}\"\n",
    "    \n",
    "    d = A1.shape[1]\n",
    "    \n",
    "    # Einsum explanation:\n",
    "    # 'id': i-th row, d-th col of A1\n",
    "    # 'jd': j-th row, d-th col of A2\n",
    "    # '-> ijd': Compute outer product for each column d\n",
    "    # Result is a 3D tensor of shape (r, r, d)\n",
    "    out_tensor = torch.einsum('id,jd->ijd', A1, A2)\n",
    "    \n",
    "    # Flatten the first two dimensions to get (r*r, d)\n",
    "    return out_tensor.reshape(-1, d)\n",
    "\n",
    "def khatri_rao_row(A1, A2):\n",
    "    \"\"\"\n",
    "    Computes the Row-wise Khatri-Rao product (Face-splitting product).\n",
    "    \n",
    "    Logic:\n",
    "        For each row k, compute the Kronecker product of row k of A1 \n",
    "        and row k of A2.\n",
    "        \n",
    "    Args:\n",
    "        A1 (torch.Tensor): Shape (r, d)\n",
    "        A2 (torch.Tensor): Shape (r, d)\n",
    "        \n",
    "    Returns:\n",
    "        torch.Tensor: Shape (r, d*d)\n",
    "    \"\"\"\n",
    "    # Check if dimensions match (number of rows must be equal)\n",
    "    assert A1.shape[0] == A2.shape[0], f\"Rows mismatch: {A1.shape[0]} vs {A2.shape[0]}\"\n",
    "    \n",
    "    r = A1.shape[0]\n",
    "    \n",
    "    # Einsum explanation:\n",
    "    # 'ri': r-th row, i-th col of A1\n",
    "    # 'rj': r-th row, j-th col of A2\n",
    "    # '-> rij': Compute outer product for each row r\n",
    "    # Result is a 3D tensor of shape (r, d, d)\n",
    "    out_tensor = torch.einsum('ri,rj->rij', A1, A2)\n",
    "    \n",
    "    # Flatten the last two dimensions to get (r, d*d)\n",
    "    return out_tensor.reshape(r, -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6248ef2b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([5, 256]) torch.Size([256, 64]) torch.Size([64, 128])\n",
      "y0 torch.Size([5, 128])\n",
      "c torch.Size([256, 128])\n",
      "A2, X torch.Size([256, 32]) torch.Size([5, 256, 1])\n",
      "torch.Size([5, 32, 32])\n",
      "h shape torch.Size([128, 32, 1]) torch.Size([5, 128, 32, 32]) torch.Size([128, 1, 32])\n",
      "y1 torch.Size([5, 128])\n",
      "sum tensor(-11390.5176, device='cuda:0')\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "N = 128\n",
    "r = 32\n",
    "M = 256\n",
    "device = 'cuda'\n",
    "B1, B2 = torch.randn(r,N, device=device), torch.randn(r,N, device=device)\n",
    "A1, A2 = torch.randn(M,r, device=device), torch.randn(M,r, device=device)\n",
    "\n",
    "x = torch.randn(5,M,device=device)\n",
    "\n",
    "B = torch.cat((B1,B2),dim=0)\n",
    "A = torch.cat((A1,A2),dim=1)\n",
    "print(x.shape, A.shape, B.shape)\n",
    "y0 = (x @ A ) @ B\n",
    "print('y0', y0.shape)\n",
    "\n",
    "C = (A1 @ B1) * (A2 @ B2)\n",
    "print('c', C.shape)\n",
    "\n",
    "X = x.unsqueeze(-1)\n",
    "print('A2, X', A2.shape, X.shape)\n",
    "h = A1.T @ (A2 * X) # @ B1.T\n",
    "print(h.shape)\n",
    "h = h.unsqueeze(-3).expand(-1,N,-1,-1)\n",
    "b1 = B1.T.unsqueeze(-2)\n",
    "b2 = B2.T.unsqueeze(-1)\n",
    "print('h shape', b2.shape, h.shape, b1.shape)\n",
    "y1 = b1 @ h @ b2\n",
    "y1 = y1.squeeze()\n",
    "print('y1', y1.shape)\n",
    "\n",
    "delta = y0-y1\n",
    "sumd = delta.sum()\n",
    "print('sum', sumd)\n",
    "\n",
    "## ===\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "ab925f89",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([2, 16]) torch.Size([16, 6]) torch.Size([6, 8])\n",
      "y0 torch.Size([2, 8])\n",
      "c torch.Size([16, 8])\n",
      "ka torch.Size([16, 9])\n",
      "ha torch.Size([2, 9])\n",
      "hb tensor([1608, 1059, 3814,  702, 1860,  660, 2159, 2795])\n",
      "A2, X torch.Size([16, 3]) torch.Size([2, 16, 1])\n",
      "h torch.Size([2, 3, 3])\n",
      "h shape torch.Size([8, 1, 3]) torch.Size([2, 8, 3, 3]) torch.Size([8, 3, 1])\n",
      "y1 torch.Size([2, 8]) tensor([[1608, 1059, 3814,  702, 1860,  660, 2159, 2795],\n",
      "        [3584, 1926, 7169, 1498, 3849, 1312, 3516, 5275]])\n",
      "sum tensor(0)\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "N = 8\n",
    "r = 3\n",
    "M = 16\n",
    "device = 'cpu'\n",
    "T = 5\n",
    "# B1, B2 = torch.randn(r,N, device=device), torch.randn(r,N, device=device)\n",
    "# A1, A2 = torch.randn(M,r, device=device), torch.randn(M,r, device=device)\n",
    "\n",
    "# x = torch.randn(2,M,device=device)\n",
    "\n",
    "B1, B2 = torch.randint(T, (r,N), device=device), torch.randint(T, (r,N), device=device)\n",
    "A1, A2 = torch.randint(T, (M,r), device=device), torch.randint(T, (M,r), device=device)\n",
    "\n",
    "x = torch.randint(T, (2,M),device=device)\n",
    "\n",
    "B = torch.cat((B1,B2),dim=0)\n",
    "A = torch.cat((A1,A2),dim=1)\n",
    "print(x.shape, A.shape, B.shape)\n",
    "y00 = (x @ A ) @ B\n",
    "print('y0', y0.shape)\n",
    "\n",
    "y0 = x @ ((A1 @ B1) * (A2 @ B2))\n",
    "\n",
    "C = (A1 @ B1) * (A2 @ B2)\n",
    "print('c', C.shape)\n",
    "##\n",
    "ka = khatri_rao_row(A1,A2)\n",
    "print('ka', ka.shape)\n",
    "ha = x @ ka\n",
    "print('ha', ha.shape)\n",
    "kb = khatri_rao_col(B1,B2)\n",
    "hb = ha @ kb\n",
    "print('hb', hb[0])\n",
    "##\n",
    "X = x.unsqueeze(-1)\n",
    "print('A2, X', A2.shape, X.shape)\n",
    "h = A1.T @ (A2 * X) # @ A1.T\n",
    "print('h', h.shape)\n",
    "# print(f'ha h, {ha[0]}, \\n {h[0]}')\n",
    "h = h.unsqueeze(-3).expand(-1,N,-1,-1)\n",
    "b1 = B1.T.unsqueeze(-2)\n",
    "b2 = B2.T.unsqueeze(-1)\n",
    "print('h shape', b1.shape, h.shape, b2.shape)\n",
    "y1 = b1 @ h @ b2\n",
    "y1 = y1.squeeze()\n",
    "print('y1', y1.shape, y1)\n",
    "\n",
    "delta = y0-y1\n",
    "sumd = delta.sum()\n",
    "print('sum', sumd)\n",
    "\n",
    "## ===\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "184cd72d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time1 0.620816707611084\n",
      "time2 0.25490880012512207\n",
      "time3 5.265974044799805\n",
      "time4 0.9792120456695557\n",
      "time5 0.10942697525024414\n",
      "time6 3.903172731399536\n",
      "1765263288.788364\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 <a href='https://aka.ms/vscodeJupyterKernelCrash'>here</a> for more info. \n",
      "\u001b[1;31mView Jupyter <a href='command:jupyter.viewOutput'>log</a> for further details."
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import time\n",
    "import sys\n",
    "\n",
    "N = 4096*2\n",
    "r = 64*4\n",
    "M = 4096\n",
    "device = 'cuda'\n",
    "T = 5\n",
    "l = 2000\n",
    "B1, B2 = torch.randn(r,N, device=device), torch.randn(r,N, device=device)\n",
    "A1, A2 = torch.randn(M,r, device=device), torch.randn(M,r, device=device)\n",
    "\n",
    "B = torch.cat((B1,B2),dim=0)\n",
    "A = torch.cat((A1,A2),dim=1)\n",
    "\n",
    "def optimized_einsum_method(x, A1, B1, A2, B2):\n",
    "    # This directly computes y = x @ C where C = (A1@B1) * (A2@B2)\n",
    "    \n",
    "    # 1. Compute the two components of C: \n",
    "    # T1_mn = A1_mk * B1_kn -> (M, N)\n",
    "    T1 = torch.einsum('mk,kn->mn', A1, B1)\n",
    "    # T2_mn = A2_ml * B2_ln -> (M, N)\n",
    "    T2 = torch.einsum('ml,ln->mn', A2, B2)\n",
    "    \n",
    "    # 2. Hadamard Product: C_mn = T1_mn * T2_mn\n",
    "    C = T1 * T2\n",
    "    \n",
    "    # 3. Final Matrix Multiplication: y_in = x_im * C_mn\n",
    "    y = torch.einsum('im,mn->in', x, C)\n",
    "    \n",
    "    return y\n",
    "\n",
    "x = torch.randn(20,M,device=device)\n",
    "start1 = time.time()\n",
    "for _ in range(l):\n",
    "    y0 = x @ ((A1 @ B1) * (A2 @ B2))\n",
    "print('time1', time.time()-start1)\n",
    "\n",
    "start2 = time.time()\n",
    "X = x.unsqueeze(-1)\n",
    "for _ in range(l):\n",
    "    h = A1.T @ (A2 * X)\n",
    "    # h = h.unsqueeze(-3).expand(-1,N,-1,-1)\n",
    "    # b1 = B1.T.unsqueeze(-2)\n",
    "    # b2 = B2.T.unsqueeze(-1)\n",
    "    # y1 = b1 @ h @ b2\n",
    "    # y1 = y1.squeeze()\n",
    "print('time2', time.time()-start2)\n",
    "\n",
    "start3 = time.time()\n",
    "for _ in range(l):\n",
    "    ka = khatri_rao_row(A1,A2)\n",
    "    ha = x @ ka\n",
    "    kb = khatri_rao_col(B1,B2)\n",
    "    hb = ha @ kb\n",
    "print('time3', time.time()-start3)\n",
    "\n",
    "start4 = time.time()\n",
    "for _ in range(l):\n",
    "    y = optimized_einsum_method(x, A1, B1, A2, B2)\n",
    "print('time4', time.time()-start4)\n",
    "\n",
    "start5 = time.time()\n",
    "for _ in range(l):\n",
    "    y = (x @ A) @ B\n",
    "print('time5', time.time()-start5)\n",
    "\n",
    "\n",
    "start6 = time.time()\n",
    "for _ in range(l):\n",
    "    h = torch.einsum('...d, dr, dq -> ...rq', x, A1, A2)\n",
    "    y = torch.einsum('kn, ln, ...kl -> ...n', B1, B2, h)\n",
    "print('time6', time.time()-start6)\n",
    "\n",
    "print(time.time())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b2303f31",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time6 0.20953679084777832\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "# Assume input tensors:\n",
    "# N, r = 1024, 4\n",
    "A = torch.randn(N, r, device='cuda')\n",
    "B = torch.randn(N, r, device='cuda')\n",
    "# h (vector r^2) is reshaped to H (matrix r, r)\n",
    "H_flat = torch.randn(r*r, device='cuda')\n",
    "H = H_flat.reshape(r, r) # H: (r, r)\n",
    "\n",
    "# --- Optimized Method using Einsum ---\n",
    "# 'nk, nl, kl -> n'\n",
    "# Sum over k and l (the two r dimensions), resulting in vector y size N\n",
    "start6 = time.time()\n",
    "for _ in range(l):\n",
    "    H = \n",
    "    y_optimized = torch.einsum('nk, nl, kl -> n', A, B, H)\n",
    "print('time6', time.time()-start6)\n",
    "\n",
    "\n",
    "# y_optimized has shape (N,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "85a802ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of DebertaV2ForSequenceClassification were not initialized from the model checkpoint at microsoft/deberta-v3-base and are newly initialized: ['classifier.bias', 'classifier.weight', 'pooler.dense.bias', 'pooler.dense.weight']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 832,516 || all params: 185,174,792 || trainable%: 0.4496\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.query_proj.base_layer.bias, torch.Size([768])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.query_proj.boft_R.default, torch.Size([1, 96, 8, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.query_proj.boft_s.default, torch.Size([768, 1])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Loading checkpoint shards: 100%|██████████| 2/2 [00:04<00:00,  2.16s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 4,456,448 || all params: 6,742,872,064 || trainable%: 0.0661\n"
     ]
    }
   ],
   "source": [
    "import transformers\n",
    "\n",
    "from transformers import AutoModelForSeq2SeqLM\n",
    "from peft import BOFTConfig, get_peft_model, BOFTConfig\n",
    "\n",
    "config = BOFTConfig(\n",
    "    boft_block_size=8,\n",
    "    boft_n_butterfly_factor=20,\n",
    "    target_modules=[\"query_proj\", \"value_proj\", \"key_proj\", 'attention.output.dense', 'intermediate.dense', 'output.dense'],\n",
    "    boft_dropout=0.1,\n",
    "    bias=\"boft_only\",\n",
    "    modules_to_save=[\"classifier\"],\n",
    ")\n",
    "\n",
    "model = transformers.AutoModelForSequenceClassification.from_pretrained(\n",
    "    \"microsoft/deberta-v3-base\",\n",
    "    num_labels=4,\n",
    ")\n",
    "boft_model = get_peft_model(model, config)\n",
    "boft_model.print_trainable_parameters()\n",
    "# print(boft_model)\n",
    "\n",
    "for n, p in boft_model.named_parameters():\n",
    "    if p.requires_grad and 'layer.2' in n and 'query_' in n:\n",
    "        print(f'n = {n}, {p.shape}')\n",
    "\n",
    "config2 = BOFTConfig(\n",
    "    boft_block_size=16,\n",
    "    boft_n_butterfly_factor=4,\n",
    "    # target_modules=[\"q_proj\", \"v_proj\", \"v_proj\", \"o_proj\", \"gate_proj\",\"up_proj\",\"down_proj\"],\n",
    "    target_modules=[\"q_proj\", \"v_proj\"],\n",
    "    boft_dropout=0.1,\n",
    "    bias=\"boft_only\",\n",
    "    modules_to_save=[\"classifier\"],\n",
    ")\n",
    "\n",
    "model2 = transformers.AutoModelForCausalLM.from_pretrained(\n",
    "    \"meta-llama/Llama-2-7b-hf\",\n",
    ")\n",
    "boft_model2 = get_peft_model(model2, config2)\n",
    "boft_model2.print_trainable_parameters()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "f93152c4",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/transformers/convert_slow_tokenizer.py:566: UserWarning: The sentencepiece tokenizer that you are converting to a fast tokenizer uses the byte fallback option which is not implemented in the fast tokenizers. In practice this means that the fast version of the tokenizer can produce unknown tokens whereas the sentencepiece version would have converted these unknown tokens into a sequence of byte tokens matching the original piece of text.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 629,378 || all params: 186,380,164 || trainable%: 0.3377\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/peft/mapping_func.py:72: UserWarning: You are trying to modify a model with PEFT for a second time. If you want to reload the model with a different config, make sure to call `.unload()` before.\n",
      "  warnings.warn(\n",
      "/home/work/miniconda3/envs/allm/lib/python3.11/site-packages/peft/tuners/tuners_utils.py:282: UserWarning: Already found a `peft_config` attribute in the model. This will lead to having multiple adapters in the model. Make sure to know what you are doing!\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig\n",
    "import transformers\n",
    "\n",
    "from peft import OFTConfig\n",
    "\n",
    "\n",
    "model_name = \"microsoft/deberta-v3-base\"\n",
    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
    "\n",
    "# Configure OFT\n",
    "config = OFTConfig(\n",
    "    oft_block_size=16,\n",
    "    use_cayley_neumann=True,\n",
    "    target_modules=\"all-linear\",\n",
    "    bias=\"none\",\n",
    "    task_type=\"SEQ_CLS\",\n",
    ")\n",
    "\n",
    "# model = transformers.AutoModelForSequenceClassification.from_pretrained(\n",
    "#     \"microsoft/deberta-v3-base\",\n",
    "#     num_labels=4,\n",
    "# )\n",
    "model = get_peft_model(model, config)\n",
    "model.print_trainable_parameters()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "cd42ba2f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of DebertaV2ForSequenceClassification were not initialized from the model checkpoint at microsoft/deberta-v3-base and are newly initialized: ['classifier.bias', 'classifier.weight', 'pooler.dense.bias', 'pooler.dense.weight']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 1,328,642 || all params: 185,752,324 || trainable%: 0.7153\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.query_proj.lora_A.default.weight, torch.Size([8, 768])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.query_proj.lora_B.default.weight, torch.Size([768, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.key_proj.lora_A.default.weight, torch.Size([8, 768])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.key_proj.lora_B.default.weight, torch.Size([768, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.value_proj.lora_A.default.weight, torch.Size([8, 768])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.self.value_proj.lora_B.default.weight, torch.Size([768, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.output.dense.lora_A.default.weight, torch.Size([8, 768])\n",
      "n = base_model.model.deberta.encoder.layer.2.attention.output.dense.lora_B.default.weight, torch.Size([768, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.intermediate.dense.lora_A.default.weight, torch.Size([8, 768])\n",
      "n = base_model.model.deberta.encoder.layer.2.intermediate.dense.lora_B.default.weight, torch.Size([3072, 8])\n",
      "n = base_model.model.deberta.encoder.layer.2.output.dense.lora_A.default.weight, torch.Size([8, 3072])\n",
      "n = base_model.model.deberta.encoder.layer.2.output.dense.lora_B.default.weight, torch.Size([768, 8])\n"
     ]
    }
   ],
   "source": [
    "from peft import LoraConfig, get_peft_model\n",
    "import transformers\n",
    "\n",
    "model = transformers.AutoModelForSequenceClassification.from_pretrained(\n",
    "    \"microsoft/deberta-v3-base\",\n",
    "    num_labels=2,\n",
    ")\n",
    "peft_config = LoraConfig(\n",
    "                r=8,\n",
    "                lora_alpha=32,\n",
    "                lora_dropout=0.05,\n",
    "                bias=\"none\",\n",
    "                task_type=\"SEQ_CLS\",\n",
    "                target_modules=[\"query_proj\", \"value_proj\", \"key_proj\", 'attention.output.dense', 'intermediate.dense', 'output.dense']\n",
    "                )\n",
    "\n",
    "model = get_peft_model(model, peft_config)\n",
    "model.print_trainable_parameters()\n",
    "\n",
    "for n, p in model.named_parameters():\n",
    "    if p.requires_grad and 'layer.2' in n:\n",
    "        print(f'n = {n}, {p.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d2f8a801",
   "metadata": {},
   "outputs": [],
   "source": [
    "rotation_layers = filter(\n",
    "            lambda p: p.requires_grad, model.parameters()\n",
    "        )    \n",
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained(\n",
    "    mainCfg.model.model_name,\n",
    "    model_max_length=mainCfg.model.model_max_seq_length,\n",
    "    padding_side=\"right\",\n",
    "    use_fast=True,\n",
    ")\n",
    "\n",
    "if tokenizer.pad_token is None:\n",
    "    if tokenizer.unk_token_id is not None:\n",
    "        tokenizer.pad_token_id = tokenizer.unk_token_id\n",
    "        tokenizer.pad_token = tokenizer.unk_token\n",
    "        print(\"Set PAD token to UNK token.\")\n",
    "    elif tokenizer.eos_token_id is not None:\n",
    "        tokenizer.pad_token_id = tokenizer.eos_token_id\n",
    "        tokenizer.pad_token = tokenizer.eos_token\n",
    "        print(\"Set PAD token to EOS token.\")\n",
    "\n",
    "    if model is not None:\n",
    "        model.config.pad_token_id = tokenizer.pad_token_id\n",
    "        if model.config.pad_token_id != tokenizer.pad_token_id:\n",
    "            raise ValueError(\"Failed to sync pad_token_id between tokenizer and model config\")\n",
    "\n",
    "# local MetaMathQA-40K\n",
    "raw_datasets = load_dataset(\"json\", data_files=mainCfg.data.path, split=mainCfg.data.dataset_split)\n",
    "#raw_train_datasets = load_dataset(\"MetaMathQA-40K\", split=mainCfg.data.dataset_split)\n",
    "# print('raw', type(raw_train_datasets), len(raw_train_datasets))\n",
    "\n",
    "# split a single set\n",
    "# split_ratio = mainCfg.data.split_ratio\n",
    "# split_data = raw_datasets.train_test_split(test_size=split_ratio, seed=42)\n",
    "# raw_train_datasets = split_data['train']\n",
    "# raw_valid_datasets = split_data['test']\n",
    "\n",
    "train_dataset = raw_datasets.map(\n",
    "    train_tokenize_function,\n",
    "    batched=True,\n",
    "    batch_size=30000,\n",
    "    num_proc=32,\n",
    "    remove_columns=raw_datasets.column_names,\n",
    "    load_from_cache_file=True,\n",
    "    desc=\"Running tokenizer on train dataset\",\n",
    "    fn_kwargs={\"tokenizer\": tokenizer, \"query\": mainCfg.data.dataset_field[0],\n",
    "                \"response\": mainCfg.data.dataset_field[1]}\n",
    ")\n",
    "\n",
    "print('- dataset size: ', len(train_dataset))\n",
    "\n",
    "\n",
    "data_collator = DataCollatorForSupervisedDataset(tokenizer=tokenizer, max_length=mainCfg.model.model_max_seq_length, \n",
    "                                                    #mode=mainCfg.model.data_collator_mode,\n",
    "                                                    )\n",
    "data_module = dict(train_dataset=train_dataset, data_collator=data_collator)\n",
    "\n",
    "optimizer = optim.AdamW(\n",
    "    rotation_layers, \n",
    "    lr=mainCfg.trainer_args.learning_rate, #\n",
    "    eps=1e-8\n",
    ")\n",
    "# print('model x', model)\n",
    "start_time = datetime.now()\n",
    "print('start time: ', start_time.strftime(\"%Y-%m-%d %H:%M:%S\"))\n",
    "trainer = MyTrainer(model=model, processing_class=tokenizer,\n",
    "                    lamda=mainCfg.model.lambda_reg,\n",
    "                    optimizers=(optimizer, None),\n",
    "                    args=training_args, **data_module)\n",
    "model.config.use_cache = False"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "allm",
   "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}