diff --git "a/02_stage1_training.ipynb" "b/02_stage1_training.ipynb" new file mode 100644--- /dev/null +++ "b/02_stage1_training.ipynb" @@ -0,0 +1,11029 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Notebook 2: Model Architecture & Stage 1 Training\n", + "\n", + "This notebook handles:\n", + "1. Setting up the Swin Transformer V2 Tiny model with a regression head\n", + "2. Creating the PyTorch Dataset and DataLoader for spectrum images\n", + "3. **Stage 1**: Pre-training on the pure spectrum dataset (15,392 samples)\n", + "4. Saving the Stage 1 checkpoint for transfer learning in Notebook 3\n", + "\n", + "**Prerequisites**: Run Notebook 1 first to generate the data.\n", + "\n", + "---" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.1 Imports & Configuration" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Using device: cuda\n", + "GPU: NVIDIA GeForce RTX 4060\n" + ] + } + ], + "source": [ + "import torch\n", + "import torch.nn as nn\n", + "from torch.utils.data import Dataset, DataLoader, random_split\n", + "import torchvision.transforms as T\n", + "import torchvision.models as models\n", + "from prodigyopt import Prodigy\n", + "import numpy as np\n", + "import json\n", + "import os\n", + "from PIL import Image\n", + "from pathlib import Path\n", + "from tqdm import tqdm\n", + "import matplotlib.pyplot as plt\n", + "import time\n", + "import cv2\n", + "\n", + "# Device setup\n", + "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n", + "print(f\"Using device: {device}\")\n", + "if device.type == 'cuda':\n", + " print(f\"GPU: {torch.cuda.get_device_name(0)}\")\n", + " # print(f\"Memory: {torch.cuda.get_device_properties(0).total_mem / 1e9:.1f} GB\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.2 Define the Spectrum Dataset Class\n", + "\n", + "This custom PyTorch Dataset loads spectrum images from disk along with their\n", + "corresponding labels (EC, EL, EJ). Labels are normalized to [0, 1] using the\n", + "known parameter ranges so that MSE loss treats all three parameters equally." + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Dataset class defined.\n" + ] + } + ], + "source": [ + "# Parameter ranges for normalization\n", + "PARAM_RANGES = {\n", + " 'EC': (0.5, 3.0),\n", + " 'EL': (0.1, 2.0),\n", + " 'EJ': (2.0, 10.0)\n", + "}\n", + "\n", + "def normalize_labels(ec, el, ej):\n", + " \"\"\"Normalize parameters to [0, 1] range.\"\"\"\n", + " ec_n = (ec - PARAM_RANGES['EC'][0]) / (PARAM_RANGES['EC'][1] - PARAM_RANGES['EC'][0])\n", + " el_n = (el - PARAM_RANGES['EL'][0]) / (PARAM_RANGES['EL'][1] - PARAM_RANGES['EL'][0])\n", + " ej_n = (ej - PARAM_RANGES['EJ'][0]) / (PARAM_RANGES['EJ'][1] - PARAM_RANGES['EJ'][0])\n", + " return np.array([ec_n, el_n, ej_n], dtype=np.float32)\n", + "\n", + "def denormalize_labels(pred):\n", + " \"\"\"Convert normalized predictions back to physical units (GHz).\"\"\"\n", + " ec = pred[0] * (PARAM_RANGES['EC'][1] - PARAM_RANGES['EC'][0]) + PARAM_RANGES['EC'][0]\n", + " el = pred[1] * (PARAM_RANGES['EL'][1] - PARAM_RANGES['EL'][0]) + PARAM_RANGES['EL'][0]\n", + " ej = pred[2] * (PARAM_RANGES['EJ'][1] - PARAM_RANGES['EJ'][0]) + PARAM_RANGES['EJ'][0]\n", + " return np.array([ec, el, ej])\n", + "\n", + "\n", + "class FluxoniumSpectrumDataset(Dataset):\n", + " \"\"\"\n", + " PyTorch Dataset for fluxonium spectrum images.\n", + " \n", + " Loads grayscale PNGs, converts to 3-channel RGB (for ImageNet-pretrained model),\n", + " applies standard ImageNet normalization, and returns normalized labels.\n", + " \"\"\"\n", + " def __init__(self, data_dir, transform=None, augment=False):\n", + " self.data_dir = Path(data_dir)\n", + " self.img_dir = self.data_dir / 'images'\n", + " \n", + " with open(self.data_dir / 'labels.json') as f:\n", + " self.labels = json.load(f)\n", + " \n", + " # Only include samples that have both image and label\n", + " self.indices = sorted(self.labels.keys())\n", + " self.indices = [idx for idx in self.indices \n", + " if (self.img_dir / f'{int(idx):05d}.png').exists()]\n", + " \n", + " self.transform = transform\n", + " self.augment = augment\n", + " \n", + " # Default transform: convert to tensor + ImageNet normalization\n", + " if self.transform is None:\n", + " transforms_list = []\n", + " if augment:\n", + " transforms_list.extend([\n", + " T.RandomHorizontalFlip(p=0.5),\n", + " T.RandomRotation(degrees=3),\n", + " ])\n", + " transforms_list.extend([\n", + " T.Resize((256, 256)),\n", + " T.ToTensor(),\n", + " T.Normalize(mean=[0.485, 0.456, 0.406],\n", + " std=[0.229, 0.224, 0.225])\n", + " ])\n", + " self.transform = T.Compose(transforms_list)\n", + " \n", + " def __len__(self):\n", + " return len(self.indices)\n", + " \n", + " def __getitem__(self, i):\n", + " idx = self.indices[i]\n", + " img_path = self.img_dir / f'{int(idx):05d}.png'\n", + " \n", + " # # Load grayscale and convert to RGB (3 channels)\n", + " # img = Image.open(img_path).convert('RGB')\n", + " # img = self.transform(img)\n", + "\n", + " # OpenCV version - 3-5x faster image loading\n", + " img = cv2.imread(str(img_path), cv2.IMREAD_GRAYSCALE)\n", + " img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)\n", + " img = cv2.resize(img, (256, 256), interpolation=cv2.INTER_LANCZOS4)\n", + " img = img.astype(np.float32) / 255.0\n", + " img = (img - np.array([0.485, 0.456, 0.406])) / np.array([0.229, 0.224, 0.225])\n", + " img = torch.from_numpy(img.transpose(2, 0, 1))\n", + " \n", + " label_dict = self.labels[idx]\n", + " label = normalize_labels(label_dict['EC'], label_dict['EL'], label_dict['EJ'])\n", + " \n", + " return img, torch.tensor(label, dtype=torch.float32)\n", + "\n", + "\n", + "print(\"Dataset class defined.\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.3 Build the Swin Transformer V2 Model\n", + "\n", + "We load a Swin Transformer V2 Tiny pretrained on ImageNet and replace the\n", + "classification head (1000 classes) with a regression head (3 outputs: EC, EL, EJ).\n", + "\n", + "The model expects 256ร—256 input with window size 8." + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Total parameters: 27,584,877\n", + "Trainable parameters: 2,307\n" + ] + } + ], + "source": [ + "def build_model():\n", + " \"\"\"\n", + " Build Swin Transformer V2 Tiny with a 3-output regression head.\n", + " \n", + " Architecture:\n", + " - Backbone: Swin V2 Tiny (pretrained on ImageNet-1K)\n", + " - Head: Linear(768, 3) for predicting [EC, EL, EJ]\n", + " \"\"\"\n", + " model = models.swin_v2_t(weights=models.Swin_V2_T_Weights.IMAGENET1K_V1)\n", + " \n", + " # Replace classification head with regression head\n", + " in_features = model.head.in_features # 768 for Swin V2 Tiny\n", + " model.head = nn.Sequential(\n", + " nn.Linear(in_features, 3),\n", + " nn.Sigmoid()\n", + " )\n", + " \n", + " # Initialize the new head with small weights\n", + " nn.init.xavier_uniform_(model.head[0].weight)\n", + " nn.init.zeros_(model.head[0].bias)\n", + " \n", + " return model\n", + "\n", + "\n", + "model = build_model().to(device)\n", + "\n", + "# Freeze all pretrained layers\n", + "for param in model.parameters():\n", + " param.requires_grad = False\n", + "\n", + "# Train only the regression head\n", + "for param in model.head.parameters():\n", + " param.requires_grad = True\n", + "\n", + "# Count parameters\n", + "total_params = sum(p.numel() for p in model.parameters())\n", + "trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)\n", + "print(f\"Total parameters: {total_params:,}\")\n", + "print(f\"Trainable parameters: {trainable_params:,}\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.4 Prepare DataLoaders\n", + "\n", + "Split the pure spectrum dataset into 90% training and 10% validation.\n", + "Training data gets augmentation (horizontal flips, small rotations);\n", + "validation data gets no augmentation." + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "id": "25879a32", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "๐Ÿ” INSPECTING DATASET...\n", + "โœ… Dataset size: 15329 samples\n", + "๐Ÿ“Š Train samples: 13796 | Val samples: 1533\n", + "โœ… FIXED LOADERS:\n", + " Train batches: 1725\n", + " Val batches: 192\n" + ] + } + ], + "source": [ + "import os, json\n", + "from PIL import Image\n", + "import torch\n", + "from torch.utils.data import Dataset, DataLoader, random_split\n", + "import torchvision.transforms as T\n", + "\n", + "class SpectrumDataset(Dataset):\n", + " def __init__(self, image_dir, label_path, transform=None):\n", + " self.image_dir = image_dir\n", + " self.transform = transform\n", + "\n", + " with open(label_path, \"r\") as f:\n", + " self.labels = json.load(f)\n", + "\n", + " self.keys = sorted(self.labels.keys(), key=lambda x: int(x))\n", + "\n", + " def __len__(self):\n", + " return len(self.keys)\n", + "\n", + " def __getitem__(self, idx):\n", + " key = self.keys[idx]\n", + " img_path = os.path.join(self.image_dir, f\"{int(key):05d}.png\")\n", + "\n", + " img = Image.open(img_path).convert(\"RGB\")\n", + " if self.transform is not None:\n", + " img = self.transform(img)\n", + "\n", + " y = self.labels[key]\n", + "\n", + " ec_n, el_n, ej_n = normalize_labels(\n", + " y[\"EC\"],\n", + " y[\"EL\"],\n", + " y[\"EJ\"]\n", + " )\n", + "\n", + " label = torch.tensor(\n", + " [ec_n, el_n, ej_n],\n", + " dtype=torch.float32\n", + " )\n", + "\n", + " return img, label\n", + "\n", + "transform = T.Compose([\n", + " T.Resize((224, 224)),\n", + " T.ToTensor(),\n", + "])\n", + "\n", + "dataset = SpectrumDataset(\n", + " image_dir=\"data/pure_spectrum/images\",\n", + " label_path=\"data/pure_spectrum/labels.json\",\n", + " transform=transform\n", + ")\n", + "\n", + "train_size = int(0.9 * len(dataset))\n", + "val_size = len(dataset) - train_size\n", + "train_ds, val_ds = random_split(dataset, [train_size, val_size])\n", + "\n", + "train_loader = DataLoader(train_ds, batch_size=8, shuffle=True, num_workers=0)\n", + "val_loader = DataLoader(val_ds, batch_size=8, shuffle=False, num_workers=0)\n", + "\n", + "print(\"๐Ÿ” INSPECTING DATASET...\")\n", + "print(f\"โœ… Dataset size: {len(dataset)} samples\")\n", + "print(f\"๐Ÿ“Š Train samples: {len(train_ds)} | Val samples: {len(val_ds)}\")\n", + "print(\"โœ… FIXED LOADERS:\")\n", + "print(f\" Train batches: {len(train_loader)}\")\n", + "print(f\" Val batches: {len(val_loader)}\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.5 Training Utilities\n", + "\n", + "Define the loss function (MSE), the paper's custom accuracy metric, and\n", + "a training/validation loop with logging." + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Training utilities defined.\n" + ] + } + ], + "source": [ + "# Loss function: Mean Squared Error\n", + "criterion = nn.SmoothL1Loss(beta=0.05)\n", + "\n", + "\n", + "def compute_accuracy(preds, targets):\n", + " \"\"\"\n", + " Compute the paper's custom accuracy metric.\n", + " \n", + " Acc(E_v) = (1/N) * sum(1 - |pred - true| / range)\n", + " \n", + " Since labels are normalized to [0,1], the range is just 1.0,\n", + " so: Acc = 1 - mean(|pred - true|)\n", + " \n", + " Returns per-parameter and overall accuracy.\n", + " \"\"\"\n", + " errors = torch.abs(preds - targets) # shape: (N, 3)\n", + " per_param_acc = (1.0 - errors.mean(dim=0)) * 100 # percentage\n", + " overall_acc = per_param_acc.mean()\n", + " return {\n", + " 'EC_acc': per_param_acc[0].item(),\n", + " 'EL_acc': per_param_acc[1].item(),\n", + " 'EJ_acc': per_param_acc[2].item(),\n", + " 'overall_acc': overall_acc.item()\n", + " }\n", + "\n", + "\n", + "def train_one_epoch(model, loader, optimizer, scheduler, device):\n", + " \"\"\"Train for one epoch. Returns average loss.\"\"\"\n", + " model.train()\n", + " total_loss = 0.0\n", + " n_batches = 0\n", + " \n", + " for imgs, labels in loader:\n", + " imgs, labels = imgs.to(device), labels.to(device)\n", + " \n", + " optimizer.zero_grad()\n", + " preds = model(imgs)\n", + " loss = criterion(preds, labels)\n", + " loss.backward()\n", + " optimizer.step()\n", + " if scheduler is not None:\n", + " scheduler.step()\n", + " \n", + " total_loss += loss.item()\n", + " n_batches += 1\n", + " \n", + " return total_loss / n_batches\n", + "\n", + "\n", + "@torch.no_grad()\n", + "def validate(model, loader, device):\n", + " \"\"\"Evaluate on validation set. Returns loss and accuracy metrics.\"\"\"\n", + " model.eval()\n", + " total_loss = 0.0\n", + " all_preds = []\n", + " all_targets = []\n", + " \n", + " for imgs, labels in loader:\n", + " imgs, labels = imgs.to(device), labels.to(device)\n", + " preds = model(imgs)\n", + " loss = criterion(preds, labels)\n", + " total_loss += loss.item()\n", + " all_preds.append(preds.cpu())\n", + " all_targets.append(labels.cpu())\n", + " \n", + " all_preds = torch.cat(all_preds)\n", + " all_targets = torch.cat(all_targets)\n", + " avg_loss = total_loss / len(loader)\n", + " acc = compute_accuracy(all_preds, all_targets)\n", + " \n", + " return avg_loss, acc\n", + "\n", + "\n", + "print(\"Training utilities defined.\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.6 Stage 1: Train on Pure Spectrum Dataset\n", + "\n", + "This is the first stage of the two-stage transfer learning approach:\n", + "- **Optimizer**: Prodigy (parameter-free, lr=1.0)\n", + "- **Scheduler**: Cosine annealing\n", + "- **Data**: 15,392 pure spectrum images\n", + "- **Epochs**: Train until validation loss plateaus (~100-200 epochs)\n", + "\n", + "Prodigy automatically adapts the learning rate, so no manual lr tuning is needed." + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "id": "192c7446", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "โœ… Augmented training subset created:\n", + " Full train dataset: 15329 samples\n", + " train_aug_subset: 13796 samples\n" + ] + } + ], + "source": [ + "# --- Create augmented training subset ---\n", + "from torch.utils.data import Subset\n", + "\n", + "# Define augmentation transform (stronger than basic validation transform)\n", + "train_transform = T.Compose([\n", + " T.Resize((256, 256)),\n", + " T.RandomHorizontalFlip(p=0.5),\n", + " T.RandomRotation(degrees=5),\n", + " T.ColorJitter(brightness=0.1, contrast=0.1),\n", + " T.ToTensor(),\n", + "])\n", + "\n", + "# Create full augmented training dataset\n", + "train_dataset_full = SpectrumDataset(\n", + " image_dir=\"data/pure_spectrum/images\",\n", + " label_path=\"data/pure_spectrum/labels.json\",\n", + " transform=train_transform # โ† Augmented!\n", + ")\n", + "\n", + "# Take 90% for training (subset to avoid memory issues during dev)\n", + "train_size = int(0.9 * len(train_dataset_full))\n", + "train_aug_subset = Subset(train_dataset_full, range(train_size))\n", + "\n", + "print(f\"โœ… Augmented training subset created:\")\n", + "print(f\" Full train dataset: {len(train_dataset_full)} samples\")\n", + "print(f\" train_aug_subset: {len(train_aug_subset)} samples\")" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "id": "a7c8a65e", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Train base samples: 13796\n", + "Val samples: 1533\n" + ] + } + ], + "source": [ + "from torch.utils.data import DataLoader, random_split, Subset\n", + "import torchvision.transforms as T\n", + "\n", + "# Validation / no-augmentation transform\n", + "val_transform = T.Compose([\n", + " T.Resize((256, 256)),\n", + " T.ToTensor(),\n", + "])\n", + "\n", + "# Full dataset for validation split\n", + "full_dataset = SpectrumDataset(\n", + " image_dir=\"data/pure_spectrum/images\",\n", + " label_path=\"data/pure_spectrum/labels.json\",\n", + " transform=val_transform\n", + ")\n", + "\n", + "train_size = int(0.9 * len(full_dataset))\n", + "val_size = len(full_dataset) - train_size\n", + "\n", + "train_dataset_base, val_dataset = random_split(full_dataset, [train_size, val_size])\n", + "\n", + "print(f\"Train base samples: {len(train_dataset_base)}\")\n", + "print(f\"Val samples: {len(val_dataset)}\")" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "id": "3975dce3", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "๐Ÿš€ FIXED LOADERS READY!\n", + "Train batches: 863 | Val batches: 96\n", + "Using decoupled weight decay\n", + "Stage 1 Configuration:\n", + " Epochs: 100\n", + " Optimizer: Prodigy (lr=1.0, wd=0.01)\n", + " Scheduler: CosineAnnealingLR (T_max=86300)\n", + " Early stopping patience: 30\n" + ] + } + ], + "source": [ + "# --- Stage 1 Hyperparameters ---\n", + "BATCH_SIZE = 16\n", + "\n", + "train_loader = DataLoader(train_aug_subset, batch_size=BATCH_SIZE, shuffle=True,\n", + " num_workers=0, pin_memory=False) # โ† 0 workers!\n", + "val_loader = DataLoader(val_dataset, batch_size=BATCH_SIZE, shuffle=False,\n", + " num_workers=0, pin_memory=False)\n", + "\n", + "print(f\"๐Ÿš€ FIXED LOADERS READY!\")\n", + "print(f\"Train batches: {len(train_loader)} | Val batches: {len(val_loader)}\")\n", + "NUM_EPOCHS_STAGE1 = 100 \n", + "PATIENCE = 30 \n", + "\n", + "# Prodigy optimizer โ€” parameter-free, just set lr=1.0\n", + "optimizer = Prodigy(model.parameters(), lr=1.0, weight_decay=0.01)\n", + "\n", + "# Cosine annealing scheduler\n", + "total_steps = NUM_EPOCHS_STAGE1 * len(train_loader)\n", + "scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=total_steps)\n", + "\n", + "print(f\"Stage 1 Configuration:\")\n", + "print(f\" Epochs: {NUM_EPOCHS_STAGE1}\")\n", + "print(f\" Optimizer: Prodigy (lr=1.0, wd=0.01)\")\n", + "print(f\" Scheduler: CosineAnnealingLR (T_max={total_steps})\")\n", + "print(f\" Early stopping patience: {PATIENCE}\")" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "id": "e32469e3", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "๐Ÿงช SINGLE BATCH TEST - Should finish instantly\n", + "โœ… BATCH SHAPE OK: torch.Size([16, 3, 256, 256])\n", + "โœ… FORWARD PASS OK: torch.Size([16, 3])\n", + "โœ… LOSS OK: 0.2248\n", + "๐ŸŽ‰ ALL SYSTEMS GO!\n" + ] + } + ], + "source": [ + "print(\"๐Ÿงช SINGLE BATCH TEST - Should finish instantly\")\n", + "model.train()\n", + "imgs, labels = next(iter(train_loader))\n", + "print(f\"โœ… BATCH SHAPE OK: {imgs.shape}\") # Must be [8, 3, 3, 256, 256]\n", + "\n", + "imgs, labels = imgs.to(device), labels.to(device)\n", + "preds = model(imgs)\n", + "print(f\"โœ… FORWARD PASS OK: {preds.shape}\") # [8, 3]\n", + "loss = criterion(preds, labels)\n", + "print(f\"โœ… LOSS OK: {loss.item():.4f}\")\n", + "print(\"๐ŸŽ‰ ALL SYSTEMS GO!\")" + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": {}, + "outputs": [], + "source": [ + "# import gc\n", + "# import os\n", + "# import time\n", + "# import torch\n", + "# import matplotlib.pyplot as plt\n", + "\n", + "# # === CRITICAL VRAM RESET ===\n", + "# torch.cuda.empty_cache()\n", + "# gc.collect()\n", + "# os.environ[\"PYTORCH_CUDA_ALLOC_CONF\"] = \"expandable_segments:True\"\n", + "\n", + "# BATCH_SIZE = 8\n", + "\n", + "# print(\"=\" * 80)\n", + "# print(\"๐Ÿš€ STAGE 1 TRAINING STARTED - LIVE MONITORING MODE\")\n", + "# print(\"=\" * 80)\n", + "\n", + "# train_losses, val_losses, val_accuracies = [], [], []\n", + "# best_val_loss, patience_counter = float('inf'), 0\n", + "\n", + "# # GPU memory monitoring\n", + "# if device.type == 'cuda':\n", + "# initial_mem = torch.cuda.memory_allocated() / 1e9\n", + "# print(f\"๐Ÿ“Š Initial GPU memory: {initial_mem:.1f} GB\")\n", + "# total_vram = torch.cuda.get_device_properties(0).total_memory / 1e9\n", + "# else:\n", + "# initial_mem, total_vram = 0, 0\n", + "\n", + "# print(f\"๐Ÿ“ˆ Train batches: {len(train_loader)} | Val batches: {len(val_loader)}\")\n", + "# print()\n", + "\n", + "# # =====================================================\n", + "# # ๐Ÿ”ด LIVE BATCH MONITORING - Shows progress EVERY 10 batches\n", + "# # =====================================================\n", + "# def train_one_epoch_live(model, loader, optimizer, scheduler, device, epoch):\n", + "# model.train()\n", + "# total_loss = 0.0\n", + "# n_batches = 0\n", + " \n", + "# print(f\" {'Batch':<4} {'Loss':<8} {'Time':<6} {'GPU':<6} {'ETA'}\")\n", + "# print(f\" {'-'*40}\")\n", + " \n", + "# batch_start = time.time()\n", + "# for batch_idx, (imgs, labels) in enumerate(loader):\n", + "# # === VRAM SAFETY CHECK ===\n", + "# if device.type == 'cuda' and torch.cuda.memory_allocated() / 1e9 > 7.0:\n", + "# torch.cuda.empty_cache()\n", + " \n", + "# imgs, labels = imgs.to(device, non_blocking=True), labels.to(device, non_blocking=True)\n", + " \n", + "# optimizer.zero_grad()\n", + "# preds = model(imgs)\n", + "# loss = criterion(preds, labels)\n", + "# loss.backward()\n", + "# optimizer.step()\n", + "# if scheduler: scheduler.step()\n", + " \n", + "# total_loss += loss.item()\n", + "# n_batches += 1\n", + " \n", + "# # === LIVE PROGRESS EVERY 10 BATCHES ===\n", + "# if batch_idx % 10 == 0:\n", + "# batch_time = time.time() - batch_start\n", + "# batches_per_sec = batch_idx / batch_time if batch_time > 0 else 0\n", + "# eta_epoch = (len(loader) - batch_idx) / batches_per_sec if batches_per_sec > 0 else 0\n", + " \n", + "# gpu_mem = torch.cuda.memory_allocated() / 1e9 if device.type == 'cuda' else 0\n", + "# mem_usage = ((gpu_mem - initial_mem) / total_vram * 100) if total_vram > 0 else 0\n", + " \n", + "# print(f\" {batch_idx:<4} {loss.item():<8.4f} {batch_time/10:.1f}s {'{:>6.1f}%'.format(mem_usage):<6} \"\n", + "# f\"{eta_epoch/60:.0f}m\")\n", + " \n", + "# avg_loss = total_loss / n_batches\n", + "# print(f\" {'-'*40}\")\n", + "# print(f\" โœ… Epoch {epoch} COMPLETE: {time.time()-batch_start:.0f}s total\")\n", + "# return avg_loss\n", + "\n", + "# # =====================================================\n", + "# # MAIN TRAINING LOOP WITH CONFIDENCE CHECKS\n", + "# # =====================================================\n", + "# for epoch in range(1, NUM_EPOCHS_STAGE1 + 1):\n", + "\n", + "# if epoch == 5:\n", + "# print(\"๐Ÿ”“ Unfreezing backbone for fine-tuning...\")\n", + "\n", + "# for param in model.parameters():\n", + "# param.requires_grad = True\n", + "\n", + "# optimizer = torch.optim.AdamW(\n", + "# model.parameters(),\n", + "# lr=3e-5,\n", + "# weight_decay=1e-4\n", + "# )\n", + "\n", + "# print(f\"\\n๐Ÿ”„ Epoch {epoch:3d}/{NUM_EPOCHS_STAGE1}\")\n", + " \n", + "# # ==================== TRAIN WITH LIVE MONITORING ====================\n", + "# epoch_start = time.time()\n", + "# train_loss = train_one_epoch_live(model, train_loader, optimizer, scheduler, device, epoch)\n", + "# train_losses.append(train_loss)\n", + " \n", + "# # ==================== VALIDATE ====================\n", + "# val_loss, val_acc = validate(model, val_loader, device)\n", + "# val_losses.append(val_loss)\n", + "# val_accuracies.append(val_acc)\n", + " \n", + "# # ==================== RESULTS ====================\n", + "# epoch_time = time.time() - epoch_start\n", + " \n", + "# print(f\"โœ… EPOCH {epoch} SUMMARY:\")\n", + "# print(f\" โฑ๏ธ Time: {epoch_time:.0f}s ({epoch_time/60:.1f}m)\")\n", + "# print(f\" ๐Ÿ“‰ Loss: Train={train_loss:.5f} โ†’ Val={val_loss:.5f}\")\n", + "# print(f\" ๐Ÿ“Š Acc: EC={val_acc['EC_acc']:4.1f}% EL={val_acc['EL_acc']:4.1f}% \"\n", + "# f\"EJ={val_acc['EJ_acc']:4.1f}% Overall={val_acc['overall_acc']:4.1f}%\")\n", + " \n", + "# # ==================== CHECKPOINT ====================\n", + "# if val_loss < best_val_loss:\n", + "# best_val_loss = val_loss\n", + "# patience_counter = 0\n", + " \n", + "# torch.save({\n", + "# 'epoch': epoch, 'model_state_dict': model.state_dict(),\n", + "# 'optimizer_state_dict': optimizer.state_dict(),\n", + "# 'scheduler_state_dict': scheduler.state_dict(),\n", + "# 'train_loss': train_loss, 'val_loss': val_loss, 'val_acc': val_acc,\n", + "# }, 'models/checkpoints/stage1_best.pt')\n", + "# print(f\" โญ SAVED NEW BEST! val_loss={val_loss:.5f}\")\n", + "# else:\n", + "# patience_counter += 1\n", + "# if patience_counter % 5 == 0:\n", + "# print(f\" โš ๏ธ No improvement for {patience_counter}/{PATIENCE} epochs\")\n", + " \n", + "# if patience_counter >= PATIENCE:\n", + "# print(f\"\\n๐Ÿ EARLY STOPPING at epoch {epoch}\")\n", + "# break\n", + " \n", + "# # ==================== ETA ====================\n", + "# remaining_epochs = NUM_EPOCHS_STAGE1 - epoch\n", + "# eta_total = epoch_time * remaining_epochs / 60\n", + "# print(f\" โฑ๏ธ ETA remaining: {eta_total:.0f}m | Progress: \"\n", + "# f\"[{'โ–ˆ' * min(epoch//3, 30)}{'โ–‘' * (30-min(epoch//3, 30))}]\")\n", + "# print()\n", + "\n", + "# # ==================== FINAL SUMMARY ====================\n", + "# print(\"=\" * 80)\n", + "# print(\"๐ŸŽ‰ STAGE 1 TRAINING COMPLETE!\")\n", + "# print(f\"๐Ÿ† BEST Validation Loss: {best_val_loss:.5f}\")\n", + "# print(f\"โญ BEST Model saved: models/checkpoints/stage1_best.pt\")\n", + "# print(f\"๐Ÿ“Š Final Acc: EC={val_accuracies[-1]['EC_acc']:.1f}% \"\n", + "# f\"EL={val_accuracies[-1]['EL_acc']:.1f}% \"\n", + "# f\"EJ={val_accuracies[-1]['EJ_acc']:.1f}% \"\n", + "# f\"Overall={val_accuracies[-1]['overall_acc']:.1f}%\")\n", + "# print(\"=\" * 80)\n", + "\n", + "# # Plot\n", + "# plt.figure(figsize=(12, 4))\n", + "# plt.subplot(1, 2, 1)\n", + "# plt.plot(train_losses, label='Train Loss', alpha=0.8)\n", + "# plt.plot(val_losses, label='Val Loss', alpha=0.8)\n", + "# plt.yscale('log')\n", + "# plt.xlabel('Epoch'); plt.ylabel('MSE Loss'); plt.legend(); plt.grid(True, alpha=0.3)\n", + "# plt.title('Stage 1 Training Progress')\n", + "\n", + "# plt.subplot(1, 2, 2)\n", + "# epochs_range = range(1, len(val_accuracies) + 1)\n", + "# plt.plot(epochs_range, [a['overall_acc'] for a in val_accuracies], linewidth=3, label='Overall')\n", + "# plt.plot(epochs_range, [a['EC_acc'] for a in val_accuracies], alpha=0.7, label='EC')\n", + "# plt.plot(epochs_range, [a['EL_acc'] for a in val_accuracies], alpha=0.7, label='EL')\n", + "# plt.plot(epochs_range, [a['EJ_acc'] for a in val_accuracies], alpha=0.7, label='EJ')\n", + "# plt.xlabel('Epoch'); plt.ylabel('Accuracy (%)'); plt.legend(); plt.grid(True, alpha=0.3)\n", + "# plt.title('Validation Accuracy')\n", + "\n", + "# plt.tight_layout()\n", + "# plt.savefig('results/figures/stage1_training.png', dpi=150, bbox_inches='tight')\n", + "# plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "id": "0f793d79", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "================================================================================\n", + "๐Ÿš€ STAGE 1 TRAINING STARTED - LIVE MONITORING MODE\n", + "================================================================================\n", + "๐Ÿ“Š Initial GPU memory: 1.64 GB / 8.6 GB\n", + "๐Ÿ“ˆ Train batches: 863 | Val batches: 96\n", + "\n", + "\n", + "๐Ÿ”„ Epoch 1/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2472 19.2% 1.5m\n", + " 10 0.2530 19.2% 1.3m\n", + " 20 0.2106 19.2% 1.3m\n", + " 30 0.3079 19.2% 1.3m\n", + " 40 0.2624 19.2% 1.3m\n", + " 50 0.2414 19.2% 1.3m\n", + " 60 0.2128 19.2% 1.3m\n", + " 70 0.2891 19.2% 1.3m\n", + " 80 0.2388 19.2% 1.3m\n", + " 90 0.2552 19.2% 1.2m\n", + " 100 0.2163 19.2% 1.2m\n", + " 110 0.2621 19.2% 1.2m\n", + " 120 0.2535 19.2% 1.2m\n", + " 130 0.1826 19.2% 1.2m\n", + " 140 0.2213 19.2% 1.1m\n", + " 150 0.2754 19.2% 1.1m\n", + " 160 0.2216 19.2% 1.1m\n", + " 170 0.2112 19.2% 1.1m\n", + " 180 0.1986 19.2% 1.1m\n", + " 190 0.2261 19.2% 1.0m\n", + " 200 0.1861 19.2% 1.0m\n", + " 210 0.2357 19.2% 1.0m\n", + " 220 0.2550 19.2% 1.0m\n", + " 230 0.2148 19.2% 1.0m\n", + " 240 0.2270 19.2% 1.0m\n", + " 250 0.2278 19.2% 1.0m\n", + " 260 0.2238 19.2% 0.9m\n", + " 270 0.2087 19.2% 0.9m\n", + " 280 0.2150 19.2% 0.9m\n", + " 290 0.1921 19.2% 0.9m\n", + " 300 0.2098 19.2% 0.9m\n", + " 310 0.1830 19.2% 0.9m\n", + " 320 0.1743 19.2% 0.8m\n", + " 330 0.2499 19.2% 0.8m\n", + " 340 0.2387 19.2% 0.8m\n", + " 350 0.2276 19.2% 0.8m\n", + " 360 0.2084 19.2% 0.8m\n", + " 370 0.2172 19.2% 0.8m\n", + " 380 0.2323 19.2% 0.7m\n", + " 390 0.2215 19.2% 0.7m\n", + " 400 0.2447 19.2% 0.7m\n", + " 410 0.1710 19.2% 0.7m\n", + " 420 0.2082 19.2% 0.7m\n", + " 430 0.1806 19.2% 0.7m\n", + " 440 0.2105 19.2% 0.7m\n", + " 450 0.2332 19.2% 0.6m\n", + " 460 0.1841 19.2% 0.6m\n", + " 470 0.1738 19.2% 0.6m\n", + " 480 0.1980 19.2% 0.6m\n", + " 490 0.2546 19.2% 0.6m\n", + " 500 0.2016 19.2% 0.6m\n", + " 510 0.2215 19.2% 0.5m\n", + " 520 0.2222 19.2% 0.5m\n", + " 530 0.2472 19.2% 0.5m\n", + " 540 0.1979 19.2% 0.5m\n", + " 550 0.2304 19.2% 0.5m\n", + " 560 0.2276 19.2% 0.5m\n", + " 570 0.2253 19.2% 0.5m\n", + " 580 0.2406 19.2% 0.4m\n", + " 590 0.2102 19.2% 0.4m\n", + " 600 0.1855 19.2% 0.4m\n", + " 610 0.2386 19.2% 0.4m\n", + " 620 0.2382 19.2% 0.4m\n", + " 630 0.2180 19.2% 0.4m\n", + " 640 0.1727 19.2% 0.3m\n", + " 650 0.1503 19.2% 0.3m\n", + " 660 0.1909 19.2% 0.3m\n", + " 670 0.2351 19.2% 0.3m\n", + " 680 0.2098 19.2% 0.3m\n", + " 690 0.2044 19.2% 0.3m\n", + " 700 0.1990 19.2% 0.3m\n", + " 710 0.2280 19.2% 0.2m\n", + " 720 0.2191 19.2% 0.2m\n", + " 730 0.2016 19.2% 0.2m\n", + " 740 0.2028 19.2% 0.2m\n", + " 750 0.1706 19.2% 0.2m\n", + " 760 0.1927 19.2% 0.2m\n", + " 770 0.2189 19.2% 0.1m\n", + " 780 0.2066 19.2% 0.1m\n", + " 790 0.1971 19.2% 0.1m\n", + " 800 0.1929 19.2% 0.1m\n", + " 810 0.2116 19.2% 0.1m\n", + " 820 0.2683 19.2% 0.1m\n", + " 830 0.1682 19.2% 0.1m\n", + " 840 0.2175 19.2% 0.0m\n", + " 850 0.2405 19.2% 0.0m\n", + " 860 0.2575 19.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 1 COMPLETE in 79s\n", + "\n", + "โœ… EPOCH 1 SUMMARY\n", + " โฑ๏ธ Time: 87s (1.5m)\n", + " ๐Ÿ“‰ Loss: Train=0.21985 โ†’ Val=0.19346\n", + " ๐Ÿ“Š Acc: EC=77.3% EL=75.3% EJ=82.3% Overall=78.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.19346)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 144m\n", + "\n", + "๐Ÿ”„ Epoch 2/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1957 19.2% 1.3m\n", + " 10 0.1924 19.2% 1.2m\n", + " 20 0.2241 19.2% 1.3m\n", + " 30 0.2031 19.2% 1.2m\n", + " 40 0.2044 19.2% 1.2m\n", + " 50 0.2263 19.2% 1.2m\n", + " 60 0.1737 19.2% 1.2m\n", + " 70 0.2003 19.2% 1.2m\n", + " 80 0.2021 19.2% 1.2m\n", + " 90 0.1890 19.2% 1.2m\n", + " 100 0.2434 19.2% 1.1m\n", + " 110 0.2150 19.2% 1.1m\n", + " 120 0.1761 19.2% 1.1m\n", + " 130 0.2068 19.2% 1.1m\n", + " 140 0.2333 19.2% 1.1m\n", + " 150 0.1633 19.2% 1.1m\n", + " 160 0.2011 19.2% 1.0m\n", + " 170 0.2208 19.2% 1.0m\n", + " 180 0.2214 19.2% 1.0m\n", + " 190 0.2023 19.2% 1.0m\n", + " 200 0.1579 19.2% 1.0m\n", + " 210 0.2278 19.2% 1.0m\n", + " 220 0.2410 19.2% 1.0m\n", + " 230 0.1984 19.2% 0.9m\n", + " 240 0.2335 19.2% 0.9m\n", + " 250 0.2127 19.2% 0.9m\n", + " 260 0.2159 19.2% 0.9m\n", + " 270 0.1875 19.2% 0.9m\n", + " 280 0.2565 19.2% 0.9m\n", + " 290 0.2090 19.2% 0.8m\n", + " 300 0.1988 19.2% 0.8m\n", + " 310 0.2123 19.2% 0.8m\n", + " 320 0.2031 19.2% 0.8m\n", + " 330 0.1956 19.2% 0.8m\n", + " 340 0.1886 19.2% 0.8m\n", + " 350 0.1824 19.2% 0.8m\n", + " 360 0.2352 19.2% 0.7m\n", + " 370 0.2007 19.2% 0.7m\n", + " 380 0.1815 19.2% 0.7m\n", + " 390 0.2212 19.2% 0.7m\n", + " 400 0.1897 19.2% 0.7m\n", + " 410 0.2151 19.2% 0.7m\n", + " 420 0.2301 19.2% 0.7m\n", + " 430 0.1826 19.2% 0.6m\n", + " 440 0.1970 19.2% 0.6m\n", + " 450 0.1973 19.2% 0.6m\n", + " 460 0.1855 19.2% 0.6m\n", + " 470 0.1907 19.2% 0.6m\n", + " 480 0.1751 19.2% 0.6m\n", + " 490 0.2161 19.2% 0.6m\n", + " 500 0.2178 19.2% 0.5m\n", + " 510 0.2260 19.2% 0.5m\n", + " 520 0.2518 19.2% 0.5m\n", + " 530 0.1946 19.2% 0.5m\n", + " 540 0.2219 19.2% 0.5m\n", + " 550 0.2079 19.2% 0.5m\n", + " 560 0.2136 19.2% 0.4m\n", + " 570 0.2298 19.2% 0.4m\n", + " 580 0.2167 19.2% 0.4m\n", + " 590 0.1925 19.2% 0.4m\n", + " 600 0.2294 19.2% 0.4m\n", + " 610 0.1767 19.2% 0.4m\n", + " 620 0.2161 19.2% 0.4m\n", + " 630 0.2409 19.2% 0.3m\n", + " 640 0.1891 19.2% 0.3m\n", + " 650 0.2312 19.2% 0.3m\n", + " 660 0.2072 19.2% 0.3m\n", + " 670 0.1964 19.2% 0.3m\n", + " 680 0.1837 19.2% 0.3m\n", + " 690 0.2415 19.2% 0.3m\n", + " 700 0.2107 19.2% 0.2m\n", + " 710 0.2132 19.2% 0.2m\n", + " 720 0.2043 19.2% 0.2m\n", + " 730 0.1911 19.2% 0.2m\n", + " 740 0.1735 19.2% 0.2m\n", + " 750 0.2563 19.2% 0.2m\n", + " 760 0.1817 19.2% 0.2m\n", + " 770 0.2153 19.2% 0.1m\n", + " 780 0.1810 19.2% 0.1m\n", + " 790 0.2070 19.2% 0.1m\n", + " 800 0.1926 19.2% 0.1m\n", + " 810 0.2020 19.2% 0.1m\n", + " 820 0.2237 19.2% 0.1m\n", + " 830 0.2141 19.2% 0.0m\n", + " 840 0.1639 19.2% 0.0m\n", + " 850 0.2139 19.2% 0.0m\n", + " 860 0.2221 19.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 2 COMPLETE in 76s\n", + "\n", + "โœ… EPOCH 2 SUMMARY\n", + " โฑ๏ธ Time: 84s (1.4m)\n", + " ๐Ÿ“‰ Loss: Train=0.20724 โ†’ Val=0.18636\n", + " ๐Ÿ“Š Acc: EC=78.4% EL=75.2% EJ=83.2% Overall=79.0%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.18636)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 138m\n", + "\n", + "๐Ÿ”„ Epoch 3/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1889 19.2% 1.3m\n", + " 10 0.2106 19.2% 1.3m\n", + " 20 0.1685 19.2% 1.3m\n", + " 30 0.1724 19.2% 1.2m\n", + " 40 0.2081 19.2% 1.2m\n", + " 50 0.1747 19.2% 1.2m\n", + " 60 0.2175 19.2% 1.2m\n", + " 70 0.1841 19.2% 1.2m\n", + " 80 0.1916 19.2% 1.2m\n", + " 90 0.1906 19.2% 1.2m\n", + " 100 0.2225 19.2% 1.1m\n", + " 110 0.2124 19.2% 1.1m\n", + " 120 0.1825 19.2% 1.1m\n", + " 130 0.2043 19.2% 1.1m\n", + " 140 0.2111 19.2% 1.1m\n", + " 150 0.2374 19.2% 1.1m\n", + " 160 0.2118 19.2% 1.0m\n", + " 170 0.1914 19.2% 1.0m\n", + " 180 0.2101 19.2% 1.0m\n", + " 190 0.1893 19.2% 1.0m\n", + " 200 0.2016 19.2% 1.0m\n", + " 210 0.2081 19.2% 1.0m\n", + " 220 0.1912 19.2% 1.0m\n", + " 230 0.2012 19.2% 0.9m\n", + " 240 0.2293 19.2% 0.9m\n", + " 250 0.2005 19.2% 0.9m\n", + " 260 0.2061 19.2% 0.9m\n", + " 270 0.1731 19.2% 0.9m\n", + " 280 0.2000 19.2% 0.9m\n", + " 290 0.2016 19.2% 0.8m\n", + " 300 0.2155 19.2% 0.8m\n", + " 310 0.1786 19.2% 0.8m\n", + " 320 0.1791 19.2% 0.8m\n", + " 330 0.1749 19.2% 0.8m\n", + " 340 0.1867 19.2% 0.8m\n", + " 350 0.2014 19.2% 0.8m\n", + " 360 0.2502 19.2% 0.7m\n", + " 370 0.1833 19.2% 0.7m\n", + " 380 0.1713 19.2% 0.7m\n", + " 390 0.2069 19.2% 0.7m\n", + " 400 0.2151 19.2% 0.7m\n", + " 410 0.1956 19.2% 0.7m\n", + " 420 0.2243 19.2% 0.7m\n", + " 430 0.2419 19.2% 0.6m\n", + " 440 0.2135 19.2% 0.6m\n", + " 450 0.2303 19.2% 0.6m\n", + " 460 0.2491 19.2% 0.6m\n", + " 470 0.1884 19.2% 0.6m\n", + " 480 0.2148 19.2% 0.6m\n", + " 490 0.1923 19.2% 0.6m\n", + " 500 0.2285 19.2% 0.5m\n", + " 510 0.1824 19.2% 0.5m\n", + " 520 0.1904 19.2% 0.5m\n", + " 530 0.1852 19.2% 0.5m\n", + " 540 0.1858 19.2% 0.5m\n", + " 550 0.2038 19.2% 0.5m\n", + " 560 0.1833 19.2% 0.4m\n", + " 570 0.2278 19.2% 0.4m\n", + " 580 0.1955 19.2% 0.4m\n", + " 590 0.2186 19.2% 0.4m\n", + " 600 0.2532 19.2% 0.4m\n", + " 610 0.1682 19.2% 0.4m\n", + " 620 0.2185 19.2% 0.4m\n", + " 630 0.2217 19.2% 0.3m\n", + " 640 0.2038 19.2% 0.3m\n", + " 650 0.2187 19.2% 0.3m\n", + " 660 0.1561 19.2% 0.3m\n", + " 670 0.1932 19.2% 0.3m\n", + " 680 0.1941 19.2% 0.3m\n", + " 690 0.1747 19.2% 0.3m\n", + " 700 0.2077 19.2% 0.2m\n", + " 710 0.1857 19.2% 0.2m\n", + " 720 0.1813 19.2% 0.2m\n", + " 730 0.1908 19.2% 0.2m\n", + " 740 0.1704 19.2% 0.2m\n", + " 750 0.2433 19.2% 0.2m\n", + " 760 0.1707 19.2% 0.2m\n", + " 770 0.2054 19.2% 0.1m\n", + " 780 0.1889 19.2% 0.1m\n", + " 790 0.2201 19.2% 0.1m\n", + " 800 0.1982 19.2% 0.1m\n", + " 810 0.2110 19.2% 0.1m\n", + " 820 0.2376 19.2% 0.1m\n", + " 830 0.2255 19.2% 0.0m\n", + " 840 0.2102 19.2% 0.0m\n", + " 850 0.1913 19.2% 0.0m\n", + " 860 0.1935 19.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 3 COMPLETE in 76s\n", + "\n", + "โœ… EPOCH 3 SUMMARY\n", + " โฑ๏ธ Time: 84s (1.4m)\n", + " ๐Ÿ“‰ Loss: Train=0.20445 โ†’ Val=0.19048\n", + " ๐Ÿ“Š Acc: EC=77.6% EL=75.1% EJ=82.9% Overall=78.6%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 136m\n", + "\n", + "๐Ÿ”„ Epoch 4/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1941 19.2% 1.3m\n", + " 10 0.1870 19.2% 1.2m\n", + " 20 0.2053 19.2% 1.2m\n", + " 30 0.1908 19.2% 1.2m\n", + " 40 0.2362 19.2% 1.2m\n", + " 50 0.1640 19.2% 1.2m\n", + " 60 0.2297 19.2% 1.2m\n", + " 70 0.2164 19.2% 1.2m\n", + " 80 0.2061 19.2% 1.2m\n", + " 90 0.2465 19.2% 1.1m\n", + " 100 0.2115 19.2% 1.1m\n", + " 110 0.2441 19.2% 1.1m\n", + " 120 0.2018 19.2% 1.1m\n", + " 130 0.2199 19.2% 1.1m\n", + " 140 0.2079 19.2% 1.1m\n", + " 150 0.2007 19.2% 1.0m\n", + " 160 0.2301 19.2% 1.0m\n", + " 170 0.1729 19.2% 1.0m\n", + " 180 0.2324 19.2% 1.0m\n", + " 190 0.2004 19.2% 1.0m\n", + " 200 0.1640 19.2% 1.0m\n", + " 210 0.2141 19.2% 1.0m\n", + " 220 0.1790 19.2% 0.9m\n", + " 230 0.2066 19.2% 0.9m\n", + " 240 0.2202 19.2% 0.9m\n", + " 250 0.2172 19.2% 0.9m\n", + " 260 0.1974 19.2% 0.9m\n", + " 270 0.1866 19.2% 0.9m\n", + " 280 0.2331 19.2% 0.9m\n", + " 290 0.2659 19.2% 0.8m\n", + " 300 0.2369 19.2% 0.8m\n", + " 310 0.1786 19.2% 0.8m\n", + " 320 0.2020 19.2% 0.8m\n", + " 330 0.1955 19.2% 0.8m\n", + " 340 0.2216 19.2% 0.8m\n", + " 350 0.1974 19.2% 0.8m\n", + " 360 0.2357 19.2% 0.7m\n", + " 370 0.2388 19.2% 0.7m\n", + " 380 0.2122 19.2% 0.7m\n", + " 390 0.2393 19.2% 0.7m\n", + " 400 0.2170 19.2% 0.7m\n", + " 410 0.2160 19.2% 0.7m\n", + " 420 0.2353 19.2% 0.7m\n", + " 430 0.2140 19.2% 0.6m\n", + " 440 0.2248 19.2% 0.6m\n", + " 450 0.2467 19.2% 0.6m\n", + " 460 0.1934 19.2% 0.6m\n", + " 470 0.1917 19.2% 0.6m\n", + " 480 0.1795 19.2% 0.6m\n", + " 490 0.2194 19.2% 0.5m\n", + " 500 0.1613 19.2% 0.5m\n", + " 510 0.1675 19.2% 0.5m\n", + " 520 0.1758 19.2% 0.5m\n", + " 530 0.2239 19.2% 0.5m\n", + " 540 0.2261 19.2% 0.5m\n", + " 550 0.2230 19.2% 0.5m\n", + " 560 0.2120 19.2% 0.4m\n", + " 570 0.2024 19.2% 0.4m\n", + " 580 0.2100 19.2% 0.4m\n", + " 590 0.2093 19.2% 0.4m\n", + " 600 0.2004 19.2% 0.4m\n", + " 610 0.2229 19.2% 0.4m\n", + " 620 0.1803 19.2% 0.4m\n", + " 630 0.1972 19.2% 0.3m\n", + " 640 0.2400 19.2% 0.3m\n", + " 650 0.2080 19.2% 0.3m\n", + " 660 0.2110 19.2% 0.3m\n", + " 670 0.2119 19.2% 0.3m\n", + " 680 0.1772 19.2% 0.3m\n", + " 690 0.2010 19.2% 0.3m\n", + " 700 0.1864 19.2% 0.2m\n", + " 710 0.2035 19.2% 0.2m\n", + " 720 0.1645 19.2% 0.2m\n", + " 730 0.1959 19.2% 0.2m\n", + " 740 0.2193 19.2% 0.2m\n", + " 750 0.2278 19.2% 0.2m\n", + " 760 0.1874 19.2% 0.2m\n", + " 770 0.2198 19.2% 0.1m\n", + " 780 0.2554 19.2% 0.1m\n", + " 790 0.1878 19.2% 0.1m\n", + " 800 0.2034 19.2% 0.1m\n", + " 810 0.1927 19.2% 0.1m\n", + " 820 0.2278 19.2% 0.1m\n", + " 830 0.1840 19.2% 0.0m\n", + " 840 0.2607 19.2% 0.0m\n", + " 850 0.1977 19.2% 0.0m\n", + " 860 0.2148 19.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 4 COMPLETE in 76s\n", + "\n", + "โœ… EPOCH 4 SUMMARY\n", + " โฑ๏ธ Time: 84s (1.4m)\n", + " ๐Ÿ“‰ Loss: Train=0.20333 โ†’ Val=0.18967\n", + " ๐Ÿ“Š Acc: EC=77.1% EL=75.4% EJ=83.4% Overall=78.6%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 134m\n", + "๐Ÿ”“ Unfreezing backbone for fine-tuning\n", + "\n", + "๐Ÿ”„ Epoch 5/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1647 20.6% 2.7m\n", + " 10 0.2265 23.2% 2.6m\n", + " 20 0.2122 23.2% 2.6m\n", + " 30 0.1643 23.2% 2.5m\n", + " 40 0.1648 23.2% 2.5m\n", + " 50 0.1915 23.2% 2.5m\n", + " 60 0.2168 23.2% 2.4m\n", + " 70 0.1688 23.2% 2.4m\n", + " 80 0.1754 23.2% 2.4m\n", + " 90 0.2131 23.2% 2.4m\n", + " 100 0.2259 23.2% 2.3m\n", + " 110 0.1676 23.2% 2.3m\n", + " 120 0.1537 23.2% 2.3m\n", + " 130 0.1873 23.2% 2.3m\n", + " 140 0.1738 23.2% 2.2m\n", + " 150 0.2028 23.2% 2.2m\n", + " 160 0.1707 23.2% 2.2m\n", + " 170 0.1613 23.2% 2.1m\n", + " 180 0.1924 23.2% 2.1m\n", + " 190 0.1458 23.2% 2.1m\n", + " 200 0.2189 23.2% 2.0m\n", + " 210 0.1808 23.2% 2.0m\n", + " 220 0.1322 23.2% 2.0m\n", + " 230 0.1389 23.2% 1.9m\n", + " 240 0.1998 23.2% 1.9m\n", + " 250 0.1621 23.2% 1.9m\n", + " 260 0.1275 23.2% 1.8m\n", + " 270 0.1741 23.2% 1.8m\n", + " 280 0.1372 23.2% 1.8m\n", + " 290 0.1768 23.2% 1.8m\n", + " 300 0.1751 23.2% 1.7m\n", + " 310 0.1716 23.2% 1.7m\n", + " 320 0.1709 23.2% 1.7m\n", + " 330 0.1710 23.2% 1.6m\n", + " 340 0.1700 23.2% 1.6m\n", + " 350 0.1885 23.2% 1.6m\n", + " 360 0.1631 23.2% 1.5m\n", + " 370 0.1928 23.2% 1.5m\n", + " 380 0.1583 23.2% 1.5m\n", + " 390 0.1921 23.2% 1.5m\n", + " 400 0.1574 23.2% 1.4m\n", + " 410 0.1638 23.2% 1.4m\n", + " 420 0.2143 23.2% 1.4m\n", + " 430 0.1529 23.2% 1.3m\n", + " 440 0.1594 23.2% 1.3m\n", + " 450 0.1604 23.2% 1.3m\n", + " 460 0.1800 23.2% 1.2m\n", + " 470 0.1400 23.2% 1.2m\n", + " 480 0.2051 23.2% 1.2m\n", + " 490 0.1685 23.2% 1.1m\n", + " 500 0.1498 23.2% 1.1m\n", + " 510 0.1781 23.2% 1.1m\n", + " 520 0.1942 23.2% 1.1m\n", + " 530 0.1812 23.2% 1.0m\n", + " 540 0.1415 23.2% 1.0m\n", + " 550 0.1788 23.2% 1.0m\n", + " 560 0.1731 23.2% 0.9m\n", + " 570 0.1748 23.2% 0.9m\n", + " 580 0.1937 23.2% 0.9m\n", + " 590 0.2392 23.2% 0.8m\n", + " 600 0.1820 23.2% 0.8m\n", + " 610 0.1918 23.2% 0.8m\n", + " 620 0.1318 23.2% 0.7m\n", + " 630 0.1313 23.2% 0.7m\n", + " 640 0.1687 23.2% 0.7m\n", + " 650 0.1414 23.2% 0.7m\n", + " 660 0.1874 23.2% 0.6m\n", + " 670 0.1427 23.2% 0.6m\n", + " 680 0.1797 23.2% 0.6m\n", + " 690 0.1713 23.2% 0.5m\n", + " 700 0.1618 23.2% 0.5m\n", + " 710 0.2211 23.2% 0.5m\n", + " 720 0.1729 23.2% 0.4m\n", + " 730 0.1626 23.2% 0.4m\n", + " 740 0.1670 23.2% 0.4m\n", + " 750 0.2094 23.2% 0.3m\n", + " 760 0.1432 23.2% 0.3m\n", + " 770 0.1765 23.2% 0.3m\n", + " 780 0.1197 23.2% 0.3m\n", + " 790 0.1961 23.2% 0.2m\n", + " 800 0.1428 23.2% 0.2m\n", + " 810 0.1643 23.2% 0.2m\n", + " 820 0.2214 23.2% 0.1m\n", + " 830 0.1887 23.2% 0.1m\n", + " 840 0.1608 23.2% 0.1m\n", + " 850 0.1819 23.2% 0.0m\n", + " 860 0.1558 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 5 COMPLETE in 160s\n", + "\n", + "โœ… EPOCH 5 SUMMARY\n", + " โฑ๏ธ Time: 168s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.17745 โ†’ Val=0.16768\n", + " ๐Ÿ“Š Acc: EC=81.2% EL=75.4% EJ=86.0% Overall=80.9%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16768)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 266m\n", + "\n", + "๐Ÿ”„ Epoch 6/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1224 23.2% 2.7m\n", + " 10 0.2069 23.2% 2.7m\n", + " 20 0.1790 23.2% 2.6m\n", + " 30 0.1669 23.2% 2.6m\n", + " 40 0.1552 23.2% 2.5m\n", + " 50 0.1759 23.2% 2.5m\n", + " 60 0.1684 23.2% 2.5m\n", + " 70 0.2020 23.2% 2.5m\n", + " 80 0.1706 23.2% 2.4m\n", + " 90 0.1793 23.2% 2.4m\n", + " 100 0.1473 23.2% 2.4m\n", + " 110 0.1710 23.2% 2.4m\n", + " 120 0.1773 23.2% 2.3m\n", + " 130 0.1645 23.2% 2.3m\n", + " 140 0.1725 23.2% 2.3m\n", + " 150 0.1565 23.2% 2.2m\n", + " 160 0.1655 23.2% 2.2m\n", + " 170 0.1797 23.2% 2.2m\n", + " 180 0.1860 23.2% 2.1m\n", + " 190 0.1225 23.2% 2.1m\n", + " 200 0.1517 23.2% 2.1m\n", + " 210 0.1820 23.2% 2.0m\n", + " 220 0.2083 23.2% 2.0m\n", + " 230 0.1628 23.2% 2.0m\n", + " 240 0.1466 23.2% 2.0m\n", + " 250 0.2110 23.2% 1.9m\n", + " 260 0.2070 23.2% 1.9m\n", + " 270 0.1650 23.2% 1.9m\n", + " 280 0.1719 23.2% 1.8m\n", + " 290 0.1572 23.2% 1.8m\n", + " 300 0.1631 23.2% 1.8m\n", + " 310 0.1570 23.2% 1.7m\n", + " 320 0.1614 23.2% 1.7m\n", + " 330 0.1804 23.2% 1.7m\n", + " 340 0.1464 23.2% 1.6m\n", + " 350 0.1767 23.2% 1.6m\n", + " 360 0.1653 23.2% 1.6m\n", + " 370 0.1973 23.2% 1.5m\n", + " 380 0.1688 23.2% 1.5m\n", + " 390 0.1952 23.2% 1.5m\n", + " 400 0.1667 23.2% 1.4m\n", + " 410 0.1650 23.2% 1.4m\n", + " 420 0.1923 23.2% 1.4m\n", + " 430 0.1600 23.2% 1.4m\n", + " 440 0.1673 23.2% 1.3m\n", + " 450 0.1230 23.2% 1.3m\n", + " 460 0.1602 23.2% 1.3m\n", + " 470 0.1792 23.2% 1.2m\n", + " 480 0.1338 23.2% 1.2m\n", + " 490 0.1822 23.2% 1.2m\n", + " 500 0.1925 23.2% 1.1m\n", + " 510 0.1782 23.2% 1.1m\n", + " 520 0.1841 23.2% 1.1m\n", + " 530 0.1302 23.2% 1.0m\n", + " 540 0.1160 23.2% 1.0m\n", + " 550 0.1820 23.2% 1.0m\n", + " 560 0.2117 23.2% 0.9m\n", + " 570 0.1935 23.2% 0.9m\n", + " 580 0.1783 23.2% 0.9m\n", + " 590 0.1795 23.2% 0.9m\n", + " 600 0.1757 23.2% 0.8m\n", + " 610 0.1738 23.2% 0.8m\n", + " 620 0.1989 23.2% 0.8m\n", + " 630 0.1435 23.2% 0.7m\n", + " 640 0.1543 23.2% 0.7m\n", + " 650 0.1546 23.2% 0.7m\n", + " 660 0.1883 23.2% 0.6m\n", + " 670 0.1533 23.2% 0.6m\n", + " 680 0.1625 23.2% 0.6m\n", + " 690 0.1613 23.2% 0.5m\n", + " 700 0.1968 23.2% 0.5m\n", + " 710 0.1474 23.2% 0.5m\n", + " 720 0.1448 23.2% 0.4m\n", + " 730 0.1878 23.2% 0.4m\n", + " 740 0.1963 23.2% 0.4m\n", + " 750 0.1499 23.2% 0.4m\n", + " 760 0.1782 23.2% 0.3m\n", + " 770 0.1655 23.2% 0.3m\n", + " 780 0.1742 23.2% 0.3m\n", + " 790 0.1271 23.2% 0.2m\n", + " 800 0.1560 23.2% 0.2m\n", + " 810 0.1827 23.2% 0.2m\n", + " 820 0.2088 23.2% 0.1m\n", + " 830 0.1657 23.2% 0.1m\n", + " 840 0.1881 23.2% 0.1m\n", + " 850 0.1889 23.2% 0.0m\n", + " 860 0.2020 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 6 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 6 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.17035 โ†’ Val=0.16544\n", + " ๐Ÿ“Š Acc: EC=81.3% EL=75.3% EJ=86.7% Overall=81.1%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16544)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 266m\n", + "\n", + "๐Ÿ”„ Epoch 7/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1527 23.2% 2.7m\n", + " 10 0.1547 23.2% 2.7m\n", + " 20 0.1891 23.2% 2.6m\n", + " 30 0.1783 23.2% 2.6m\n", + " 40 0.1913 23.2% 2.6m\n", + " 50 0.1636 23.2% 2.6m\n", + " 60 0.1826 23.2% 2.5m\n", + " 70 0.2061 23.2% 2.5m\n", + " 80 0.1859 23.2% 2.5m\n", + " 90 0.1370 23.2% 2.4m\n", + " 100 0.1648 23.2% 2.4m\n", + " 110 0.1569 23.2% 2.4m\n", + " 120 0.1455 23.2% 2.3m\n", + " 130 0.2262 23.2% 2.3m\n", + " 140 0.1581 23.2% 2.3m\n", + " 150 0.1313 23.2% 2.2m\n", + " 160 0.1886 23.2% 2.2m\n", + " 170 0.1827 23.2% 2.2m\n", + " 180 0.1788 23.2% 2.1m\n", + " 190 0.1826 23.2% 2.1m\n", + " 200 0.1542 23.2% 2.1m\n", + " 210 0.1719 23.2% 2.0m\n", + " 220 0.1915 23.2% 2.0m\n", + " 230 0.1539 23.2% 2.0m\n", + " 240 0.1461 23.2% 1.9m\n", + " 250 0.1604 23.2% 1.9m\n", + " 260 0.1411 23.2% 1.9m\n", + " 270 0.1768 23.2% 1.8m\n", + " 280 0.1726 23.2% 1.8m\n", + " 290 0.1375 23.2% 1.8m\n", + " 300 0.1597 23.2% 1.7m\n", + " 310 0.1941 23.2% 1.7m\n", + " 320 0.1386 23.2% 1.7m\n", + " 330 0.1929 23.2% 1.7m\n", + " 340 0.1490 23.2% 1.6m\n", + " 350 0.1856 23.2% 1.6m\n", + " 360 0.1838 23.2% 1.6m\n", + " 370 0.2080 23.2% 1.5m\n", + " 380 0.1830 23.2% 1.5m\n", + " 390 0.2111 23.2% 1.5m\n", + " 400 0.1426 23.2% 1.4m\n", + " 410 0.1630 23.2% 1.4m\n", + " 420 0.1643 23.2% 1.4m\n", + " 430 0.1377 23.2% 1.3m\n", + " 440 0.1500 23.2% 1.3m\n", + " 450 0.2040 23.2% 1.3m\n", + " 460 0.1881 23.2% 1.2m\n", + " 470 0.1708 23.2% 1.2m\n", + " 480 0.2086 23.2% 1.2m\n", + " 490 0.1470 23.2% 1.2m\n", + " 500 0.1580 23.2% 1.1m\n", + " 510 0.1527 23.2% 1.1m\n", + " 520 0.1878 23.2% 1.1m\n", + " 530 0.1865 23.2% 1.0m\n", + " 540 0.1688 23.2% 1.0m\n", + " 550 0.1740 23.2% 1.0m\n", + " 560 0.1454 23.2% 0.9m\n", + " 570 0.1439 23.2% 0.9m\n", + " 580 0.1491 23.2% 0.9m\n", + " 590 0.1763 23.2% 0.8m\n", + " 600 0.1415 23.2% 0.8m\n", + " 610 0.2339 23.2% 0.8m\n", + " 620 0.1749 23.2% 0.8m\n", + " 630 0.1545 23.2% 0.7m\n", + " 640 0.1543 23.2% 0.7m\n", + " 650 0.1541 23.2% 0.7m\n", + " 660 0.1424 23.2% 0.6m\n", + " 670 0.1550 23.2% 0.6m\n", + " 680 0.1654 23.2% 0.6m\n", + " 690 0.1821 23.2% 0.5m\n", + " 700 0.2063 23.2% 0.5m\n", + " 710 0.1811 23.2% 0.5m\n", + " 720 0.1990 23.2% 0.4m\n", + " 730 0.1541 23.2% 0.4m\n", + " 740 0.1697 23.2% 0.4m\n", + " 750 0.1705 23.2% 0.4m\n", + " 760 0.1930 23.2% 0.3m\n", + " 770 0.1469 23.2% 0.3m\n", + " 780 0.1646 23.2% 0.3m\n", + " 790 0.1629 23.2% 0.2m\n", + " 800 0.1858 23.2% 0.2m\n", + " 810 0.1742 23.2% 0.2m\n", + " 820 0.1652 23.2% 0.1m\n", + " 830 0.1601 23.2% 0.1m\n", + " 840 0.1278 23.2% 0.1m\n", + " 850 0.1747 23.2% 0.0m\n", + " 860 0.1841 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 7 COMPLETE in 160s\n", + "\n", + "โœ… EPOCH 7 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16914 โ†’ Val=0.16483\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.3% EJ=86.6% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16483)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 261m\n", + "\n", + "๐Ÿ”„ Epoch 8/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1742 23.2% 2.7m\n", + " 10 0.1522 23.2% 2.7m\n", + " 20 0.1706 23.2% 2.7m\n", + " 30 0.1478 23.2% 2.6m\n", + " 40 0.1669 23.2% 2.6m\n", + " 50 0.1899 23.2% 2.5m\n", + " 60 0.1736 23.2% 2.5m\n", + " 70 0.1573 23.2% 2.5m\n", + " 80 0.1481 23.2% 2.4m\n", + " 90 0.2096 23.2% 2.4m\n", + " 100 0.1901 23.2% 2.4m\n", + " 110 0.1602 23.2% 2.3m\n", + " 120 0.1652 23.2% 2.3m\n", + " 130 0.1901 23.2% 2.3m\n", + " 140 0.0976 23.2% 2.3m\n", + " 150 0.1770 23.2% 2.2m\n", + " 160 0.1776 23.2% 2.2m\n", + " 170 0.1605 23.2% 2.2m\n", + " 180 0.1729 23.2% 2.1m\n", + " 190 0.1597 23.2% 2.1m\n", + " 200 0.1834 23.2% 2.1m\n", + " 210 0.1644 23.2% 2.0m\n", + " 220 0.1683 23.2% 2.0m\n", + " 230 0.1616 23.2% 2.0m\n", + " 240 0.1931 23.2% 1.9m\n", + " 250 0.1750 23.2% 1.9m\n", + " 260 0.2016 23.2% 1.9m\n", + " 270 0.1975 23.2% 1.9m\n", + " 280 0.1892 23.2% 1.8m\n", + " 290 0.1634 23.2% 1.8m\n", + " 300 0.1873 23.2% 1.8m\n", + " 310 0.1670 23.2% 1.7m\n", + " 320 0.1430 23.2% 1.7m\n", + " 330 0.1718 23.2% 1.7m\n", + " 340 0.1747 23.2% 1.6m\n", + " 350 0.2115 23.2% 1.6m\n", + " 360 0.1595 23.2% 1.6m\n", + " 370 0.1721 23.2% 1.5m\n", + " 380 0.1878 23.2% 1.5m\n", + " 390 0.1661 23.2% 1.5m\n", + " 400 0.1870 23.2% 1.4m\n", + " 410 0.1861 23.2% 1.4m\n", + " 420 0.1335 23.2% 1.4m\n", + " 430 0.1474 23.2% 1.4m\n", + " 440 0.1445 23.2% 1.3m\n", + " 450 0.1833 23.2% 1.3m\n", + " 460 0.2387 23.2% 1.3m\n", + " 470 0.1802 23.2% 1.2m\n", + " 480 0.1716 23.2% 1.2m\n", + " 490 0.1594 23.2% 1.2m\n", + " 500 0.2219 23.2% 1.1m\n", + " 510 0.1742 23.2% 1.1m\n", + " 520 0.1777 23.2% 1.1m\n", + " 530 0.2004 23.2% 1.0m\n", + " 540 0.1726 23.2% 1.0m\n", + " 550 0.1769 23.2% 1.0m\n", + " 560 0.1660 23.2% 0.9m\n", + " 570 0.1448 23.2% 0.9m\n", + " 580 0.1389 23.2% 0.9m\n", + " 590 0.2252 23.2% 0.9m\n", + " 600 0.1344 23.2% 0.8m\n", + " 610 0.1604 23.2% 0.8m\n", + " 620 0.1626 23.2% 0.8m\n", + " 630 0.1546 23.2% 0.7m\n", + " 640 0.1575 23.2% 0.7m\n", + " 650 0.1731 23.2% 0.7m\n", + " 660 0.1365 23.2% 0.6m\n", + " 670 0.1499 23.2% 0.6m\n", + " 680 0.1679 23.2% 0.6m\n", + " 690 0.1781 23.2% 0.5m\n", + " 700 0.1454 23.2% 0.5m\n", + " 710 0.1717 23.2% 0.5m\n", + " 720 0.1881 23.2% 0.4m\n", + " 730 0.1569 23.2% 0.4m\n", + " 740 0.1837 23.2% 0.4m\n", + " 750 0.1454 23.2% 0.4m\n", + " 760 0.1800 23.2% 0.3m\n", + " 770 0.1911 23.2% 0.3m\n", + " 780 0.1956 23.2% 0.3m\n", + " 790 0.1554 23.2% 0.2m\n", + " 800 0.2131 23.2% 0.2m\n", + " 810 0.1367 23.2% 0.2m\n", + " 820 0.1910 23.2% 0.1m\n", + " 830 0.1583 23.2% 0.1m\n", + " 840 0.1837 23.2% 0.1m\n", + " 850 0.1243 23.2% 0.0m\n", + " 860 0.1725 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 8 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 8 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16872 โ†’ Val=0.16462\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.3% EJ=86.7% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16462)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 260m\n", + "\n", + "๐Ÿ”„ Epoch 9/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1987 23.2% 2.6m\n", + " 10 0.1602 23.2% 2.7m\n", + " 20 0.1346 23.2% 2.6m\n", + " 30 0.1503 23.2% 2.6m\n", + " 40 0.2123 23.2% 2.6m\n", + " 50 0.1551 23.2% 2.5m\n", + " 60 0.1605 23.2% 2.5m\n", + " 70 0.1789 23.2% 2.5m\n", + " 80 0.2188 23.2% 2.5m\n", + " 90 0.2130 23.2% 2.4m\n", + " 100 0.1747 23.2% 2.4m\n", + " 110 0.1931 23.2% 2.4m\n", + " 120 0.1262 23.2% 2.3m\n", + " 130 0.1677 23.2% 2.3m\n", + " 140 0.1761 23.2% 2.3m\n", + " 150 0.1845 23.2% 2.2m\n", + " 160 0.1625 23.2% 2.2m\n", + " 170 0.2198 23.2% 2.2m\n", + " 180 0.2066 23.2% 2.1m\n", + " 190 0.1935 23.2% 2.1m\n", + " 200 0.1487 23.2% 2.1m\n", + " 210 0.1492 23.2% 2.1m\n", + " 220 0.1675 23.2% 2.0m\n", + " 230 0.1370 23.2% 2.0m\n", + " 240 0.1846 23.2% 2.0m\n", + " 250 0.1908 23.2% 1.9m\n", + " 260 0.1714 23.2% 1.9m\n", + " 270 0.2204 23.2% 1.9m\n", + " 280 0.1506 23.2% 1.8m\n", + " 290 0.1656 23.2% 1.8m\n", + " 300 0.1809 23.2% 1.8m\n", + " 310 0.1823 23.2% 1.7m\n", + " 320 0.2135 23.2% 1.7m\n", + " 330 0.1683 23.2% 1.7m\n", + " 340 0.1414 23.2% 1.6m\n", + " 350 0.1525 23.2% 1.6m\n", + " 360 0.1418 23.2% 1.6m\n", + " 370 0.1615 23.2% 1.5m\n", + " 380 0.1519 23.2% 1.5m\n", + " 390 0.1542 23.2% 1.5m\n", + " 400 0.1671 23.2% 1.4m\n", + " 410 0.1447 23.2% 1.4m\n", + " 420 0.1919 23.2% 1.4m\n", + " 430 0.1775 23.2% 1.4m\n", + " 440 0.1292 23.2% 1.3m\n", + " 450 0.1705 23.2% 1.3m\n", + " 460 0.1612 23.2% 1.3m\n", + " 470 0.1411 23.2% 1.2m\n", + " 480 0.1811 23.2% 1.2m\n", + " 490 0.1568 23.2% 1.2m\n", + " 500 0.1620 23.2% 1.1m\n", + " 510 0.1325 23.2% 1.1m\n", + " 520 0.1648 23.2% 1.1m\n", + " 530 0.1879 23.2% 1.0m\n", + " 540 0.1389 23.2% 1.0m\n", + " 550 0.2039 23.2% 1.0m\n", + " 560 0.1540 23.2% 0.9m\n", + " 570 0.1761 23.2% 0.9m\n", + " 580 0.1954 23.2% 0.9m\n", + " 590 0.1674 23.2% 0.9m\n", + " 600 0.1364 23.2% 0.8m\n", + " 610 0.1759 23.2% 0.8m\n", + " 620 0.1405 23.2% 0.8m\n", + " 630 0.1626 23.2% 0.7m\n", + " 640 0.1465 23.2% 0.7m\n", + " 650 0.1625 23.2% 0.7m\n", + " 660 0.1322 23.2% 0.6m\n", + " 670 0.1865 23.2% 0.6m\n", + " 680 0.1678 23.2% 0.6m\n", + " 690 0.1984 23.2% 0.5m\n", + " 700 0.1849 23.2% 0.5m\n", + " 710 0.1475 23.2% 0.5m\n", + " 720 0.1596 23.2% 0.4m\n", + " 730 0.1622 23.2% 0.4m\n", + " 740 0.1518 23.2% 0.4m\n", + " 750 0.1639 23.2% 0.4m\n", + " 760 0.1780 23.2% 0.3m\n", + " 770 0.1538 23.2% 0.3m\n", + " 780 0.1435 23.2% 0.3m\n", + " 790 0.2093 23.2% 0.2m\n", + " 800 0.1631 23.2% 0.2m\n", + " 810 0.1409 23.2% 0.2m\n", + " 820 0.1543 23.2% 0.1m\n", + " 830 0.2182 23.2% 0.1m\n", + " 840 0.1900 23.2% 0.1m\n", + " 850 0.1487 23.2% 0.0m\n", + " 860 0.1688 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 9 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 9 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16827 โ†’ Val=0.16457\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16457)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 259m\n", + "\n", + "๐Ÿ”„ Epoch 10/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2100 23.2% 2.9m\n", + " 10 0.1349 23.2% 2.7m\n", + " 20 0.2062 23.2% 2.6m\n", + " 30 0.1748 23.2% 2.6m\n", + " 40 0.2162 23.2% 2.6m\n", + " 50 0.1512 23.2% 2.5m\n", + " 60 0.1605 23.2% 2.5m\n", + " 70 0.1274 23.2% 2.5m\n", + " 80 0.1513 23.2% 2.4m\n", + " 90 0.1661 23.2% 2.4m\n", + " 100 0.1788 23.2% 2.4m\n", + " 110 0.1581 23.2% 2.3m\n", + " 120 0.2010 23.2% 2.3m\n", + " 130 0.2075 23.2% 2.3m\n", + " 140 0.1675 23.2% 2.3m\n", + " 150 0.1771 23.2% 2.2m\n", + " 160 0.1508 23.2% 2.2m\n", + " 170 0.1531 23.2% 2.2m\n", + " 180 0.1819 23.2% 2.1m\n", + " 190 0.1734 23.2% 2.1m\n", + " 200 0.1534 23.2% 2.1m\n", + " 210 0.1805 23.2% 2.0m\n", + " 220 0.2114 23.2% 2.0m\n", + " 230 0.1469 23.2% 2.0m\n", + " 240 0.1779 23.2% 1.9m\n", + " 250 0.1907 23.2% 1.9m\n", + " 260 0.1590 23.2% 1.9m\n", + " 270 0.1864 23.2% 1.8m\n", + " 280 0.1716 23.2% 1.8m\n", + " 290 0.1662 23.2% 1.8m\n", + " 300 0.2090 23.2% 1.8m\n", + " 310 0.1423 23.2% 1.7m\n", + " 320 0.1638 23.2% 1.7m\n", + " 330 0.1921 23.2% 1.7m\n", + " 340 0.1450 23.2% 1.6m\n", + " 350 0.0950 23.2% 1.6m\n", + " 360 0.1588 23.2% 1.6m\n", + " 370 0.1827 23.2% 1.5m\n", + " 380 0.1529 23.2% 1.5m\n", + " 390 0.1517 23.2% 1.5m\n", + " 400 0.1330 23.2% 1.4m\n", + " 410 0.1471 23.2% 1.4m\n", + " 420 0.1544 23.2% 1.4m\n", + " 430 0.1969 23.2% 1.3m\n", + " 440 0.1895 23.2% 1.3m\n", + " 450 0.1822 23.2% 1.3m\n", + " 460 0.1488 23.2% 1.3m\n", + " 470 0.1775 23.2% 1.2m\n", + " 480 0.1643 23.2% 1.2m\n", + " 490 0.1765 23.2% 1.2m\n", + " 500 0.2084 23.2% 1.1m\n", + " 510 0.1585 23.2% 1.1m\n", + " 520 0.1718 23.2% 1.1m\n", + " 530 0.1493 23.2% 1.0m\n", + " 540 0.1809 23.2% 1.0m\n", + " 550 0.1897 23.2% 1.0m\n", + " 560 0.1753 23.2% 0.9m\n", + " 570 0.1647 23.2% 0.9m\n", + " 580 0.1941 23.2% 0.9m\n", + " 590 0.1590 23.2% 0.9m\n", + " 600 0.1255 23.2% 0.8m\n", + " 610 0.1601 23.2% 0.8m\n", + " 620 0.1754 23.2% 0.8m\n", + " 630 0.1607 23.2% 0.7m\n", + " 640 0.1518 23.2% 0.7m\n", + " 650 0.1944 23.2% 0.7m\n", + " 660 0.1893 23.2% 0.6m\n", + " 670 0.1472 23.2% 0.6m\n", + " 680 0.1599 23.2% 0.6m\n", + " 690 0.1550 23.2% 0.5m\n", + " 700 0.2074 23.2% 0.5m\n", + " 710 0.1710 23.2% 0.5m\n", + " 720 0.1407 23.2% 0.4m\n", + " 730 0.1582 23.2% 0.4m\n", + " 740 0.1903 23.2% 0.4m\n", + " 750 0.1845 23.2% 0.4m\n", + " 760 0.1785 23.2% 0.3m\n", + " 770 0.1534 23.2% 0.3m\n", + " 780 0.1850 23.2% 0.3m\n", + " 790 0.1808 23.2% 0.2m\n", + " 800 0.1526 23.2% 0.2m\n", + " 810 0.1536 23.2% 0.2m\n", + " 820 0.1552 23.2% 0.1m\n", + " 830 0.1570 23.2% 0.1m\n", + " 840 0.1532 23.2% 0.1m\n", + " 850 0.1726 23.2% 0.0m\n", + " 860 0.1579 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 10 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 10 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16809 โ†’ Val=0.16463\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.3% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 255m\n", + "\n", + "๐Ÿ”„ Epoch 11/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1571 23.2% 2.7m\n", + " 10 0.1832 23.2% 2.7m\n", + " 20 0.1595 23.2% 2.7m\n", + " 30 0.1742 23.2% 2.6m\n", + " 40 0.1385 23.2% 2.6m\n", + " 50 0.1570 23.2% 2.6m\n", + " 60 0.1732 23.2% 2.5m\n", + " 70 0.1712 23.2% 2.5m\n", + " 80 0.1352 23.2% 2.5m\n", + " 90 0.1722 23.2% 2.4m\n", + " 100 0.1781 23.2% 2.4m\n", + " 110 0.1554 23.2% 2.4m\n", + " 120 0.1675 23.2% 2.3m\n", + " 130 0.1578 23.2% 2.3m\n", + " 140 0.1515 23.2% 2.3m\n", + " 150 0.1746 23.2% 2.2m\n", + " 160 0.1826 23.2% 2.2m\n", + " 170 0.1449 23.2% 2.2m\n", + " 180 0.1769 23.2% 2.1m\n", + " 190 0.1862 23.2% 2.1m\n", + " 200 0.1338 23.2% 2.1m\n", + " 210 0.1800 23.2% 2.1m\n", + " 220 0.1755 23.2% 2.0m\n", + " 230 0.1567 23.2% 2.0m\n", + " 240 0.1785 23.2% 2.0m\n", + " 250 0.1805 23.2% 1.9m\n", + " 260 0.2086 23.2% 1.9m\n", + " 270 0.1576 23.2% 1.9m\n", + " 280 0.2110 23.2% 1.8m\n", + " 290 0.1933 23.2% 1.8m\n", + " 300 0.1869 23.2% 1.8m\n", + " 310 0.2021 23.2% 1.7m\n", + " 320 0.1810 23.2% 1.7m\n", + " 330 0.2239 23.2% 1.7m\n", + " 340 0.1622 23.2% 1.6m\n", + " 350 0.1429 23.2% 1.6m\n", + " 360 0.1276 23.2% 1.6m\n", + " 370 0.1675 23.2% 1.5m\n", + " 380 0.1540 23.2% 1.5m\n", + " 390 0.1616 23.2% 1.5m\n", + " 400 0.1523 23.2% 1.4m\n", + " 410 0.1408 23.2% 1.4m\n", + " 420 0.1746 23.2% 1.4m\n", + " 430 0.1775 23.2% 1.4m\n", + " 440 0.1835 23.2% 1.3m\n", + " 450 0.1875 23.2% 1.3m\n", + " 460 0.1665 23.2% 1.3m\n", + " 470 0.1855 23.2% 1.2m\n", + " 480 0.1686 23.2% 1.2m\n", + " 490 0.1950 23.2% 1.2m\n", + " 500 0.2058 23.2% 1.1m\n", + " 510 0.1265 23.2% 1.1m\n", + " 520 0.1567 23.2% 1.1m\n", + " 530 0.1556 23.2% 1.0m\n", + " 540 0.1939 23.2% 1.0m\n", + " 550 0.1636 23.2% 1.0m\n", + " 560 0.1621 23.2% 0.9m\n", + " 570 0.1722 23.2% 0.9m\n", + " 580 0.1622 23.2% 0.9m\n", + " 590 0.2033 23.2% 0.9m\n", + " 600 0.1561 23.2% 0.8m\n", + " 610 0.1355 23.2% 0.8m\n", + " 620 0.1950 23.2% 0.8m\n", + " 630 0.1973 23.2% 0.7m\n", + " 640 0.1441 23.2% 0.7m\n", + " 650 0.2151 23.2% 0.7m\n", + " 660 0.1776 23.2% 0.6m\n", + " 670 0.1530 23.2% 0.6m\n", + " 680 0.1591 23.2% 0.6m\n", + " 690 0.2029 23.2% 0.5m\n", + " 700 0.1803 23.2% 0.5m\n", + " 710 0.1785 23.2% 0.5m\n", + " 720 0.1845 23.2% 0.4m\n", + " 730 0.1525 23.2% 0.4m\n", + " 740 0.1814 23.2% 0.4m\n", + " 750 0.1865 23.2% 0.4m\n", + " 760 0.1612 23.2% 0.3m\n", + " 770 0.1588 23.2% 0.3m\n", + " 780 0.2003 23.2% 0.3m\n", + " 790 0.1841 23.2% 0.2m\n", + " 800 0.1485 23.2% 0.2m\n", + " 810 0.1531 23.2% 0.2m\n", + " 820 0.1871 23.2% 0.1m\n", + " 830 0.1916 23.2% 0.1m\n", + " 840 0.1659 23.2% 0.1m\n", + " 850 0.1619 23.2% 0.0m\n", + " 860 0.1805 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 11 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 11 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16840 โ†’ Val=0.16673\n", + " ๐Ÿ“Š Acc: EC=81.2% EL=75.4% EJ=86.3% Overall=81.0%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 252m\n", + "\n", + "๐Ÿ”„ Epoch 12/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1480 23.2% 2.8m\n", + " 10 0.1615 23.2% 2.6m\n", + " 20 0.1849 23.2% 2.6m\n", + " 30 0.1763 23.2% 2.6m\n", + " 40 0.1607 23.2% 2.6m\n", + " 50 0.1781 23.2% 2.5m\n", + " 60 0.1617 23.2% 2.5m\n", + " 70 0.1620 23.2% 2.5m\n", + " 80 0.1608 23.2% 2.4m\n", + " 90 0.1746 23.2% 2.4m\n", + " 100 0.1818 23.2% 2.4m\n", + " 110 0.1936 23.2% 2.3m\n", + " 120 0.1345 23.2% 2.3m\n", + " 130 0.1715 23.2% 2.3m\n", + " 140 0.1602 23.2% 2.3m\n", + " 150 0.1943 23.2% 2.2m\n", + " 160 0.1608 23.2% 2.2m\n", + " 170 0.1923 23.2% 2.2m\n", + " 180 0.1710 23.2% 2.1m\n", + " 190 0.1896 23.2% 2.1m\n", + " 200 0.1756 23.2% 2.1m\n", + " 210 0.1771 23.2% 2.0m\n", + " 220 0.1923 23.2% 2.0m\n", + " 230 0.1824 23.2% 2.0m\n", + " 240 0.1767 23.2% 1.9m\n", + " 250 0.1673 23.2% 1.9m\n", + " 260 0.2080 23.2% 1.9m\n", + " 270 0.1829 23.2% 1.8m\n", + " 280 0.1313 23.2% 1.8m\n", + " 290 0.1485 23.2% 1.8m\n", + " 300 0.1461 23.2% 1.7m\n", + " 310 0.1562 23.2% 1.7m\n", + " 320 0.1653 23.2% 1.7m\n", + " 330 0.1645 23.2% 1.7m\n", + " 340 0.1879 23.2% 1.6m\n", + " 350 0.2023 23.2% 1.6m\n", + " 360 0.1667 23.2% 1.6m\n", + " 370 0.1716 23.2% 1.5m\n", + " 380 0.1384 23.2% 1.5m\n", + " 390 0.1239 23.2% 1.5m\n", + " 400 0.1819 23.2% 1.4m\n", + " 410 0.1426 23.2% 1.4m\n", + " 420 0.1600 23.2% 1.4m\n", + " 430 0.1440 23.2% 1.3m\n", + " 440 0.1577 23.2% 1.3m\n", + " 450 0.1035 23.2% 1.3m\n", + " 460 0.1790 23.2% 1.3m\n", + " 470 0.1561 23.2% 1.2m\n", + " 480 0.1459 23.2% 1.2m\n", + " 490 0.2077 23.2% 1.2m\n", + " 500 0.1673 23.2% 1.1m\n", + " 510 0.1476 23.2% 1.1m\n", + " 520 0.1705 23.2% 1.1m\n", + " 530 0.1513 23.2% 1.0m\n", + " 540 0.1850 23.2% 1.0m\n", + " 550 0.1716 23.2% 1.0m\n", + " 560 0.1902 23.2% 0.9m\n", + " 570 0.1532 23.2% 0.9m\n", + " 580 0.1758 23.2% 0.9m\n", + " 590 0.1603 23.2% 0.8m\n", + " 600 0.1805 23.2% 0.8m\n", + " 610 0.1631 23.2% 0.8m\n", + " 620 0.1644 23.2% 0.8m\n", + " 630 0.1303 23.2% 0.7m\n", + " 640 0.1847 23.2% 0.7m\n", + " 650 0.2193 23.2% 0.7m\n", + " 660 0.1776 23.2% 0.6m\n", + " 670 0.1877 23.2% 0.6m\n", + " 680 0.2168 23.2% 0.6m\n", + " 690 0.1832 23.2% 0.5m\n", + " 700 0.2000 23.2% 0.5m\n", + " 710 0.1929 23.2% 0.5m\n", + " 720 0.1811 23.2% 0.4m\n", + " 730 0.1426 23.2% 0.4m\n", + " 740 0.1761 23.2% 0.4m\n", + " 750 0.1625 23.2% 0.4m\n", + " 760 0.1988 23.2% 0.3m\n", + " 770 0.1426 23.2% 0.3m\n", + " 780 0.1891 23.2% 0.3m\n", + " 790 0.1987 23.2% 0.2m\n", + " 800 0.1717 23.2% 0.2m\n", + " 810 0.1638 23.2% 0.2m\n", + " 820 0.2034 23.2% 0.1m\n", + " 830 0.1579 23.2% 0.1m\n", + " 840 0.1778 23.2% 0.1m\n", + " 850 0.1434 23.2% 0.0m\n", + " 860 0.1565 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 12 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 12 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16793 โ†’ Val=0.16486\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.5% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 247m\n", + "\n", + "๐Ÿ”„ Epoch 13/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1801 23.2% 2.6m\n", + " 10 0.1931 23.2% 2.7m\n", + " 20 0.1537 23.2% 2.6m\n", + " 30 0.1275 23.2% 2.6m\n", + " 40 0.1524 23.2% 2.6m\n", + " 50 0.1467 23.2% 2.5m\n", + " 60 0.1676 23.2% 2.5m\n", + " 70 0.1078 23.2% 2.5m\n", + " 80 0.2014 23.2% 2.4m\n", + " 90 0.1609 23.2% 2.4m\n", + " 100 0.1896 23.2% 2.4m\n", + " 110 0.1937 23.2% 2.3m\n", + " 120 0.1647 23.2% 2.3m\n", + " 130 0.1719 23.2% 2.3m\n", + " 140 0.1402 23.2% 2.2m\n", + " 150 0.1454 23.2% 2.2m\n", + " 160 0.1518 23.2% 2.2m\n", + " 170 0.1606 23.2% 2.1m\n", + " 180 0.1054 23.2% 2.1m\n", + " 190 0.1329 23.2% 2.1m\n", + " 200 0.1587 23.2% 2.1m\n", + " 210 0.1524 23.2% 2.0m\n", + " 220 0.1723 23.2% 2.0m\n", + " 230 0.1516 23.2% 2.0m\n", + " 240 0.1611 23.2% 1.9m\n", + " 250 0.1404 23.2% 1.9m\n", + " 260 0.1680 23.2% 1.9m\n", + " 270 0.1662 23.2% 1.8m\n", + " 280 0.1404 23.2% 1.8m\n", + " 290 0.1705 23.2% 1.8m\n", + " 300 0.1210 23.2% 1.7m\n", + " 310 0.1708 23.2% 1.7m\n", + " 320 0.1586 23.2% 1.7m\n", + " 330 0.1994 23.2% 1.7m\n", + " 340 0.1764 23.2% 1.6m\n", + " 350 0.1409 23.2% 1.6m\n", + " 360 0.1718 23.2% 1.6m\n", + " 370 0.1499 23.2% 1.5m\n", + " 380 0.1763 23.2% 1.5m\n", + " 390 0.1350 23.2% 1.5m\n", + " 400 0.1565 23.2% 1.4m\n", + " 410 0.1778 23.2% 1.4m\n", + " 420 0.1646 23.2% 1.4m\n", + " 430 0.1480 23.2% 1.3m\n", + " 440 0.1822 23.2% 1.3m\n", + " 450 0.1989 23.2% 1.3m\n", + " 460 0.1334 23.2% 1.3m\n", + " 470 0.1482 23.2% 1.2m\n", + " 480 0.1819 23.2% 1.2m\n", + " 490 0.1690 23.2% 1.2m\n", + " 500 0.1650 23.2% 1.1m\n", + " 510 0.1502 23.2% 1.1m\n", + " 520 0.1430 23.2% 1.1m\n", + " 530 0.1773 23.2% 1.0m\n", + " 540 0.1999 23.2% 1.0m\n", + " 550 0.1365 23.2% 1.0m\n", + " 560 0.1411 23.2% 0.9m\n", + " 570 0.1607 23.2% 0.9m\n", + " 580 0.1563 23.2% 0.9m\n", + " 590 0.1657 23.2% 0.8m\n", + " 600 0.1737 23.2% 0.8m\n", + " 610 0.1621 23.2% 0.8m\n", + " 620 0.1703 23.2% 0.8m\n", + " 630 0.1709 23.2% 0.7m\n", + " 640 0.1561 23.2% 0.7m\n", + " 650 0.1884 23.2% 0.7m\n", + " 660 0.1860 23.2% 0.6m\n", + " 670 0.1605 23.2% 0.6m\n", + " 680 0.1496 23.2% 0.6m\n", + " 690 0.1248 23.2% 0.5m\n", + " 700 0.1470 23.2% 0.5m\n", + " 710 0.1412 23.2% 0.5m\n", + " 720 0.1708 23.2% 0.4m\n", + " 730 0.1634 23.2% 0.4m\n", + " 740 0.1408 23.2% 0.4m\n", + " 750 0.1837 23.2% 0.4m\n", + " 760 0.1499 23.2% 0.3m\n", + " 770 0.1384 23.2% 0.3m\n", + " 780 0.1911 23.2% 0.3m\n", + " 790 0.1668 23.2% 0.2m\n", + " 800 0.1420 23.2% 0.2m\n", + " 810 0.1805 23.2% 0.2m\n", + " 820 0.1515 23.2% 0.1m\n", + " 830 0.1977 23.2% 0.1m\n", + " 840 0.1726 23.2% 0.1m\n", + " 850 0.1769 23.2% 0.0m\n", + " 860 0.1891 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 13 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 13 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16810 โ†’ Val=0.16447\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16447)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 245m\n", + "\n", + "๐Ÿ”„ Epoch 14/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1733 23.2% 2.6m\n", + " 10 0.1724 23.2% 2.7m\n", + " 20 0.1398 23.2% 2.6m\n", + " 30 0.1736 23.2% 2.6m\n", + " 40 0.1610 23.2% 2.6m\n", + " 50 0.1770 23.2% 2.6m\n", + " 60 0.1334 23.2% 2.5m\n", + " 70 0.1495 23.2% 2.5m\n", + " 80 0.1680 23.2% 2.5m\n", + " 90 0.1669 23.2% 2.4m\n", + " 100 0.1441 23.2% 2.4m\n", + " 110 0.1703 23.2% 2.4m\n", + " 120 0.1478 23.2% 2.3m\n", + " 130 0.1643 23.2% 2.3m\n", + " 140 0.2238 23.2% 2.3m\n", + " 150 0.1503 23.2% 2.2m\n", + " 160 0.1215 23.2% 2.2m\n", + " 170 0.1278 23.2% 2.2m\n", + " 180 0.1561 23.2% 2.1m\n", + " 190 0.1627 23.2% 2.1m\n", + " 200 0.1763 23.2% 2.1m\n", + " 210 0.1783 23.2% 2.0m\n", + " 220 0.1455 23.2% 2.0m\n", + " 230 0.1454 23.2% 2.0m\n", + " 240 0.1599 23.2% 1.9m\n", + " 250 0.1799 23.2% 1.9m\n", + " 260 0.1976 23.2% 1.9m\n", + " 270 0.1276 23.2% 1.9m\n", + " 280 0.1492 23.2% 1.8m\n", + " 290 0.1493 23.2% 1.8m\n", + " 300 0.1809 23.2% 1.8m\n", + " 310 0.1699 23.2% 1.7m\n", + " 320 0.2242 23.2% 1.7m\n", + " 330 0.2003 23.2% 1.7m\n", + " 340 0.1416 23.2% 1.6m\n", + " 350 0.1788 23.2% 1.6m\n", + " 360 0.1925 23.2% 1.6m\n", + " 370 0.1572 23.2% 1.5m\n", + " 380 0.1598 23.2% 1.5m\n", + " 390 0.2008 23.2% 1.5m\n", + " 400 0.1622 23.2% 1.4m\n", + " 410 0.1560 23.2% 1.4m\n", + " 420 0.1694 23.2% 1.4m\n", + " 430 0.1380 23.2% 1.4m\n", + " 440 0.1747 23.2% 1.3m\n", + " 450 0.1694 23.2% 1.3m\n", + " 460 0.1806 23.2% 1.3m\n", + " 470 0.1583 23.2% 1.2m\n", + " 480 0.1991 23.2% 1.2m\n", + " 490 0.1962 23.2% 1.2m\n", + " 500 0.1558 23.2% 1.1m\n", + " 510 0.1570 23.2% 1.1m\n", + " 520 0.1633 23.2% 1.1m\n", + " 530 0.1446 23.2% 1.0m\n", + " 540 0.1640 23.2% 1.0m\n", + " 550 0.1889 23.2% 1.0m\n", + " 560 0.2044 23.2% 0.9m\n", + " 570 0.1762 23.2% 0.9m\n", + " 580 0.1690 23.2% 0.9m\n", + " 590 0.1867 23.2% 0.9m\n", + " 600 0.1599 23.2% 0.8m\n", + " 610 0.1622 23.2% 0.8m\n", + " 620 0.1693 23.2% 0.8m\n", + " 630 0.1770 23.2% 0.7m\n", + " 640 0.1684 23.2% 0.7m\n", + " 650 0.1568 23.2% 0.7m\n", + " 660 0.1415 23.2% 0.6m\n", + " 670 0.1354 23.2% 0.6m\n", + " 680 0.1753 23.2% 0.6m\n", + " 690 0.2086 23.2% 0.5m\n", + " 700 0.1682 23.2% 0.5m\n", + " 710 0.1756 23.2% 0.5m\n", + " 720 0.1393 23.2% 0.4m\n", + " 730 0.1768 23.2% 0.4m\n", + " 740 0.1869 23.2% 0.4m\n", + " 750 0.1791 23.2% 0.4m\n", + " 760 0.1677 23.2% 0.3m\n", + " 770 0.1552 23.2% 0.3m\n", + " 780 0.1536 23.2% 0.3m\n", + " 790 0.1443 23.2% 0.2m\n", + " 800 0.1571 23.2% 0.2m\n", + " 810 0.1748 23.2% 0.2m\n", + " 820 0.1520 23.2% 0.1m\n", + " 830 0.1710 23.2% 0.1m\n", + " 840 0.1524 23.2% 0.1m\n", + " 850 0.1872 23.2% 0.0m\n", + " 860 0.1801 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 14 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 14 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16764 โ†’ Val=0.16430\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16430)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 243m\n", + "\n", + "๐Ÿ”„ Epoch 15/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1213 23.2% 2.7m\n", + " 10 0.1947 23.2% 2.7m\n", + " 20 0.1757 23.2% 2.6m\n", + " 30 0.1984 23.2% 2.6m\n", + " 40 0.1921 23.2% 2.6m\n", + " 50 0.1507 23.2% 2.5m\n", + " 60 0.1914 23.2% 2.5m\n", + " 70 0.1584 23.2% 2.5m\n", + " 80 0.1516 23.2% 2.4m\n", + " 90 0.1551 23.2% 2.4m\n", + " 100 0.1837 23.2% 2.4m\n", + " 110 0.1469 23.2% 2.3m\n", + " 120 0.1775 23.2% 2.3m\n", + " 130 0.1803 23.2% 2.3m\n", + " 140 0.1901 23.2% 2.3m\n", + " 150 0.2057 23.2% 2.2m\n", + " 160 0.1445 23.2% 2.2m\n", + " 170 0.1793 23.2% 2.2m\n", + " 180 0.1749 23.2% 2.1m\n", + " 190 0.1558 23.2% 2.1m\n", + " 200 0.1989 23.2% 2.1m\n", + " 210 0.1324 23.2% 2.0m\n", + " 220 0.1846 23.2% 2.0m\n", + " 230 0.1780 23.2% 2.0m\n", + " 240 0.1510 23.2% 1.9m\n", + " 250 0.1597 23.2% 1.9m\n", + " 260 0.1376 23.2% 1.9m\n", + " 270 0.1605 23.2% 1.8m\n", + " 280 0.1626 23.2% 1.8m\n", + " 290 0.1551 23.2% 1.8m\n", + " 300 0.1498 23.2% 1.8m\n", + " 310 0.1539 23.2% 1.7m\n", + " 320 0.1848 23.2% 1.7m\n", + " 330 0.1837 23.2% 1.7m\n", + " 340 0.1522 23.2% 1.6m\n", + " 350 0.1691 23.2% 1.6m\n", + " 360 0.1562 23.2% 1.6m\n", + " 370 0.1549 23.2% 1.5m\n", + " 380 0.1071 23.2% 1.5m\n", + " 390 0.1609 23.2% 1.5m\n", + " 400 0.2057 23.2% 1.4m\n", + " 410 0.1887 23.2% 1.4m\n", + " 420 0.1548 23.2% 1.4m\n", + " 430 0.1383 23.2% 1.3m\n", + " 440 0.1718 23.2% 1.3m\n", + " 450 0.1534 23.2% 1.3m\n", + " 460 0.1666 23.2% 1.3m\n", + " 470 0.1629 23.2% 1.2m\n", + " 480 0.1465 23.2% 1.2m\n", + " 490 0.1593 23.2% 1.2m\n", + " 500 0.1839 23.2% 1.1m\n", + " 510 0.1772 23.2% 1.1m\n", + " 520 0.2088 23.2% 1.1m\n", + " 530 0.1726 23.2% 1.0m\n", + " 540 0.1767 23.2% 1.0m\n", + " 550 0.1969 23.2% 1.0m\n", + " 560 0.1572 23.2% 0.9m\n", + " 570 0.1444 23.2% 0.9m\n", + " 580 0.1546 23.2% 0.9m\n", + " 590 0.1811 23.2% 0.9m\n", + " 600 0.1604 23.2% 0.8m\n", + " 610 0.1600 23.2% 0.8m\n", + " 620 0.1491 23.2% 0.8m\n", + " 630 0.1681 23.2% 0.7m\n", + " 640 0.1756 23.2% 0.7m\n", + " 650 0.1672 23.2% 0.7m\n", + " 660 0.1422 23.2% 0.6m\n", + " 670 0.2159 23.2% 0.6m\n", + " 680 0.1718 23.2% 0.6m\n", + " 690 0.1756 23.2% 0.5m\n", + " 700 0.1609 23.2% 0.5m\n", + " 710 0.1828 23.2% 0.5m\n", + " 720 0.1629 23.2% 0.4m\n", + " 730 0.1768 23.2% 0.4m\n", + " 740 0.1449 23.2% 0.4m\n", + " 750 0.1862 23.2% 0.4m\n", + " 760 0.1671 23.2% 0.3m\n", + " 770 0.1656 23.2% 0.3m\n", + " 780 0.1584 23.2% 0.3m\n", + " 790 0.1804 23.2% 0.2m\n", + " 800 0.1987 23.2% 0.2m\n", + " 810 0.1659 23.2% 0.2m\n", + " 820 0.1254 23.2% 0.1m\n", + " 830 0.1781 23.2% 0.1m\n", + " 840 0.1961 23.2% 0.1m\n", + " 850 0.1777 23.2% 0.0m\n", + " 860 0.1418 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 15 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 15 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16752 โ†’ Val=0.16423\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16423)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 240m\n", + "\n", + "๐Ÿ”„ Epoch 16/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1776 23.2% 2.7m\n", + " 10 0.1261 23.2% 2.7m\n", + " 20 0.1510 23.2% 2.7m\n", + " 30 0.1475 23.2% 2.6m\n", + " 40 0.1164 23.2% 2.6m\n", + " 50 0.1710 23.2% 2.5m\n", + " 60 0.1978 23.2% 2.5m\n", + " 70 0.1741 23.2% 2.5m\n", + " 80 0.1906 23.2% 2.4m\n", + " 90 0.1674 23.2% 2.4m\n", + " 100 0.1389 23.2% 2.4m\n", + " 110 0.1707 23.2% 2.4m\n", + " 120 0.1531 23.2% 2.3m\n", + " 130 0.1621 23.2% 2.3m\n", + " 140 0.1780 23.2% 2.3m\n", + " 150 0.1696 23.2% 2.2m\n", + " 160 0.1955 23.2% 2.2m\n", + " 170 0.1680 23.2% 2.2m\n", + " 180 0.2014 23.2% 2.1m\n", + " 190 0.1442 23.2% 2.1m\n", + " 200 0.1769 23.2% 2.1m\n", + " 210 0.1997 23.2% 2.0m\n", + " 220 0.1674 23.2% 2.0m\n", + " 230 0.1731 23.2% 2.0m\n", + " 240 0.1755 23.2% 1.9m\n", + " 250 0.1701 23.2% 1.9m\n", + " 260 0.1159 23.2% 1.9m\n", + " 270 0.1705 23.2% 1.8m\n", + " 280 0.1592 23.2% 1.8m\n", + " 290 0.1964 23.2% 1.8m\n", + " 300 0.1694 23.2% 1.8m\n", + " 310 0.1599 23.2% 1.7m\n", + " 320 0.1328 23.2% 1.7m\n", + " 330 0.1384 23.2% 1.7m\n", + " 340 0.1671 23.2% 1.6m\n", + " 350 0.1473 23.2% 1.6m\n", + " 360 0.1627 23.2% 1.6m\n", + " 370 0.1812 23.2% 1.5m\n", + " 380 0.1757 23.2% 1.5m\n", + " 390 0.1306 23.2% 1.5m\n", + " 400 0.1712 23.2% 1.4m\n", + " 410 0.2058 23.2% 1.4m\n", + " 420 0.1815 23.2% 1.4m\n", + " 430 0.1758 23.2% 1.3m\n", + " 440 0.1511 23.2% 1.3m\n", + " 450 0.1906 23.2% 1.3m\n", + " 460 0.1660 23.2% 1.3m\n", + " 470 0.1550 23.2% 1.2m\n", + " 480 0.1689 23.2% 1.2m\n", + " 490 0.1517 23.2% 1.2m\n", + " 500 0.1855 23.2% 1.1m\n", + " 510 0.1939 23.2% 1.1m\n", + " 520 0.1879 23.2% 1.1m\n", + " 530 0.2014 23.2% 1.0m\n", + " 540 0.1498 23.2% 1.0m\n", + " 550 0.1892 23.2% 1.0m\n", + " 560 0.1700 23.2% 0.9m\n", + " 570 0.1746 23.2% 0.9m\n", + " 580 0.1622 23.2% 0.9m\n", + " 590 0.2206 23.2% 0.9m\n", + " 600 0.1588 23.2% 0.8m\n", + " 610 0.1644 23.2% 0.8m\n", + " 620 0.1521 23.2% 0.8m\n", + " 630 0.1796 23.2% 0.7m\n", + " 640 0.1733 23.2% 0.7m\n", + " 650 0.1863 23.2% 0.7m\n", + " 660 0.1684 23.2% 0.6m\n", + " 670 0.1585 23.2% 0.6m\n", + " 680 0.1540 23.2% 0.6m\n", + " 690 0.1837 23.2% 0.5m\n", + " 700 0.1908 23.2% 0.5m\n", + " 710 0.1924 23.2% 0.5m\n", + " 720 0.1828 23.2% 0.4m\n", + " 730 0.1615 23.2% 0.4m\n", + " 740 0.1556 23.2% 0.4m\n", + " 750 0.1587 23.2% 0.4m\n", + " 760 0.1349 23.2% 0.3m\n", + " 770 0.1602 23.2% 0.3m\n", + " 780 0.1699 23.2% 0.3m\n", + " 790 0.2097 23.2% 0.2m\n", + " 800 0.1676 23.2% 0.2m\n", + " 810 0.1674 23.2% 0.2m\n", + " 820 0.1839 23.2% 0.1m\n", + " 830 0.1960 23.2% 0.1m\n", + " 840 0.2016 23.2% 0.1m\n", + " 850 0.1541 23.2% 0.0m\n", + " 860 0.1492 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 16 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 16 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16754 โ†’ Val=0.16426\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 237m\n", + "\n", + "๐Ÿ”„ Epoch 17/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2191 23.2% 2.6m\n", + " 10 0.1685 23.2% 2.6m\n", + " 20 0.1986 23.2% 2.6m\n", + " 30 0.1550 23.2% 2.6m\n", + " 40 0.1601 23.2% 2.5m\n", + " 50 0.1753 23.2% 2.5m\n", + " 60 0.2003 23.2% 2.5m\n", + " 70 0.1954 23.2% 2.5m\n", + " 80 0.1245 23.2% 2.4m\n", + " 90 0.1686 23.2% 2.4m\n", + " 100 0.1634 23.2% 2.4m\n", + " 110 0.1892 23.2% 2.3m\n", + " 120 0.1576 23.2% 2.3m\n", + " 130 0.1707 23.2% 2.3m\n", + " 140 0.1788 23.2% 2.2m\n", + " 150 0.1457 23.2% 2.2m\n", + " 160 0.1750 23.2% 2.2m\n", + " 170 0.1746 23.2% 2.2m\n", + " 180 0.1807 23.2% 2.1m\n", + " 190 0.1521 23.2% 2.1m\n", + " 200 0.1712 23.2% 2.1m\n", + " 210 0.1661 23.2% 2.0m\n", + " 220 0.1531 23.2% 2.0m\n", + " 230 0.1657 23.2% 2.0m\n", + " 240 0.2026 23.2% 1.9m\n", + " 250 0.1883 23.2% 1.9m\n", + " 260 0.1462 23.2% 1.9m\n", + " 270 0.1477 23.2% 1.8m\n", + " 280 0.1276 23.2% 1.8m\n", + " 290 0.1808 23.2% 1.8m\n", + " 300 0.1874 23.2% 1.7m\n", + " 310 0.1709 23.2% 1.7m\n", + " 320 0.1667 23.2% 1.7m\n", + " 330 0.1972 23.2% 1.7m\n", + " 340 0.1405 23.2% 1.6m\n", + " 350 0.1795 23.2% 1.6m\n", + " 360 0.1563 23.2% 1.6m\n", + " 370 0.1779 23.2% 1.5m\n", + " 380 0.2144 23.2% 1.5m\n", + " 390 0.1508 23.2% 1.5m\n", + " 400 0.1802 23.2% 1.4m\n", + " 410 0.1820 23.2% 1.4m\n", + " 420 0.1570 23.2% 1.4m\n", + " 430 0.1452 23.2% 1.3m\n", + " 440 0.1471 23.2% 1.3m\n", + " 450 0.2202 23.2% 1.3m\n", + " 460 0.1642 23.2% 1.3m\n", + " 470 0.1478 23.2% 1.2m\n", + " 480 0.1773 23.2% 1.2m\n", + " 490 0.1531 23.2% 1.2m\n", + " 500 0.1295 23.2% 1.1m\n", + " 510 0.1641 23.2% 1.1m\n", + " 520 0.1757 23.2% 1.1m\n", + " 530 0.1528 23.2% 1.0m\n", + " 540 0.1971 23.2% 1.0m\n", + " 550 0.1881 23.2% 1.0m\n", + " 560 0.1404 23.2% 0.9m\n", + " 570 0.1977 23.2% 0.9m\n", + " 580 0.1866 23.2% 0.9m\n", + " 590 0.1585 23.2% 0.8m\n", + " 600 0.1486 23.2% 0.8m\n", + " 610 0.1709 23.2% 0.8m\n", + " 620 0.1657 23.2% 0.8m\n", + " 630 0.1673 23.2% 0.7m\n", + " 640 0.1504 23.2% 0.7m\n", + " 650 0.1854 23.2% 0.7m\n", + " 660 0.1770 23.2% 0.6m\n", + " 670 0.1978 23.2% 0.6m\n", + " 680 0.1546 23.2% 0.6m\n", + " 690 0.1756 23.2% 0.5m\n", + " 700 0.1741 23.2% 0.5m\n", + " 710 0.1592 23.2% 0.5m\n", + " 720 0.1858 23.2% 0.4m\n", + " 730 0.2055 23.2% 0.4m\n", + " 740 0.1683 23.2% 0.4m\n", + " 750 0.1456 23.2% 0.4m\n", + " 760 0.1747 23.2% 0.3m\n", + " 770 0.1249 23.2% 0.3m\n", + " 780 0.1585 23.2% 0.3m\n", + " 790 0.1785 23.2% 0.2m\n", + " 800 0.1570 23.2% 0.2m\n", + " 810 0.1632 23.2% 0.2m\n", + " 820 0.1182 23.2% 0.1m\n", + " 830 0.1586 23.2% 0.1m\n", + " 840 0.1862 23.2% 0.1m\n", + " 850 0.1738 23.2% 0.0m\n", + " 860 0.1635 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 17 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 17 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16768 โ†’ Val=0.16528\n", + " ๐Ÿ“Š Acc: EC=81.3% EL=75.4% EJ=86.7% Overall=81.1%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 233m\n", + "\n", + "๐Ÿ”„ Epoch 18/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1556 23.2% 2.6m\n", + " 10 0.2086 23.2% 2.7m\n", + " 20 0.2199 23.2% 2.7m\n", + " 30 0.1616 23.2% 2.6m\n", + " 40 0.1735 23.2% 2.6m\n", + " 50 0.2002 23.2% 2.6m\n", + " 60 0.1746 23.2% 2.5m\n", + " 70 0.1813 23.2% 2.5m\n", + " 80 0.1479 23.2% 2.5m\n", + " 90 0.1908 23.2% 2.4m\n", + " 100 0.1361 23.2% 2.4m\n", + " 110 0.1783 23.2% 2.4m\n", + " 120 0.1663 23.2% 2.3m\n", + " 130 0.1577 23.2% 2.3m\n", + " 140 0.2038 23.2% 2.3m\n", + " 150 0.1555 23.2% 2.2m\n", + " 160 0.1623 23.2% 2.2m\n", + " 170 0.1809 23.2% 2.2m\n", + " 180 0.2173 23.2% 2.1m\n", + " 190 0.1947 23.2% 2.1m\n", + " 200 0.1635 23.2% 2.1m\n", + " 210 0.1502 23.2% 2.0m\n", + " 220 0.1931 23.2% 2.0m\n", + " 230 0.1468 23.2% 2.0m\n", + " 240 0.1722 23.2% 1.9m\n", + " 250 0.1668 23.2% 1.9m\n", + " 260 0.1948 23.2% 1.9m\n", + " 270 0.1674 23.2% 1.9m\n", + " 280 0.1480 23.2% 1.8m\n", + " 290 0.1748 23.2% 1.8m\n", + " 300 0.1523 23.2% 1.8m\n", + " 310 0.1721 23.2% 1.7m\n", + " 320 0.1629 23.2% 1.7m\n", + " 330 0.1657 23.2% 1.7m\n", + " 340 0.1727 23.2% 1.6m\n", + " 350 0.1495 23.2% 1.6m\n", + " 360 0.1601 23.2% 1.6m\n", + " 370 0.1844 23.2% 1.5m\n", + " 380 0.1742 23.2% 1.5m\n", + " 390 0.1525 23.2% 1.5m\n", + " 400 0.1360 23.2% 1.4m\n", + " 410 0.1428 23.2% 1.4m\n", + " 420 0.1691 23.2% 1.4m\n", + " 430 0.1559 23.2% 1.4m\n", + " 440 0.1247 23.2% 1.3m\n", + " 450 0.1430 23.2% 1.3m\n", + " 460 0.1419 23.2% 1.3m\n", + " 470 0.1344 23.2% 1.2m\n", + " 480 0.1531 23.2% 1.2m\n", + " 490 0.1783 23.2% 1.2m\n", + " 500 0.1528 23.2% 1.1m\n", + " 510 0.1754 23.2% 1.1m\n", + " 520 0.1800 23.2% 1.1m\n", + " 530 0.1804 23.2% 1.0m\n", + " 540 0.1234 23.2% 1.0m\n", + " 550 0.1850 23.2% 1.0m\n", + " 560 0.1667 23.2% 0.9m\n", + " 570 0.1511 23.2% 0.9m\n", + " 580 0.1898 23.2% 0.9m\n", + " 590 0.1857 23.2% 0.9m\n", + " 600 0.2043 23.2% 0.8m\n", + " 610 0.1481 23.2% 0.8m\n", + " 620 0.1471 23.2% 0.8m\n", + " 630 0.1580 23.2% 0.7m\n", + " 640 0.1520 23.2% 0.7m\n", + " 650 0.2278 23.2% 0.7m\n", + " 660 0.2052 23.2% 0.6m\n", + " 670 0.1643 23.2% 0.6m\n", + " 680 0.1951 23.2% 0.6m\n", + " 690 0.1290 23.2% 0.5m\n", + " 700 0.1762 23.2% 0.5m\n", + " 710 0.1820 23.2% 0.5m\n", + " 720 0.1501 23.2% 0.4m\n", + " 730 0.1376 23.2% 0.4m\n", + " 740 0.1801 23.2% 0.4m\n", + " 750 0.1736 23.2% 0.4m\n", + " 760 0.1582 23.2% 0.3m\n", + " 770 0.1787 23.2% 0.3m\n", + " 780 0.1829 23.2% 0.3m\n", + " 790 0.1686 23.2% 0.2m\n", + " 800 0.1653 23.2% 0.2m\n", + " 810 0.1351 23.2% 0.2m\n", + " 820 0.2046 23.2% 0.1m\n", + " 830 0.1689 23.2% 0.1m\n", + " 840 0.1588 23.2% 0.1m\n", + " 850 0.1516 23.2% 0.0m\n", + " 860 0.1914 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 18 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 18 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16755 โ†’ Val=0.16436\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 231m\n", + "\n", + "๐Ÿ”„ Epoch 19/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1525 23.2% 2.7m\n", + " 10 0.1669 23.2% 2.7m\n", + " 20 0.1487 23.2% 2.6m\n", + " 30 0.2084 23.2% 2.6m\n", + " 40 0.1779 23.2% 2.6m\n", + " 50 0.1448 23.2% 2.5m\n", + " 60 0.1551 23.2% 2.5m\n", + " 70 0.1155 23.2% 2.5m\n", + " 80 0.1786 23.2% 2.5m\n", + " 90 0.1701 23.2% 2.4m\n", + " 100 0.1781 23.2% 2.4m\n", + " 110 0.1905 23.2% 2.4m\n", + " 120 0.1899 23.2% 2.3m\n", + " 130 0.1452 23.2% 2.3m\n", + " 140 0.1638 23.2% 2.3m\n", + " 150 0.1753 23.2% 2.2m\n", + " 160 0.1793 23.2% 2.2m\n", + " 170 0.1391 23.2% 2.2m\n", + " 180 0.1709 23.2% 2.1m\n", + " 190 0.1890 23.2% 2.1m\n", + " 200 0.1770 23.2% 2.1m\n", + " 210 0.1568 23.2% 2.0m\n", + " 220 0.1622 23.2% 2.0m\n", + " 230 0.1839 23.2% 2.0m\n", + " 240 0.2063 23.2% 1.9m\n", + " 250 0.1529 23.2% 1.9m\n", + " 260 0.1691 23.2% 1.9m\n", + " 270 0.1925 23.2% 1.9m\n", + " 280 0.1539 23.2% 1.8m\n", + " 290 0.2082 23.2% 1.8m\n", + " 300 0.1580 23.2% 1.8m\n", + " 310 0.1623 23.2% 1.7m\n", + " 320 0.1522 23.2% 1.7m\n", + " 330 0.1741 23.2% 1.7m\n", + " 340 0.1649 23.2% 1.6m\n", + " 350 0.2204 23.2% 1.6m\n", + " 360 0.1360 23.2% 1.6m\n", + " 370 0.2045 23.2% 1.5m\n", + " 380 0.1788 23.2% 1.5m\n", + " 390 0.1343 23.2% 1.5m\n", + " 400 0.1725 23.2% 1.4m\n", + " 410 0.1473 23.2% 1.4m\n", + " 420 0.1294 23.2% 1.4m\n", + " 430 0.1572 23.2% 1.3m\n", + " 440 0.1814 23.2% 1.3m\n", + " 450 0.1708 23.2% 1.3m\n", + " 460 0.1359 23.2% 1.3m\n", + " 470 0.1412 23.2% 1.2m\n", + " 480 0.1690 23.2% 1.2m\n", + " 490 0.1549 23.2% 1.2m\n", + " 500 0.1536 23.2% 1.1m\n", + " 510 0.1588 23.2% 1.1m\n", + " 520 0.1629 23.2% 1.1m\n", + " 530 0.1695 23.2% 1.0m\n", + " 540 0.1752 23.2% 1.0m\n", + " 550 0.1521 23.2% 1.0m\n", + " 560 0.1540 23.2% 0.9m\n", + " 570 0.1468 23.2% 0.9m\n", + " 580 0.1832 23.2% 0.9m\n", + " 590 0.1674 23.2% 0.9m\n", + " 600 0.1630 23.2% 0.8m\n", + " 610 0.1653 23.2% 0.8m\n", + " 620 0.1804 23.2% 0.8m\n", + " 630 0.1163 23.2% 0.7m\n", + " 640 0.1500 23.2% 0.7m\n", + " 650 0.1680 23.2% 0.7m\n", + " 660 0.1738 23.2% 0.6m\n", + " 670 0.1870 23.2% 0.6m\n", + " 680 0.1686 23.2% 0.6m\n", + " 690 0.1811 23.2% 0.5m\n", + " 700 0.1608 23.2% 0.5m\n", + " 710 0.1496 23.2% 0.5m\n", + " 720 0.1472 23.2% 0.4m\n", + " 730 0.1814 23.2% 0.4m\n", + " 740 0.1814 23.2% 0.4m\n", + " 750 0.1709 23.2% 0.4m\n", + " 760 0.2073 23.2% 0.3m\n", + " 770 0.1674 23.2% 0.3m\n", + " 780 0.2093 23.2% 0.3m\n", + " 790 0.1651 23.2% 0.2m\n", + " 800 0.1930 23.2% 0.2m\n", + " 810 0.1831 23.2% 0.2m\n", + " 820 0.1599 23.2% 0.1m\n", + " 830 0.2002 23.2% 0.1m\n", + " 840 0.1963 23.2% 0.1m\n", + " 850 0.1563 23.2% 0.0m\n", + " 860 0.1838 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 19 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 19 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16737 โ†’ Val=0.16442\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 229m\n", + "\n", + "๐Ÿ”„ Epoch 20/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1511 23.2% 2.6m\n", + " 10 0.1285 23.2% 2.7m\n", + " 20 0.1480 23.2% 2.6m\n", + " 30 0.1647 23.2% 2.6m\n", + " 40 0.1657 23.2% 2.6m\n", + " 50 0.1629 23.2% 2.5m\n", + " 60 0.1700 23.2% 2.5m\n", + " 70 0.1698 23.2% 2.5m\n", + " 80 0.1449 23.2% 2.4m\n", + " 90 0.2170 23.2% 2.4m\n", + " 100 0.1437 23.2% 2.4m\n", + " 110 0.1927 23.2% 2.3m\n", + " 120 0.1913 23.2% 2.3m\n", + " 130 0.1396 23.2% 2.3m\n", + " 140 0.1464 23.2% 2.3m\n", + " 150 0.1760 23.2% 2.2m\n", + " 160 0.1698 23.2% 2.2m\n", + " 170 0.1546 23.2% 2.2m\n", + " 180 0.1615 23.2% 2.1m\n", + " 190 0.1306 23.2% 2.1m\n", + " 200 0.1766 23.2% 2.1m\n", + " 210 0.1558 23.2% 2.0m\n", + " 220 0.1632 23.2% 2.0m\n", + " 230 0.1477 23.2% 2.0m\n", + " 240 0.1687 23.2% 1.9m\n", + " 250 0.2011 23.2% 1.9m\n", + " 260 0.1468 23.2% 1.9m\n", + " 270 0.1639 23.2% 1.8m\n", + " 280 0.1813 23.2% 1.8m\n", + " 290 0.1972 23.2% 1.8m\n", + " 300 0.1576 23.2% 1.8m\n", + " 310 0.1574 23.2% 1.7m\n", + " 320 0.1283 23.2% 1.7m\n", + " 330 0.1732 23.2% 1.7m\n", + " 340 0.1672 23.2% 1.6m\n", + " 350 0.1515 23.2% 1.6m\n", + " 360 0.1776 23.2% 1.6m\n", + " 370 0.1756 23.2% 1.5m\n", + " 380 0.1773 23.2% 1.5m\n", + " 390 0.1707 23.2% 1.5m\n", + " 400 0.1837 23.2% 1.4m\n", + " 410 0.1277 23.2% 1.4m\n", + " 420 0.1381 23.2% 1.4m\n", + " 430 0.1470 23.2% 1.4m\n", + " 440 0.1599 23.2% 1.3m\n", + " 450 0.1852 23.2% 1.3m\n", + " 460 0.1164 23.2% 1.3m\n", + " 470 0.1671 23.2% 1.2m\n", + " 480 0.1809 23.2% 1.2m\n", + " 490 0.1958 23.2% 1.2m\n", + " 500 0.1926 23.2% 1.1m\n", + " 510 0.1399 23.2% 1.1m\n", + " 520 0.1690 23.2% 1.1m\n", + " 530 0.2006 23.2% 1.0m\n", + " 540 0.1621 23.2% 1.0m\n", + " 550 0.1616 23.2% 1.0m\n", + " 560 0.1327 23.2% 0.9m\n", + " 570 0.1516 23.2% 0.9m\n", + " 580 0.1563 23.2% 0.9m\n", + " 590 0.1446 23.2% 0.9m\n", + " 600 0.1919 23.2% 0.8m\n", + " 610 0.1924 23.2% 0.8m\n", + " 620 0.1408 23.2% 0.8m\n", + " 630 0.1471 23.2% 0.7m\n", + " 640 0.1742 23.2% 0.7m\n", + " 650 0.1828 23.2% 0.7m\n", + " 660 0.1558 23.2% 0.6m\n", + " 670 0.1543 23.2% 0.6m\n", + " 680 0.1945 23.2% 0.6m\n", + " 690 0.1383 23.2% 0.5m\n", + " 700 0.1803 23.2% 0.5m\n", + " 710 0.1310 23.2% 0.5m\n", + " 720 0.1815 23.2% 0.4m\n", + " 730 0.1534 23.2% 0.4m\n", + " 740 0.1624 23.2% 0.4m\n", + " 750 0.1751 23.2% 0.4m\n", + " 760 0.1539 23.2% 0.3m\n", + " 770 0.1669 23.2% 0.3m\n", + " 780 0.1287 23.2% 0.3m\n", + " 790 0.1659 23.2% 0.2m\n", + " 800 0.1937 23.2% 0.2m\n", + " 810 0.1194 23.2% 0.2m\n", + " 820 0.2056 23.2% 0.1m\n", + " 830 0.1616 23.2% 0.1m\n", + " 840 0.1595 23.2% 0.1m\n", + " 850 0.1752 23.2% 0.0m\n", + " 860 0.1710 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 20 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 20 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16757 โ†’ Val=0.16412\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16412)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 226m\n", + "\n", + "๐Ÿ”„ Epoch 21/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1599 23.2% 2.7m\n", + " 10 0.1490 23.2% 2.7m\n", + " 20 0.1689 23.2% 2.7m\n", + " 30 0.1763 23.2% 2.6m\n", + " 40 0.1533 23.2% 2.6m\n", + " 50 0.1515 23.2% 2.6m\n", + " 60 0.1696 23.2% 2.5m\n", + " 70 0.1582 23.2% 2.5m\n", + " 80 0.1904 23.2% 2.5m\n", + " 90 0.1441 23.2% 2.5m\n", + " 100 0.1853 23.2% 2.4m\n", + " 110 0.1779 23.2% 2.4m\n", + " 120 0.1424 23.2% 2.3m\n", + " 130 0.1539 23.2% 2.3m\n", + " 140 0.1637 23.2% 2.3m\n", + " 150 0.1649 23.2% 2.2m\n", + " 160 0.1907 23.2% 2.2m\n", + " 170 0.1772 23.2% 2.2m\n", + " 180 0.1556 23.2% 2.1m\n", + " 190 0.1973 23.2% 2.1m\n", + " 200 0.1733 23.2% 2.1m\n", + " 210 0.1659 23.2% 2.0m\n", + " 220 0.1682 23.2% 2.0m\n", + " 230 0.1894 23.2% 2.0m\n", + " 240 0.1576 23.2% 2.0m\n", + " 250 0.1793 23.2% 1.9m\n", + " 260 0.1697 23.2% 1.9m\n", + " 270 0.1382 23.2% 1.9m\n", + " 280 0.1671 23.2% 1.8m\n", + " 290 0.1400 23.2% 1.8m\n", + " 300 0.1771 23.2% 1.8m\n", + " 310 0.1634 23.2% 1.7m\n", + " 320 0.1617 23.2% 1.7m\n", + " 330 0.1712 23.2% 1.7m\n", + " 340 0.1377 23.2% 1.6m\n", + " 350 0.1479 23.2% 1.6m\n", + " 360 0.1531 23.2% 1.6m\n", + " 370 0.1889 23.2% 1.5m\n", + " 380 0.1648 23.2% 1.5m\n", + " 390 0.1649 23.2% 1.5m\n", + " 400 0.1446 23.2% 1.4m\n", + " 410 0.1573 23.2% 1.4m\n", + " 420 0.2065 23.2% 1.4m\n", + " 430 0.1821 23.2% 1.3m\n", + " 440 0.1365 23.2% 1.3m\n", + " 450 0.1698 23.2% 1.3m\n", + " 460 0.1368 23.2% 1.3m\n", + " 470 0.1323 23.2% 1.2m\n", + " 480 0.1731 23.2% 1.2m\n", + " 490 0.1769 23.2% 1.2m\n", + " 500 0.1335 23.2% 1.1m\n", + " 510 0.2031 23.2% 1.1m\n", + " 520 0.1301 23.2% 1.1m\n", + " 530 0.1346 23.2% 1.0m\n", + " 540 0.1957 23.2% 1.0m\n", + " 550 0.1912 23.2% 1.0m\n", + " 560 0.1694 23.2% 0.9m\n", + " 570 0.1686 23.2% 0.9m\n", + " 580 0.1881 23.2% 0.9m\n", + " 590 0.1584 23.2% 0.9m\n", + " 600 0.1614 23.2% 0.8m\n", + " 610 0.1756 23.2% 0.8m\n", + " 620 0.1431 23.2% 0.8m\n", + " 630 0.1586 23.2% 0.7m\n", + " 640 0.1610 23.2% 0.7m\n", + " 650 0.1507 23.2% 0.7m\n", + " 660 0.1828 23.2% 0.6m\n", + " 670 0.1821 23.2% 0.6m\n", + " 680 0.1741 23.2% 0.6m\n", + " 690 0.1743 23.2% 0.5m\n", + " 700 0.1384 23.2% 0.5m\n", + " 710 0.1473 23.2% 0.5m\n", + " 720 0.2020 23.2% 0.4m\n", + " 730 0.1780 23.2% 0.4m\n", + " 740 0.1367 23.2% 0.4m\n", + " 750 0.1831 23.2% 0.4m\n", + " 760 0.1700 23.2% 0.3m\n", + " 770 0.1737 23.2% 0.3m\n", + " 780 0.1240 23.2% 0.3m\n", + " 790 0.1505 23.2% 0.2m\n", + " 800 0.1622 23.2% 0.2m\n", + " 810 0.1403 23.2% 0.2m\n", + " 820 0.2161 23.2% 0.1m\n", + " 830 0.1647 23.2% 0.1m\n", + " 840 0.1490 23.2% 0.1m\n", + " 850 0.1440 23.2% 0.0m\n", + " 860 0.1608 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 21 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 21 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16725 โ†’ Val=0.16459\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 223m\n", + "\n", + "๐Ÿ”„ Epoch 22/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1404 23.2% 2.6m\n", + " 10 0.1781 23.2% 2.7m\n", + " 20 0.2118 23.2% 2.6m\n", + " 30 0.1514 23.2% 2.6m\n", + " 40 0.1529 23.2% 2.6m\n", + " 50 0.1722 23.2% 2.5m\n", + " 60 0.1623 23.2% 2.5m\n", + " 70 0.1480 23.2% 2.5m\n", + " 80 0.1442 23.2% 2.4m\n", + " 90 0.1529 23.2% 2.4m\n", + " 100 0.1960 23.2% 2.4m\n", + " 110 0.1617 23.2% 2.3m\n", + " 120 0.1426 23.2% 2.3m\n", + " 130 0.1972 23.2% 2.3m\n", + " 140 0.1964 23.2% 2.3m\n", + " 150 0.1763 23.2% 2.2m\n", + " 160 0.1763 23.2% 2.2m\n", + " 170 0.1882 23.2% 2.2m\n", + " 180 0.1398 23.2% 2.1m\n", + " 190 0.1250 23.2% 2.1m\n", + " 200 0.1613 23.2% 2.1m\n", + " 210 0.1822 23.2% 2.0m\n", + " 220 0.1677 23.2% 2.0m\n", + " 230 0.1772 23.2% 2.0m\n", + " 240 0.1442 23.2% 1.9m\n", + " 250 0.1971 23.2% 1.9m\n", + " 260 0.2006 23.2% 1.9m\n", + " 270 0.1858 23.2% 1.8m\n", + " 280 0.1506 23.2% 1.8m\n", + " 290 0.1243 23.2% 1.8m\n", + " 300 0.1930 23.2% 1.8m\n", + " 310 0.1664 23.2% 1.7m\n", + " 320 0.1552 23.2% 1.7m\n", + " 330 0.1936 23.2% 1.7m\n", + " 340 0.1769 23.2% 1.6m\n", + " 350 0.1850 23.2% 1.6m\n", + " 360 0.1334 23.2% 1.6m\n", + " 370 0.1472 23.2% 1.5m\n", + " 380 0.1551 23.2% 1.5m\n", + " 390 0.1689 23.2% 1.5m\n", + " 400 0.1573 23.2% 1.4m\n", + " 410 0.1619 23.2% 1.4m\n", + " 420 0.1851 23.2% 1.4m\n", + " 430 0.2175 23.2% 1.3m\n", + " 440 0.1338 23.2% 1.3m\n", + " 450 0.1884 23.2% 1.3m\n", + " 460 0.1494 23.2% 1.3m\n", + " 470 0.1695 23.2% 1.2m\n", + " 480 0.1854 23.2% 1.2m\n", + " 490 0.1884 23.2% 1.2m\n", + " 500 0.1874 23.2% 1.1m\n", + " 510 0.1683 23.2% 1.1m\n", + " 520 0.1304 23.2% 1.1m\n", + " 530 0.1492 23.2% 1.0m\n", + " 540 0.1802 23.2% 1.0m\n", + " 550 0.1606 23.2% 1.0m\n", + " 560 0.1061 23.2% 0.9m\n", + " 570 0.1774 23.2% 0.9m\n", + " 580 0.1736 23.2% 0.9m\n", + " 590 0.1446 23.2% 0.8m\n", + " 600 0.1711 23.2% 0.8m\n", + " 610 0.1632 23.2% 0.8m\n", + " 620 0.1271 23.2% 0.8m\n", + " 630 0.1771 23.2% 0.7m\n", + " 640 0.1744 23.2% 0.7m\n", + " 650 0.1795 23.2% 0.7m\n", + " 660 0.1629 23.2% 0.6m\n", + " 670 0.1959 23.2% 0.6m\n", + " 680 0.2048 23.2% 0.6m\n", + " 690 0.1668 23.2% 0.5m\n", + " 700 0.1809 23.2% 0.5m\n", + " 710 0.1806 23.2% 0.5m\n", + " 720 0.1675 23.2% 0.4m\n", + " 730 0.1853 23.2% 0.4m\n", + " 740 0.1625 23.2% 0.4m\n", + " 750 0.1947 23.2% 0.4m\n", + " 760 0.1421 23.2% 0.3m\n", + " 770 0.1453 23.2% 0.3m\n", + " 780 0.1387 23.2% 0.3m\n", + " 790 0.1653 23.2% 0.2m\n", + " 800 0.1745 23.2% 0.2m\n", + " 810 0.1586 23.2% 0.2m\n", + " 820 0.1576 23.2% 0.1m\n", + " 830 0.1995 23.2% 0.1m\n", + " 840 0.1435 23.2% 0.1m\n", + " 850 0.1958 23.2% 0.0m\n", + " 860 0.1854 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 22 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 22 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16735 โ†’ Val=0.16429\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 220m\n", + "\n", + "๐Ÿ”„ Epoch 23/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1919 23.2% 2.7m\n", + " 10 0.1179 23.2% 2.7m\n", + " 20 0.1817 23.2% 2.6m\n", + " 30 0.1693 23.2% 2.6m\n", + " 40 0.1533 23.2% 2.5m\n", + " 50 0.1777 23.2% 2.5m\n", + " 60 0.1474 23.2% 2.5m\n", + " 70 0.1589 23.2% 2.5m\n", + " 80 0.1490 23.2% 2.4m\n", + " 90 0.1533 23.2% 2.4m\n", + " 100 0.1908 23.2% 2.4m\n", + " 110 0.1761 23.2% 2.3m\n", + " 120 0.1608 23.2% 2.3m\n", + " 130 0.1553 23.2% 2.3m\n", + " 140 0.1680 23.2% 2.2m\n", + " 150 0.1771 23.2% 2.2m\n", + " 160 0.1735 23.2% 2.2m\n", + " 170 0.1711 23.2% 2.2m\n", + " 180 0.1675 23.2% 2.1m\n", + " 190 0.1659 23.2% 2.1m\n", + " 200 0.1733 23.2% 2.1m\n", + " 210 0.1656 23.2% 2.0m\n", + " 220 0.1982 23.2% 2.0m\n", + " 230 0.1701 23.2% 2.0m\n", + " 240 0.1986 23.2% 1.9m\n", + " 250 0.1797 23.2% 1.9m\n", + " 260 0.1253 23.2% 1.9m\n", + " 270 0.1747 23.2% 1.8m\n", + " 280 0.2022 23.2% 1.8m\n", + " 290 0.1604 23.2% 1.8m\n", + " 300 0.1756 23.2% 1.7m\n", + " 310 0.1601 23.2% 1.7m\n", + " 320 0.1528 23.2% 1.7m\n", + " 330 0.1368 23.2% 1.7m\n", + " 340 0.2012 23.2% 1.6m\n", + " 350 0.1617 23.2% 1.6m\n", + " 360 0.1657 23.2% 1.6m\n", + " 370 0.1746 23.2% 1.5m\n", + " 380 0.2142 23.2% 1.5m\n", + " 390 0.1470 23.2% 1.5m\n", + " 400 0.1776 23.2% 1.4m\n", + " 410 0.1726 23.2% 1.4m\n", + " 420 0.1374 23.2% 1.4m\n", + " 430 0.1993 23.2% 1.3m\n", + " 440 0.1328 23.2% 1.3m\n", + " 450 0.1481 23.2% 1.3m\n", + " 460 0.1750 23.2% 1.3m\n", + " 470 0.1359 23.2% 1.2m\n", + " 480 0.1561 23.2% 1.2m\n", + " 490 0.1805 23.2% 1.2m\n", + " 500 0.1693 23.2% 1.1m\n", + " 510 0.1988 23.2% 1.1m\n", + " 520 0.1536 23.2% 1.1m\n", + " 530 0.1867 23.2% 1.0m\n", + " 540 0.1658 23.2% 1.0m\n", + " 550 0.1307 23.2% 1.0m\n", + " 560 0.1824 23.2% 0.9m\n", + " 570 0.1381 23.2% 0.9m\n", + " 580 0.1456 23.2% 0.9m\n", + " 590 0.1558 23.2% 0.8m\n", + " 600 0.1515 23.2% 0.8m\n", + " 610 0.1711 23.2% 0.8m\n", + " 620 0.2202 23.2% 0.8m\n", + " 630 0.1809 23.2% 0.7m\n", + " 640 0.1881 23.2% 0.7m\n", + " 650 0.1407 23.2% 0.7m\n", + " 660 0.1525 23.2% 0.6m\n", + " 670 0.1688 23.2% 0.6m\n", + " 680 0.2197 23.2% 0.6m\n", + " 690 0.1580 23.2% 0.5m\n", + " 700 0.1557 23.2% 0.5m\n", + " 710 0.1988 23.2% 0.5m\n", + " 720 0.1740 23.2% 0.4m\n", + " 730 0.1822 23.2% 0.4m\n", + " 740 0.1330 23.2% 0.4m\n", + " 750 0.1965 23.2% 0.4m\n", + " 760 0.1593 23.2% 0.3m\n", + " 770 0.1466 23.2% 0.3m\n", + " 780 0.1653 23.2% 0.3m\n", + " 790 0.1526 23.2% 0.2m\n", + " 800 0.1695 23.2% 0.2m\n", + " 810 0.2039 23.2% 0.2m\n", + " 820 0.1467 23.2% 0.1m\n", + " 830 0.1795 23.2% 0.1m\n", + " 840 0.1449 23.2% 0.1m\n", + " 850 0.1863 23.2% 0.0m\n", + " 860 0.1886 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 23 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 23 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16726 โ†’ Val=0.16436\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 217m\n", + "\n", + "๐Ÿ”„ Epoch 24/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1784 23.2% 2.8m\n", + " 10 0.1791 23.2% 2.7m\n", + " 20 0.1540 23.2% 2.6m\n", + " 30 0.1354 23.2% 2.6m\n", + " 40 0.2029 23.2% 2.6m\n", + " 50 0.1575 23.2% 2.6m\n", + " 60 0.1827 23.2% 2.5m\n", + " 70 0.1260 23.2% 2.5m\n", + " 80 0.1560 23.2% 2.5m\n", + " 90 0.1508 23.2% 2.4m\n", + " 100 0.1676 23.2% 2.4m\n", + " 110 0.1573 23.2% 2.4m\n", + " 120 0.1581 23.2% 2.3m\n", + " 130 0.1877 23.2% 2.3m\n", + " 140 0.1582 23.2% 2.3m\n", + " 150 0.2198 23.2% 2.2m\n", + " 160 0.1804 23.2% 2.2m\n", + " 170 0.1584 23.2% 2.2m\n", + " 180 0.1523 23.2% 2.1m\n", + " 190 0.1831 23.2% 2.1m\n", + " 200 0.1597 23.2% 2.1m\n", + " 210 0.1725 23.2% 2.0m\n", + " 220 0.1698 23.2% 2.0m\n", + " 230 0.1556 23.2% 2.0m\n", + " 240 0.1577 23.2% 2.0m\n", + " 250 0.1459 23.2% 1.9m\n", + " 260 0.1952 23.2% 1.9m\n", + " 270 0.1682 23.2% 1.9m\n", + " 280 0.2038 23.2% 1.8m\n", + " 290 0.2085 23.2% 1.8m\n", + " 300 0.2178 23.2% 1.8m\n", + " 310 0.1812 23.2% 1.7m\n", + " 320 0.1360 23.2% 1.7m\n", + " 330 0.1263 23.2% 1.7m\n", + " 340 0.1922 23.2% 1.6m\n", + " 350 0.1507 23.2% 1.6m\n", + " 360 0.1929 23.2% 1.6m\n", + " 370 0.1603 23.2% 1.5m\n", + " 380 0.1592 23.2% 1.5m\n", + " 390 0.1591 23.2% 1.5m\n", + " 400 0.2136 23.2% 1.4m\n", + " 410 0.1463 23.2% 1.4m\n", + " 420 0.1978 23.2% 1.4m\n", + " 430 0.2040 23.2% 1.4m\n", + " 440 0.1976 23.2% 1.3m\n", + " 450 0.1721 23.2% 1.3m\n", + " 460 0.2101 23.2% 1.3m\n", + " 470 0.1455 23.2% 1.2m\n", + " 480 0.1758 23.2% 1.2m\n", + " 490 0.1753 23.2% 1.2m\n", + " 500 0.1660 23.2% 1.1m\n", + " 510 0.1668 23.2% 1.1m\n", + " 520 0.1468 23.2% 1.1m\n", + " 530 0.1693 23.2% 1.0m\n", + " 540 0.1547 23.2% 1.0m\n", + " 550 0.1782 23.2% 1.0m\n", + " 560 0.1541 23.2% 0.9m\n", + " 570 0.1823 23.2% 0.9m\n", + " 580 0.1762 23.2% 0.9m\n", + " 590 0.1177 23.2% 0.9m\n", + " 600 0.1644 23.2% 0.8m\n", + " 610 0.1428 23.2% 0.8m\n", + " 620 0.2136 23.2% 0.8m\n", + " 630 0.1695 23.2% 0.7m\n", + " 640 0.1729 23.2% 0.7m\n", + " 650 0.1907 23.2% 0.7m\n", + " 660 0.1800 23.2% 0.6m\n", + " 670 0.1408 23.2% 0.6m\n", + " 680 0.2325 23.2% 0.6m\n", + " 690 0.1932 23.2% 0.5m\n", + " 700 0.1751 23.2% 0.5m\n", + " 710 0.1414 23.2% 0.5m\n", + " 720 0.1776 23.2% 0.4m\n", + " 730 0.1801 23.2% 0.4m\n", + " 740 0.1665 23.2% 0.4m\n", + " 750 0.1415 23.2% 0.4m\n", + " 760 0.1941 23.2% 0.3m\n", + " 770 0.1887 23.2% 0.3m\n", + " 780 0.1610 23.2% 0.3m\n", + " 790 0.1458 23.2% 0.2m\n", + " 800 0.1692 23.2% 0.2m\n", + " 810 0.2065 23.2% 0.2m\n", + " 820 0.1396 23.2% 0.1m\n", + " 830 0.1655 23.2% 0.1m\n", + " 840 0.1873 23.2% 0.1m\n", + " 850 0.2018 23.2% 0.0m\n", + " 860 0.1700 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 24 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 24 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16733 โ†’ Val=0.16471\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 215m\n", + "\n", + "๐Ÿ”„ Epoch 25/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1493 23.2% 2.7m\n", + " 10 0.1632 23.2% 2.7m\n", + " 20 0.2123 23.2% 2.6m\n", + " 30 0.1685 23.2% 2.6m\n", + " 40 0.1588 23.2% 2.6m\n", + " 50 0.1819 23.2% 2.6m\n", + " 60 0.1917 23.2% 2.5m\n", + " 70 0.1683 23.2% 2.5m\n", + " 80 0.2200 23.2% 2.5m\n", + " 90 0.1886 23.2% 2.4m\n", + " 100 0.1673 23.2% 2.4m\n", + " 110 0.1621 23.2% 2.4m\n", + " 120 0.1671 23.2% 2.3m\n", + " 130 0.1218 23.2% 2.3m\n", + " 140 0.1296 23.2% 2.3m\n", + " 150 0.1768 23.2% 2.2m\n", + " 160 0.1765 23.2% 2.2m\n", + " 170 0.1577 23.2% 2.2m\n", + " 180 0.1433 23.2% 2.1m\n", + " 190 0.1724 23.2% 2.1m\n", + " 200 0.1688 23.2% 2.1m\n", + " 210 0.1811 23.2% 2.0m\n", + " 220 0.1330 23.2% 2.0m\n", + " 230 0.1983 23.2% 2.0m\n", + " 240 0.1839 23.2% 2.0m\n", + " 250 0.1864 23.2% 1.9m\n", + " 260 0.2065 23.2% 1.9m\n", + " 270 0.1453 23.2% 1.9m\n", + " 280 0.1545 23.2% 1.8m\n", + " 290 0.1392 23.2% 1.8m\n", + " 300 0.1507 23.2% 1.8m\n", + " 310 0.1584 23.2% 1.7m\n", + " 320 0.1843 23.2% 1.7m\n", + " 330 0.1946 23.2% 1.7m\n", + " 340 0.2254 23.2% 1.6m\n", + " 350 0.2149 23.2% 1.6m\n", + " 360 0.1752 23.2% 1.6m\n", + " 370 0.1341 23.2% 1.5m\n", + " 380 0.1549 23.2% 1.5m\n", + " 390 0.1600 23.2% 1.5m\n", + " 400 0.1565 23.2% 1.4m\n", + " 410 0.1704 23.2% 1.4m\n", + " 420 0.1813 23.2% 1.4m\n", + " 430 0.1927 23.2% 1.4m\n", + " 440 0.2012 23.2% 1.3m\n", + " 450 0.1245 23.2% 1.3m\n", + " 460 0.2493 23.2% 1.3m\n", + " 470 0.1953 23.2% 1.2m\n", + " 480 0.1263 23.2% 1.2m\n", + " 490 0.1695 23.2% 1.2m\n", + " 500 0.2262 23.2% 1.1m\n", + " 510 0.1419 23.2% 1.1m\n", + " 520 0.1853 23.2% 1.1m\n", + " 530 0.1892 23.2% 1.0m\n", + " 540 0.1545 23.2% 1.0m\n", + " 550 0.1436 23.2% 1.0m\n", + " 560 0.1844 23.2% 0.9m\n", + " 570 0.1393 23.2% 0.9m\n", + " 580 0.1808 23.2% 0.9m\n", + " 590 0.1477 23.2% 0.9m\n", + " 600 0.1763 23.2% 0.8m\n", + " 610 0.1104 23.2% 0.8m\n", + " 620 0.1409 23.2% 0.8m\n", + " 630 0.1764 23.2% 0.7m\n", + " 640 0.1895 23.2% 0.7m\n", + " 650 0.1745 23.2% 0.7m\n", + " 660 0.1787 23.2% 0.6m\n", + " 670 0.1354 23.2% 0.6m\n", + " 680 0.1745 23.2% 0.6m\n", + " 690 0.1698 23.2% 0.5m\n", + " 700 0.1790 23.2% 0.5m\n", + " 710 0.1607 23.2% 0.5m\n", + " 720 0.1823 23.2% 0.4m\n", + " 730 0.1541 23.2% 0.4m\n", + " 740 0.2009 23.2% 0.4m\n", + " 750 0.1599 23.2% 0.4m\n", + " 760 0.1857 23.2% 0.3m\n", + " 770 0.1813 23.2% 0.3m\n", + " 780 0.1669 23.2% 0.3m\n", + " 790 0.1907 23.2% 0.2m\n", + " 800 0.2187 23.2% 0.2m\n", + " 810 0.1404 23.2% 0.2m\n", + " 820 0.1527 23.2% 0.1m\n", + " 830 0.1498 23.2% 0.1m\n", + " 840 0.1815 23.2% 0.1m\n", + " 850 0.1850 23.2% 0.0m\n", + " 860 0.1376 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 25 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 25 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16711 โ†’ Val=0.16487\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.5% Overall=81.1%\n", + " โš ๏ธ No improvement for 5/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 212m\n", + "\n", + "๐Ÿ”„ Epoch 26/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2221 23.2% 2.7m\n", + " 10 0.1944 23.2% 2.6m\n", + " 20 0.1644 23.2% 2.6m\n", + " 30 0.1646 23.2% 2.6m\n", + " 40 0.1971 23.2% 2.5m\n", + " 50 0.1524 23.2% 2.5m\n", + " 60 0.2110 23.2% 2.5m\n", + " 70 0.1937 23.2% 2.5m\n", + " 80 0.1856 23.2% 2.4m\n", + " 90 0.1790 23.2% 2.4m\n", + " 100 0.1949 23.2% 2.4m\n", + " 110 0.1566 23.2% 2.3m\n", + " 120 0.1779 23.2% 2.3m\n", + " 130 0.1563 23.2% 2.3m\n", + " 140 0.1808 23.2% 2.2m\n", + " 150 0.1845 23.2% 2.2m\n", + " 160 0.1809 23.2% 2.2m\n", + " 170 0.2058 23.2% 2.2m\n", + " 180 0.1545 23.2% 2.1m\n", + " 190 0.2055 23.2% 2.1m\n", + " 200 0.1601 23.2% 2.1m\n", + " 210 0.1521 23.2% 2.0m\n", + " 220 0.1692 23.2% 2.0m\n", + " 230 0.2261 23.2% 2.0m\n", + " 240 0.1344 23.2% 1.9m\n", + " 250 0.1779 23.2% 1.9m\n", + " 260 0.1467 23.2% 1.9m\n", + " 270 0.1694 23.2% 1.9m\n", + " 280 0.1634 23.2% 1.8m\n", + " 290 0.1960 23.2% 1.8m\n", + " 300 0.1492 23.2% 1.8m\n", + " 310 0.1605 23.2% 1.7m\n", + " 320 0.1496 23.2% 1.7m\n", + " 330 0.1431 23.2% 1.7m\n", + " 340 0.1434 23.2% 1.6m\n", + " 350 0.1805 23.2% 1.6m\n", + " 360 0.1617 23.2% 1.6m\n", + " 370 0.1843 23.2% 1.5m\n", + " 380 0.1719 23.2% 1.5m\n", + " 390 0.1841 23.2% 1.5m\n", + " 400 0.1793 23.2% 1.4m\n", + " 410 0.2103 23.2% 1.4m\n", + " 420 0.1476 23.2% 1.4m\n", + " 430 0.1575 23.2% 1.4m\n", + " 440 0.2092 23.2% 1.3m\n", + " 450 0.1685 23.2% 1.3m\n", + " 460 0.1829 23.2% 1.3m\n", + " 470 0.1495 23.2% 1.2m\n", + " 480 0.1809 23.2% 1.2m\n", + " 490 0.1916 23.2% 1.2m\n", + " 500 0.1371 23.2% 1.1m\n", + " 510 0.1368 23.2% 1.1m\n", + " 520 0.2084 23.2% 1.1m\n", + " 530 0.2072 23.2% 1.0m\n", + " 540 0.1386 23.2% 1.0m\n", + " 550 0.1430 23.2% 1.0m\n", + " 560 0.1523 23.2% 0.9m\n", + " 570 0.1388 23.2% 0.9m\n", + " 580 0.1677 23.2% 0.9m\n", + " 590 0.1638 23.2% 0.9m\n", + " 600 0.1618 23.2% 0.8m\n", + " 610 0.1710 23.2% 0.8m\n", + " 620 0.1793 23.2% 0.8m\n", + " 630 0.1558 23.2% 0.7m\n", + " 640 0.1704 23.2% 0.7m\n", + " 650 0.1823 23.2% 0.7m\n", + " 660 0.1511 23.2% 0.6m\n", + " 670 0.1854 23.2% 0.6m\n", + " 680 0.1827 23.2% 0.6m\n", + " 690 0.1761 23.2% 0.5m\n", + " 700 0.1420 23.2% 0.5m\n", + " 710 0.1515 23.2% 0.5m\n", + " 720 0.1795 23.2% 0.4m\n", + " 730 0.1801 23.2% 0.4m\n", + " 740 0.1424 23.2% 0.4m\n", + " 750 0.1724 23.2% 0.4m\n", + " 760 0.1362 23.2% 0.3m\n", + " 770 0.1694 23.2% 0.3m\n", + " 780 0.1684 23.2% 0.3m\n", + " 790 0.1941 23.2% 0.2m\n", + " 800 0.2180 23.2% 0.2m\n", + " 810 0.1885 23.2% 0.2m\n", + " 820 0.1626 23.2% 0.1m\n", + " 830 0.1588 23.2% 0.1m\n", + " 840 0.1796 23.2% 0.1m\n", + " 850 0.1351 23.2% 0.0m\n", + " 860 0.1502 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 26 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 26 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16711 โ†’ Val=0.16428\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 209m\n", + "\n", + "๐Ÿ”„ Epoch 27/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1562 23.2% 2.6m\n", + " 10 0.1752 23.2% 2.7m\n", + " 20 0.1764 23.2% 2.6m\n", + " 30 0.1632 23.2% 2.6m\n", + " 40 0.1670 23.2% 2.6m\n", + " 50 0.1627 23.2% 2.5m\n", + " 60 0.1945 23.2% 2.5m\n", + " 70 0.1816 23.2% 2.5m\n", + " 80 0.1603 23.2% 2.5m\n", + " 90 0.1528 23.2% 2.4m\n", + " 100 0.1803 23.2% 2.4m\n", + " 110 0.1496 23.2% 2.4m\n", + " 120 0.1663 23.2% 2.3m\n", + " 130 0.1765 23.2% 2.3m\n", + " 140 0.1604 23.2% 2.3m\n", + " 150 0.1529 23.2% 2.2m\n", + " 160 0.1489 23.2% 2.2m\n", + " 170 0.1896 23.2% 2.2m\n", + " 180 0.1708 23.2% 2.1m\n", + " 190 0.1665 23.2% 2.1m\n", + " 200 0.2061 23.2% 2.1m\n", + " 210 0.1841 23.2% 2.0m\n", + " 220 0.1537 23.2% 2.0m\n", + " 230 0.1813 23.2% 2.0m\n", + " 240 0.1516 23.2% 1.9m\n", + " 250 0.1609 23.2% 1.9m\n", + " 260 0.1333 23.2% 1.9m\n", + " 270 0.1329 23.2% 1.9m\n", + " 280 0.1408 23.2% 1.8m\n", + " 290 0.1756 23.2% 1.8m\n", + " 300 0.1836 23.2% 1.8m\n", + " 310 0.1653 23.2% 1.7m\n", + " 320 0.1851 23.2% 1.7m\n", + " 330 0.1700 23.2% 1.7m\n", + " 340 0.1840 23.2% 1.6m\n", + " 350 0.1914 23.2% 1.6m\n", + " 360 0.1736 23.2% 1.6m\n", + " 370 0.1528 23.2% 1.5m\n", + " 380 0.1839 23.2% 1.5m\n", + " 390 0.1444 23.2% 1.5m\n", + " 400 0.1347 23.2% 1.4m\n", + " 410 0.1796 23.2% 1.4m\n", + " 420 0.1514 23.2% 1.4m\n", + " 430 0.1986 23.2% 1.3m\n", + " 440 0.1933 23.2% 1.3m\n", + " 450 0.1287 23.2% 1.3m\n", + " 460 0.2167 23.2% 1.3m\n", + " 470 0.1442 23.2% 1.2m\n", + " 480 0.1719 23.2% 1.2m\n", + " 490 0.2124 23.2% 1.2m\n", + " 500 0.1808 23.2% 1.1m\n", + " 510 0.1421 23.2% 1.1m\n", + " 520 0.1756 23.2% 1.1m\n", + " 530 0.1629 23.2% 1.0m\n", + " 540 0.1592 23.2% 1.0m\n", + " 550 0.1680 23.2% 1.0m\n", + " 560 0.1536 23.2% 0.9m\n", + " 570 0.1629 23.2% 0.9m\n", + " 580 0.1917 23.2% 0.9m\n", + " 590 0.1468 23.2% 0.8m\n", + " 600 0.1858 23.2% 0.8m\n", + " 610 0.1688 23.2% 0.8m\n", + " 620 0.1302 23.2% 0.8m\n", + " 630 0.1724 23.2% 0.7m\n", + " 640 0.1609 23.2% 0.7m\n", + " 650 0.1866 23.2% 0.7m\n", + " 660 0.1578 23.2% 0.6m\n", + " 670 0.1527 23.2% 0.6m\n", + " 680 0.1867 23.2% 0.6m\n", + " 690 0.1554 23.2% 0.5m\n", + " 700 0.1917 23.2% 0.5m\n", + " 710 0.1143 23.2% 0.5m\n", + " 720 0.1607 23.2% 0.4m\n", + " 730 0.1524 23.2% 0.4m\n", + " 740 0.1674 23.2% 0.4m\n", + " 750 0.1705 23.2% 0.4m\n", + " 760 0.1419 23.2% 0.3m\n", + " 770 0.1510 23.2% 0.3m\n", + " 780 0.1632 23.2% 0.3m\n", + " 790 0.2119 23.2% 0.2m\n", + " 800 0.1885 23.2% 0.2m\n", + " 810 0.1609 23.2% 0.2m\n", + " 820 0.1223 23.2% 0.1m\n", + " 830 0.1736 23.2% 0.1m\n", + " 840 0.1556 23.2% 0.1m\n", + " 850 0.1217 23.2% 0.0m\n", + " 860 0.1458 23.2% 0.0m\n", + " ----------------------------------------\n", + " ๏ฟฝ๏ฟฝ๏ฟฝ Epoch 27 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 27 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16710 โ†’ Val=0.16395\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16395)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 206m\n", + "\n", + "๐Ÿ”„ Epoch 28/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1857 23.2% 2.6m\n", + " 10 0.1679 23.2% 2.7m\n", + " 20 0.1248 23.2% 2.6m\n", + " 30 0.1515 23.2% 2.6m\n", + " 40 0.2149 23.2% 2.6m\n", + " 50 0.1859 23.2% 2.5m\n", + " 60 0.1350 23.2% 2.5m\n", + " 70 0.1836 23.2% 2.5m\n", + " 80 0.1387 23.2% 2.4m\n", + " 90 0.1941 23.2% 2.4m\n", + " 100 0.1496 23.2% 2.4m\n", + " 110 0.2256 23.2% 2.3m\n", + " 120 0.1839 23.2% 2.3m\n", + " 130 0.1614 23.2% 2.3m\n", + " 140 0.2069 23.2% 2.2m\n", + " 150 0.1176 23.2% 2.2m\n", + " 160 0.1723 23.2% 2.2m\n", + " 170 0.1792 23.2% 2.1m\n", + " 180 0.1522 23.2% 2.1m\n", + " 190 0.1519 23.2% 2.1m\n", + " 200 0.1875 23.2% 2.1m\n", + " 210 0.2035 23.2% 2.0m\n", + " 220 0.2039 23.2% 2.0m\n", + " 230 0.2107 23.2% 2.0m\n", + " 240 0.1832 23.2% 1.9m\n", + " 250 0.1845 23.2% 1.9m\n", + " 260 0.1560 23.2% 1.9m\n", + " 270 0.1551 23.2% 1.8m\n", + " 280 0.1797 23.2% 1.8m\n", + " 290 0.1811 23.2% 1.8m\n", + " 300 0.1978 23.2% 1.7m\n", + " 310 0.1787 23.2% 1.7m\n", + " 320 0.1669 23.2% 1.7m\n", + " 330 0.1513 23.2% 1.7m\n", + " 340 0.1768 23.2% 1.6m\n", + " 350 0.1580 23.2% 1.6m\n", + " 360 0.1937 23.2% 1.6m\n", + " 370 0.1385 23.2% 1.5m\n", + " 380 0.1725 23.2% 1.5m\n", + " 390 0.2004 23.2% 1.5m\n", + " 400 0.1629 23.2% 1.4m\n", + " 410 0.1646 23.2% 1.4m\n", + " 420 0.1776 23.2% 1.4m\n", + " 430 0.1368 23.2% 1.3m\n", + " 440 0.1872 23.2% 1.3m\n", + " 450 0.1613 23.2% 1.3m\n", + " 460 0.1301 23.2% 1.3m\n", + " 470 0.1635 23.2% 1.2m\n", + " 480 0.2024 23.2% 1.2m\n", + " 490 0.1869 23.2% 1.2m\n", + " 500 0.1572 23.2% 1.1m\n", + " 510 0.1703 23.2% 1.1m\n", + " 520 0.1335 23.2% 1.1m\n", + " 530 0.1901 23.2% 1.0m\n", + " 540 0.2203 23.2% 1.0m\n", + " 550 0.1102 23.2% 1.0m\n", + " 560 0.1521 23.2% 0.9m\n", + " 570 0.1993 23.2% 0.9m\n", + " 580 0.1885 23.2% 0.9m\n", + " 590 0.1735 23.2% 0.9m\n", + " 600 0.1855 23.2% 0.8m\n", + " 610 0.1786 23.2% 0.8m\n", + " 620 0.1606 23.2% 0.8m\n", + " 630 0.1519 23.2% 0.7m\n", + " 640 0.2004 23.2% 0.7m\n", + " 650 0.1755 23.2% 0.7m\n", + " 660 0.1812 23.2% 0.6m\n", + " 670 0.1818 23.2% 0.6m\n", + " 680 0.1474 23.2% 0.6m\n", + " 690 0.1664 23.2% 0.5m\n", + " 700 0.1866 23.2% 0.5m\n", + " 710 0.1388 23.2% 0.5m\n", + " 720 0.1722 23.2% 0.4m\n", + " 730 0.1476 23.2% 0.4m\n", + " 740 0.1590 23.2% 0.4m\n", + " 750 0.1418 23.2% 0.4m\n", + " 760 0.1236 23.2% 0.3m\n", + " 770 0.1501 23.2% 0.3m\n", + " 780 0.1878 23.2% 0.3m\n", + " 790 0.1894 23.2% 0.2m\n", + " 800 0.1394 23.2% 0.2m\n", + " 810 0.1719 23.2% 0.2m\n", + " 820 0.1505 23.2% 0.1m\n", + " 830 0.1283 23.2% 0.1m\n", + " 840 0.2084 23.2% 0.1m\n", + " 850 0.1848 23.2% 0.0m\n", + " 860 0.1855 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 28 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 28 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16705 โ†’ Val=0.16416\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 203m\n", + "\n", + "๐Ÿ”„ Epoch 29/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1680 23.2% 2.9m\n", + " 10 0.1709 23.2% 2.7m\n", + " 20 0.1564 23.2% 2.6m\n", + " 30 0.1497 23.2% 2.6m\n", + " 40 0.1938 23.2% 2.5m\n", + " 50 0.1580 23.2% 2.5m\n", + " 60 0.1674 23.2% 2.5m\n", + " 70 0.1614 23.2% 2.5m\n", + " 80 0.1530 23.2% 2.4m\n", + " 90 0.1326 23.2% 2.4m\n", + " 100 0.1951 23.2% 2.4m\n", + " 110 0.2200 23.2% 2.3m\n", + " 120 0.1892 23.2% 2.3m\n", + " 130 0.1582 23.2% 2.3m\n", + " 140 0.1681 23.2% 2.3m\n", + " 150 0.1839 23.2% 2.2m\n", + " 160 0.1377 23.2% 2.2m\n", + " 170 0.1787 23.2% 2.2m\n", + " 180 0.1833 23.2% 2.1m\n", + " 190 0.1502 23.2% 2.1m\n", + " 200 0.1825 23.2% 2.1m\n", + " 210 0.1391 23.2% 2.0m\n", + " 220 0.1692 23.2% 2.0m\n", + " 230 0.1773 23.2% 2.0m\n", + " 240 0.1646 23.2% 1.9m\n", + " 250 0.1942 23.2% 1.9m\n", + " 260 0.1613 23.2% 1.9m\n", + " 270 0.1640 23.2% 1.8m\n", + " 280 0.2096 23.2% 1.8m\n", + " 290 0.1600 23.2% 1.8m\n", + " 300 0.1183 23.2% 1.8m\n", + " 310 0.1761 23.2% 1.7m\n", + " 320 0.1710 23.2% 1.7m\n", + " 330 0.1541 23.2% 1.7m\n", + " 340 0.1621 23.2% 1.6m\n", + " 350 0.1883 23.2% 1.6m\n", + " 360 0.1604 23.2% 1.6m\n", + " 370 0.1772 23.2% 1.5m\n", + " 380 0.1794 23.2% 1.5m\n", + " 390 0.1816 23.2% 1.5m\n", + " 400 0.1862 23.2% 1.4m\n", + " 410 0.1576 23.2% 1.4m\n", + " 420 0.1851 23.2% 1.4m\n", + " 430 0.1475 23.2% 1.3m\n", + " 440 0.2136 23.2% 1.3m\n", + " 450 0.1631 23.2% 1.3m\n", + " 460 0.1731 23.2% 1.3m\n", + " 470 0.1611 23.2% 1.2m\n", + " 480 0.1848 23.2% 1.2m\n", + " 490 0.1757 23.2% 1.2m\n", + " 500 0.1355 23.2% 1.1m\n", + " 510 0.1741 23.2% 1.1m\n", + " 520 0.1479 23.2% 1.1m\n", + " 530 0.1688 23.2% 1.0m\n", + " 540 0.1356 23.2% 1.0m\n", + " 550 0.2057 23.2% 1.0m\n", + " 560 0.1439 23.2% 0.9m\n", + " 570 0.1593 23.2% 0.9m\n", + " 580 0.1456 23.2% 0.9m\n", + " 590 0.1930 23.2% 0.9m\n", + " 600 0.1668 23.2% 0.8m\n", + " 610 0.1626 23.2% 0.8m\n", + " 620 0.1573 23.2% 0.8m\n", + " 630 0.2317 23.2% 0.7m\n", + " 640 0.1591 23.2% 0.7m\n", + " 650 0.2059 23.2% 0.7m\n", + " 660 0.1640 23.2% 0.6m\n", + " 670 0.1557 23.2% 0.6m\n", + " 680 0.1797 23.2% 0.6m\n", + " 690 0.1343 23.2% 0.5m\n", + " 700 0.1346 23.2% 0.5m\n", + " 710 0.1700 23.2% 0.5m\n", + " 720 0.1636 23.2% 0.4m\n", + " 730 0.1864 23.2% 0.4m\n", + " 740 0.1686 23.2% 0.4m\n", + " 750 0.1587 23.2% 0.4m\n", + " 760 0.1888 23.2% 0.3m\n", + " 770 0.1460 23.2% 0.3m\n", + " 780 0.1795 23.2% 0.3m\n", + " 790 0.1542 23.2% 0.2m\n", + " 800 0.1820 23.2% 0.2m\n", + " 810 0.1411 23.2% 0.2m\n", + " 820 0.1734 23.2% 0.1m\n", + " 830 0.1621 23.2% 0.1m\n", + " 840 0.1894 23.2% 0.1m\n", + " 850 0.1535 23.2% 0.0m\n", + " 860 0.1768 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 29 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 29 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16713 โ†’ Val=0.16424\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 201m\n", + "\n", + "๐Ÿ”„ Epoch 30/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2034 23.2% 2.7m\n", + " 10 0.1697 23.2% 2.7m\n", + " 20 0.1645 23.2% 2.6m\n", + " 30 0.2069 23.2% 2.6m\n", + " 40 0.1663 23.2% 2.6m\n", + " 50 0.1825 23.2% 2.5m\n", + " 60 0.1772 23.2% 2.5m\n", + " 70 0.1803 23.2% 2.5m\n", + " 80 0.2421 23.2% 2.4m\n", + " 90 0.1491 23.2% 2.4m\n", + " 100 0.1535 23.2% 2.4m\n", + " 110 0.1895 23.2% 2.4m\n", + " 120 0.1752 23.2% 2.3m\n", + " 130 0.1615 23.2% 2.3m\n", + " 140 0.1659 23.2% 2.3m\n", + " 150 0.1527 23.2% 2.2m\n", + " 160 0.1522 23.2% 2.2m\n", + " 170 0.1976 23.2% 2.2m\n", + " 180 0.1842 23.2% 2.1m\n", + " 190 0.1474 23.2% 2.1m\n", + " 200 0.1770 23.2% 2.1m\n", + " 210 0.1674 23.2% 2.0m\n", + " 220 0.1451 23.2% 2.0m\n", + " 230 0.1728 23.2% 2.0m\n", + " 240 0.1761 23.2% 1.9m\n", + " 250 0.1618 23.2% 1.9m\n", + " 260 0.1558 23.2% 1.9m\n", + " 270 0.1207 23.2% 1.9m\n", + " 280 0.1661 23.2% 1.8m\n", + " 290 0.1754 23.2% 1.8m\n", + " 300 0.1771 23.2% 1.8m\n", + " 310 0.1973 23.2% 1.7m\n", + " 320 0.1717 23.2% 1.7m\n", + " 330 0.1649 23.2% 1.7m\n", + " 340 0.1475 23.2% 1.6m\n", + " 350 0.1374 23.2% 1.6m\n", + " 360 0.1717 23.2% 1.6m\n", + " 370 0.1881 23.2% 1.5m\n", + " 380 0.1873 23.2% 1.5m\n", + " 390 0.1793 23.2% 1.5m\n", + " 400 0.1877 23.2% 1.4m\n", + " 410 0.1769 23.2% 1.4m\n", + " 420 0.1676 23.2% 1.4m\n", + " 430 0.1493 23.2% 1.4m\n", + " 440 0.1963 23.2% 1.3m\n", + " 450 0.1691 23.2% 1.3m\n", + " 460 0.1945 23.2% 1.3m\n", + " 470 0.1620 23.2% 1.2m\n", + " 480 0.1806 23.2% 1.2m\n", + " 490 0.2004 23.2% 1.2m\n", + " 500 0.1609 23.2% 1.1m\n", + " 510 0.1489 23.2% 1.1m\n", + " 520 0.1218 23.2% 1.1m\n", + " 530 0.1687 23.2% 1.0m\n", + " 540 0.1894 23.2% 1.0m\n", + " 550 0.1882 23.2% 1.0m\n", + " 560 0.2213 23.2% 0.9m\n", + " 570 0.1717 23.2% 0.9m\n", + " 580 0.1835 23.2% 0.9m\n", + " 590 0.2069 23.2% 0.9m\n", + " 600 0.1578 23.2% 0.8m\n", + " 610 0.1737 23.2% 0.8m\n", + " 620 0.1546 23.2% 0.8m\n", + " 630 0.1649 23.2% 0.7m\n", + " 640 0.1538 23.2% 0.7m\n", + " 650 0.2005 23.2% 0.7m\n", + " 660 0.1823 23.2% 0.6m\n", + " 670 0.1586 23.2% 0.6m\n", + " 680 0.1906 23.2% 0.6m\n", + " 690 0.1469 23.2% 0.5m\n", + " 700 0.1712 23.2% 0.5m\n", + " 710 0.1292 23.2% 0.5m\n", + " 720 0.1326 23.2% 0.4m\n", + " 730 0.1650 23.2% 0.4m\n", + " 740 0.1775 23.2% 0.4m\n", + " 750 0.1432 23.2% 0.4m\n", + " 760 0.1915 23.2% 0.3m\n", + " 770 0.1620 23.2% 0.3m\n", + " 780 0.1637 23.2% 0.3m\n", + " 790 0.1638 23.2% 0.2m\n", + " 800 0.1038 23.2% 0.2m\n", + " 810 0.1771 23.2% 0.2m\n", + " 820 0.2012 23.2% 0.1m\n", + " 830 0.1727 23.2% 0.1m\n", + " 840 0.1496 23.2% 0.1m\n", + " 850 0.1923 23.2% 0.0m\n", + " 860 0.1490 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 30 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 30 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16705 โ†’ Val=0.16420\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 198m\n", + "\n", + "๐Ÿ”„ Epoch 31/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1863 23.2% 2.7m\n", + " 10 0.1529 23.2% 2.6m\n", + " 20 0.1882 23.2% 2.7m\n", + " 30 0.1630 23.2% 2.6m\n", + " 40 0.1862 23.2% 2.6m\n", + " 50 0.1410 23.2% 2.6m\n", + " 60 0.1245 23.2% 2.5m\n", + " 70 0.1581 23.2% 2.5m\n", + " 80 0.1894 23.2% 2.5m\n", + " 90 0.1650 23.2% 2.4m\n", + " 100 0.1623 23.2% 2.4m\n", + " 110 0.1558 23.2% 2.4m\n", + " 120 0.1938 23.2% 2.3m\n", + " 130 0.2118 23.2% 2.3m\n", + " 140 0.1666 23.2% 2.3m\n", + " 150 0.1781 23.2% 2.2m\n", + " 160 0.1441 23.2% 2.2m\n", + " 170 0.1432 23.2% 2.2m\n", + " 180 0.1580 23.2% 2.1m\n", + " 190 0.1747 23.2% 2.1m\n", + " 200 0.2255 23.2% 2.1m\n", + " 210 0.1776 23.2% 2.0m\n", + " 220 0.1447 23.2% 2.0m\n", + " 230 0.1653 23.2% 2.0m\n", + " 240 0.1579 23.2% 2.0m\n", + " 250 0.1420 23.2% 1.9m\n", + " 260 0.1442 23.2% 1.9m\n", + " 270 0.1565 23.2% 1.9m\n", + " 280 0.1850 23.2% 1.8m\n", + " 290 0.1594 23.2% 1.8m\n", + " 300 0.1471 23.2% 1.8m\n", + " 310 0.1602 23.2% 1.7m\n", + " 320 0.1841 23.2% 1.7m\n", + " 330 0.1437 23.2% 1.7m\n", + " 340 0.1933 23.2% 1.6m\n", + " 350 0.1712 23.2% 1.6m\n", + " 360 0.1745 23.2% 1.6m\n", + " 370 0.1512 23.2% 1.5m\n", + " 380 0.2005 23.2% 1.5m\n", + " 390 0.1537 23.2% 1.5m\n", + " 400 0.1755 23.2% 1.5m\n", + " 410 0.1608 23.2% 1.4m\n", + " 420 0.2001 23.2% 1.4m\n", + " 430 0.1779 23.2% 1.4m\n", + " 440 0.1595 23.2% 1.3m\n", + " 450 0.1370 23.2% 1.3m\n", + " 460 0.1567 23.2% 1.3m\n", + " 470 0.1976 23.2% 1.2m\n", + " 480 0.1667 23.2% 1.2m\n", + " 490 0.1321 23.2% 1.2m\n", + " 500 0.1460 23.2% 1.1m\n", + " 510 0.1716 23.2% 1.1m\n", + " 520 0.1794 23.2% 1.1m\n", + " 530 0.1521 23.2% 1.0m\n", + " 540 0.2002 23.2% 1.0m\n", + " 550 0.1924 23.2% 1.0m\n", + " 560 0.1721 23.2% 0.9m\n", + " 570 0.1303 23.2% 0.9m\n", + " 580 0.1926 23.2% 0.9m\n", + " 590 0.1247 23.2% 0.9m\n", + " 600 0.1295 23.2% 0.8m\n", + " 610 0.1723 23.2% 0.8m\n", + " 620 0.1801 23.2% 0.8m\n", + " 630 0.1380 23.2% 0.7m\n", + " 640 0.1388 23.2% 0.7m\n", + " 650 0.1617 23.2% 0.7m\n", + " 660 0.1853 23.2% 0.6m\n", + " 670 0.1473 23.2% 0.6m\n", + " 680 0.1662 23.2% 0.6m\n", + " 690 0.1854 23.2% 0.5m\n", + " 700 0.1758 23.2% 0.5m\n", + " 710 0.1572 23.2% 0.5m\n", + " 720 0.1678 23.2% 0.4m\n", + " 730 0.2120 23.2% 0.4m\n", + " 740 0.1704 23.2% 0.4m\n", + " 750 0.1927 23.2% 0.4m\n", + " 760 0.1193 23.2% 0.3m\n", + " 770 0.1589 23.2% 0.3m\n", + " 780 0.1886 23.2% 0.3m\n", + " 790 0.1760 23.2% 0.2m\n", + " 800 0.1531 23.2% 0.2m\n", + " 810 0.1324 23.2% 0.2m\n", + " 820 0.1753 23.2% 0.1m\n", + " 830 0.1279 23.2% 0.1m\n", + " 840 0.2046 23.2% 0.1m\n", + " 850 0.1533 23.2% 0.0m\n", + " 860 0.1762 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 31 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 31 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16706 โ†’ Val=0.16419\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 196m\n", + "\n", + "๐Ÿ”„ Epoch 32/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1265 23.2% 2.7m\n", + " 10 0.1536 23.2% 2.7m\n", + " 20 0.1997 23.2% 2.6m\n", + " 30 0.1824 23.2% 2.6m\n", + " 40 0.1682 23.2% 2.6m\n", + " 50 0.1518 23.2% 2.5m\n", + " 60 0.1798 23.2% 2.5m\n", + " 70 0.1855 23.2% 2.5m\n", + " 80 0.1872 23.2% 2.4m\n", + " 90 0.2019 23.2% 2.4m\n", + " 100 0.1860 23.2% 2.4m\n", + " 110 0.2063 23.2% 2.4m\n", + " 120 0.1519 23.2% 2.3m\n", + " 130 0.1684 23.2% 2.3m\n", + " 140 0.1791 23.2% 2.3m\n", + " 150 0.1749 23.2% 2.2m\n", + " 160 0.2285 23.2% 2.2m\n", + " 170 0.1674 23.2% 2.2m\n", + " 180 0.1583 23.2% 2.1m\n", + " 190 0.1589 23.2% 2.1m\n", + " 200 0.1706 23.2% 2.1m\n", + " 210 0.2116 23.2% 2.0m\n", + " 220 0.1893 23.2% 2.0m\n", + " 230 0.2097 23.2% 2.0m\n", + " 240 0.1768 23.2% 1.9m\n", + " 250 0.1670 23.2% 1.9m\n", + " 260 0.1877 23.2% 1.9m\n", + " 270 0.1561 23.2% 1.8m\n", + " 280 0.1670 23.2% 1.8m\n", + " 290 0.2080 23.2% 1.8m\n", + " 300 0.1598 23.2% 1.8m\n", + " 310 0.1579 23.2% 1.7m\n", + " 320 0.1596 23.2% 1.7m\n", + " 330 0.1993 23.2% 1.7m\n", + " 340 0.1870 23.2% 1.6m\n", + " 350 0.2102 23.2% 1.6m\n", + " 360 0.1709 23.2% 1.6m\n", + " 370 0.1754 23.2% 1.5m\n", + " 380 0.1749 23.2% 1.5m\n", + " 390 0.1611 23.2% 1.5m\n", + " 400 0.1552 23.2% 1.4m\n", + " 410 0.1693 23.2% 1.4m\n", + " 420 0.1471 23.2% 1.4m\n", + " 430 0.1483 23.2% 1.3m\n", + " 440 0.1668 23.2% 1.3m\n", + " 450 0.1157 23.2% 1.3m\n", + " 460 0.1614 23.2% 1.3m\n", + " 470 0.1659 23.2% 1.2m\n", + " 480 0.1799 23.2% 1.2m\n", + " 490 0.2190 23.2% 1.2m\n", + " 500 0.2133 23.2% 1.1m\n", + " 510 0.1692 23.2% 1.1m\n", + " 520 0.1575 23.2% 1.1m\n", + " 530 0.1841 23.2% 1.0m\n", + " 540 0.1910 23.2% 1.0m\n", + " 550 0.1608 23.2% 1.0m\n", + " 560 0.1653 23.2% 0.9m\n", + " 570 0.1987 23.2% 0.9m\n", + " 580 0.1565 23.2% 0.9m\n", + " 590 0.1529 23.2% 0.9m\n", + " 600 0.1742 23.2% 0.8m\n", + " 610 0.1556 23.2% 0.8m\n", + " 620 0.1364 23.2% 0.8m\n", + " 630 0.1264 23.2% 0.7m\n", + " 640 0.1711 23.2% 0.7m\n", + " 650 0.1773 23.2% 0.7m\n", + " 660 0.1309 23.2% 0.6m\n", + " 670 0.1618 23.2% 0.6m\n", + " 680 0.1459 23.2% 0.6m\n", + " 690 0.1396 23.2% 0.5m\n", + " 700 0.1794 23.2% 0.5m\n", + " 710 0.1561 23.2% 0.5m\n", + " 720 0.1431 23.2% 0.4m\n", + " 730 0.1158 23.2% 0.4m\n", + " 740 0.1594 23.2% 0.4m\n", + " 750 0.1422 23.2% 0.4m\n", + " 760 0.1671 23.2% 0.3m\n", + " 770 0.1985 23.2% 0.3m\n", + " 780 0.1540 23.2% 0.3m\n", + " 790 0.1865 23.2% 0.2m\n", + " 800 0.1933 23.2% 0.2m\n", + " 810 0.1578 23.2% 0.2m\n", + " 820 0.1806 23.2% 0.1m\n", + " 830 0.1490 23.2% 0.1m\n", + " 840 0.1610 23.2% 0.1m\n", + " 850 0.1182 23.2% 0.0m\n", + " 860 0.1655 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 32 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 32 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16698 โ†’ Val=0.16457\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โš ๏ธ No improvement for 5/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 192m\n", + "\n", + "๐Ÿ”„ Epoch 33/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1600 23.2% 2.7m\n", + " 10 0.1318 23.2% 2.6m\n", + " 20 0.1765 23.2% 2.6m\n", + " 30 0.1813 23.2% 2.6m\n", + " 40 0.1414 23.2% 2.5m\n", + " 50 0.1817 23.2% 2.5m\n", + " 60 0.1615 23.2% 2.5m\n", + " 70 0.1521 23.2% 2.5m\n", + " 80 0.1576 23.2% 2.4m\n", + " 90 0.1554 23.2% 2.4m\n", + " 100 0.1591 23.2% 2.4m\n", + " 110 0.1940 23.2% 2.3m\n", + " 120 0.2145 23.2% 2.3m\n", + " 130 0.1729 23.2% 2.3m\n", + " 140 0.1999 23.2% 2.3m\n", + " 150 0.1958 23.2% 2.2m\n", + " 160 0.1316 23.2% 2.2m\n", + " 170 0.1748 23.2% 2.2m\n", + " 180 0.1625 23.2% 2.1m\n", + " 190 0.1913 23.2% 2.1m\n", + " 200 0.1879 23.2% 2.1m\n", + " 210 0.1560 23.2% 2.0m\n", + " 220 0.1644 23.2% 2.0m\n", + " 230 0.1572 23.2% 2.0m\n", + " 240 0.1598 23.2% 1.9m\n", + " 250 0.1593 23.2% 1.9m\n", + " 260 0.1373 23.2% 1.9m\n", + " 270 0.1640 23.2% 1.9m\n", + " 280 0.1748 23.2% 1.8m\n", + " 290 0.1556 23.2% 1.8m\n", + " 300 0.2012 23.2% 1.8m\n", + " 310 0.1498 23.2% 1.7m\n", + " 320 0.1369 23.2% 1.7m\n", + " 330 0.1826 23.2% 1.7m\n", + " 340 0.1510 23.2% 1.6m\n", + " 350 0.2045 23.2% 1.6m\n", + " 360 0.1960 23.2% 1.6m\n", + " 370 0.1472 23.2% 1.5m\n", + " 380 0.2035 23.2% 1.5m\n", + " 390 0.1869 23.2% 1.5m\n", + " 400 0.1913 23.2% 1.4m\n", + " 410 0.1520 23.2% 1.4m\n", + " 420 0.1386 23.2% 1.4m\n", + " 430 0.2162 23.2% 1.4m\n", + " 440 0.1556 23.2% 1.3m\n", + " 450 0.1746 23.2% 1.3m\n", + " 460 0.1887 23.2% 1.3m\n", + " 470 0.1561 23.2% 1.2m\n", + " 480 0.1377 23.2% 1.2m\n", + " 490 0.1697 23.2% 1.2m\n", + " 500 0.1517 23.2% 1.1m\n", + " 510 0.1788 23.2% 1.1m\n", + " 520 0.1831 23.2% 1.1m\n", + " 530 0.1480 23.2% 1.0m\n", + " 540 0.1769 23.2% 1.0m\n", + " 550 0.1611 23.2% 1.0m\n", + " 560 0.1922 23.2% 0.9m\n", + " 570 0.1743 23.2% 0.9m\n", + " 580 0.1694 23.2% 0.9m\n", + " 590 0.1896 23.2% 0.9m\n", + " 600 0.1763 23.2% 0.8m\n", + " 610 0.1779 23.2% 0.8m\n", + " 620 0.1530 23.2% 0.8m\n", + " 630 0.1991 23.2% 0.7m\n", + " 640 0.1663 23.2% 0.7m\n", + " 650 0.2019 23.2% 0.7m\n", + " 660 0.1849 23.2% 0.6m\n", + " 670 0.1726 23.2% 0.6m\n", + " 680 0.1713 23.2% 0.6m\n", + " 690 0.1680 23.2% 0.5m\n", + " 700 0.1534 23.2% 0.5m\n", + " 710 0.2081 23.2% 0.5m\n", + " 720 0.1809 23.2% 0.4m\n", + " 730 0.2037 23.2% 0.4m\n", + " 740 0.1732 23.2% 0.4m\n", + " 750 0.1678 23.2% 0.4m\n", + " 760 0.1856 23.2% 0.3m\n", + " 770 0.1643 23.2% 0.3m\n", + " 780 0.1558 23.2% 0.3m\n", + " 790 0.1214 23.2% 0.2m\n", + " 800 0.1543 23.2% 0.2m\n", + " 810 0.1774 23.2% 0.2m\n", + " 820 0.1800 23.2% 0.1m\n", + " 830 0.1744 23.2% 0.1m\n", + " 840 0.1910 23.2% 0.1m\n", + " 850 0.1508 23.2% 0.0m\n", + " 860 0.1386 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 33 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 33 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16691 โ†’ Val=0.16443\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 190m\n", + "\n", + "๐Ÿ”„ Epoch 34/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1677 23.2% 2.7m\n", + " 10 0.1956 23.2% 2.7m\n", + " 20 0.1511 23.2% 2.6m\n", + " 30 0.1735 23.2% 2.6m\n", + " 40 0.1202 23.2% 2.6m\n", + " 50 0.1290 23.2% 2.6m\n", + " 60 0.1483 23.2% 2.5m\n", + " 70 0.1380 23.2% 2.5m\n", + " 80 0.1899 23.2% 2.5m\n", + " 90 0.1736 23.2% 2.4m\n", + " 100 0.1472 23.2% 2.4m\n", + " 110 0.1510 23.2% 2.4m\n", + " 120 0.1481 23.2% 2.3m\n", + " 130 0.1813 23.2% 2.3m\n", + " 140 0.1760 23.2% 2.3m\n", + " 150 0.2084 23.2% 2.2m\n", + " 160 0.1524 23.2% 2.2m\n", + " 170 0.1802 23.2% 2.2m\n", + " 180 0.1473 23.2% 2.1m\n", + " 190 0.1584 23.2% 2.1m\n", + " 200 0.1649 23.2% 2.1m\n", + " 210 0.1284 23.2% 2.0m\n", + " 220 0.1270 23.2% 2.0m\n", + " 230 0.1915 23.2% 2.0m\n", + " 240 0.1637 23.2% 2.0m\n", + " 250 0.1635 23.2% 1.9m\n", + " 260 0.1505 23.2% 1.9m\n", + " 270 0.1487 23.2% 1.9m\n", + " 280 0.1714 23.2% 1.8m\n", + " 290 0.1800 23.2% 1.8m\n", + " 300 0.1610 23.2% 1.8m\n", + " 310 0.1557 23.2% 1.7m\n", + " 320 0.1673 23.2% 1.7m\n", + " 330 0.1652 23.2% 1.7m\n", + " 340 0.1539 23.2% 1.6m\n", + " 350 0.1420 23.2% 1.6m\n", + " 360 0.1244 23.2% 1.6m\n", + " 370 0.1538 23.2% 1.5m\n", + " 380 0.1333 23.2% 1.5m\n", + " 390 0.1268 23.2% 1.5m\n", + " 400 0.1880 23.2% 1.5m\n", + " 410 0.1844 23.2% 1.4m\n", + " 420 0.1627 23.2% 1.4m\n", + " 430 0.1932 23.2% 1.4m\n", + " 440 0.1498 23.2% 1.3m\n", + " 450 0.2086 23.2% 1.3m\n", + " 460 0.1478 23.2% 1.3m\n", + " 470 0.1782 23.2% 1.2m\n", + " 480 0.1280 23.2% 1.2m\n", + " 490 0.1548 23.2% 1.2m\n", + " 500 0.1677 23.2% 1.1m\n", + " 510 0.1690 23.2% 1.1m\n", + " 520 0.1383 23.2% 1.1m\n", + " 530 0.1668 23.2% 1.0m\n", + " 540 0.1964 23.2% 1.0m\n", + " 550 0.1818 23.2% 1.0m\n", + " 560 0.1844 23.2% 1.0m\n", + " 570 0.1972 23.2% 0.9m\n", + " 580 0.1450 23.2% 0.9m\n", + " 590 0.1772 23.2% 0.9m\n", + " 600 0.2183 23.2% 0.8m\n", + " 610 0.1573 23.2% 0.8m\n", + " 620 0.1851 23.2% 0.8m\n", + " 630 0.1582 23.2% 0.7m\n", + " 640 0.1668 23.2% 0.7m\n", + " 650 0.1153 23.2% 0.7m\n", + " 660 0.1431 23.2% 0.6m\n", + " 670 0.1574 23.2% 0.6m\n", + " 680 0.1633 23.2% 0.6m\n", + " 690 0.1619 23.2% 0.5m\n", + " 700 0.1700 23.2% 0.5m\n", + " 710 0.1734 23.2% 0.5m\n", + " 720 0.1543 23.2% 0.4m\n", + " 730 0.1577 23.2% 0.4m\n", + " 740 0.1693 23.2% 0.4m\n", + " 750 0.2101 23.2% 0.4m\n", + " 760 0.1352 23.2% 0.3m\n", + " 770 0.1853 23.2% 0.3m\n", + " 780 0.1437 23.2% 0.3m\n", + " 790 0.1454 23.2% 0.2m\n", + " 800 0.1618 23.2% 0.2m\n", + " 810 0.1795 23.2% 0.2m\n", + " 820 0.2156 23.2% 0.1m\n", + " 830 0.1548 23.2% 0.1m\n", + " 840 0.1455 23.2% 0.1m\n", + " 850 0.1930 23.2% 0.0m\n", + " 860 0.1759 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 34 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 34 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16700 โ†’ Val=0.16430\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 188m\n", + "\n", + "๐Ÿ”„ Epoch 35/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1306 23.2% 2.6m\n", + " 10 0.1704 23.2% 2.6m\n", + " 20 0.1881 23.2% 2.6m\n", + " 30 0.1984 23.2% 2.6m\n", + " 40 0.1357 23.2% 2.6m\n", + " 50 0.2121 23.2% 2.5m\n", + " 60 0.1809 23.2% 2.5m\n", + " 70 0.1674 23.2% 2.5m\n", + " 80 0.1829 23.2% 2.5m\n", + " 90 0.1758 23.2% 2.4m\n", + " 100 0.1772 23.2% 2.4m\n", + " 110 0.1613 23.2% 2.4m\n", + " 120 0.2180 23.2% 2.3m\n", + " 130 0.1601 23.2% 2.3m\n", + " 140 0.1662 23.2% 2.3m\n", + " 150 0.1839 23.2% 2.2m\n", + " 160 0.1839 23.2% 2.2m\n", + " 170 0.1275 23.2% 2.2m\n", + " 180 0.1994 23.2% 2.1m\n", + " 190 0.1965 23.2% 2.1m\n", + " 200 0.1381 23.2% 2.1m\n", + " 210 0.1511 23.2% 2.0m\n", + " 220 0.1788 23.2% 2.0m\n", + " 230 0.1442 23.2% 2.0m\n", + " 240 0.1875 23.2% 2.0m\n", + " 250 0.1450 23.2% 1.9m\n", + " 260 0.1626 23.2% 1.9m\n", + " 270 0.1909 23.2% 1.9m\n", + " 280 0.1741 23.2% 1.8m\n", + " 290 0.1472 23.2% 1.8m\n", + " 300 0.1746 23.2% 1.8m\n", + " 310 0.1929 23.2% 1.7m\n", + " 320 0.1347 23.2% 1.7m\n", + " 330 0.1992 23.2% 1.7m\n", + " 340 0.1713 23.2% 1.6m\n", + " 350 0.1907 23.2% 1.6m\n", + " 360 0.1924 23.2% 1.6m\n", + " 370 0.1692 23.2% 1.5m\n", + " 380 0.2092 23.2% 1.5m\n", + " 390 0.1545 23.2% 1.5m\n", + " 400 0.1724 23.2% 1.5m\n", + " 410 0.1947 23.2% 1.4m\n", + " 420 0.1221 23.2% 1.4m\n", + " 430 0.1385 23.2% 1.4m\n", + " 440 0.1727 23.2% 1.3m\n", + " 450 0.1969 23.2% 1.3m\n", + " 460 0.1697 23.2% 1.3m\n", + " 470 0.1585 23.2% 1.2m\n", + " 480 0.1974 23.2% 1.2m\n", + " 490 0.1821 23.2% 1.2m\n", + " 500 0.1376 23.2% 1.1m\n", + " 510 0.2054 23.2% 1.1m\n", + " 520 0.1494 23.2% 1.1m\n", + " 530 0.1467 23.2% 1.0m\n", + " 540 0.1923 23.2% 1.0m\n", + " 550 0.1574 23.2% 1.0m\n", + " 560 0.1722 23.2% 0.9m\n", + " 570 0.1572 23.2% 0.9m\n", + " 580 0.2052 23.2% 0.9m\n", + " 590 0.1497 23.2% 0.9m\n", + " 600 0.1619 23.2% 0.8m\n", + " 610 0.1459 23.2% 0.8m\n", + " 620 0.2018 23.2% 0.8m\n", + " 630 0.1502 23.2% 0.7m\n", + " 640 0.1208 23.2% 0.7m\n", + " 650 0.1291 23.2% 0.7m\n", + " 660 0.1696 23.2% 0.6m\n", + " 670 0.1812 23.2% 0.6m\n", + " 680 0.1816 23.2% 0.6m\n", + " 690 0.1505 23.2% 0.5m\n", + " 700 0.1750 23.2% 0.5m\n", + " 710 0.1633 23.2% 0.5m\n", + " 720 0.1397 23.2% 0.4m\n", + " 730 0.1486 23.2% 0.4m\n", + " 740 0.1784 23.2% 0.4m\n", + " 750 0.1523 23.2% 0.4m\n", + " 760 0.1414 23.2% 0.3m\n", + " 770 0.1936 23.2% 0.3m\n", + " 780 0.1401 23.2% 0.3m\n", + " 790 0.2011 23.2% 0.2m\n", + " 800 0.1819 23.2% 0.2m\n", + " 810 0.1427 23.2% 0.2m\n", + " 820 0.1907 23.2% 0.1m\n", + " 830 0.1735 23.2% 0.1m\n", + " 840 0.1859 23.2% 0.1m\n", + " 850 0.1649 23.2% 0.0m\n", + " 860 0.1618 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 35 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 35 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16700 โ†’ Val=0.16439\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 184m\n", + "\n", + "๐Ÿ”„ Epoch 36/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1503 23.2% 2.7m\n", + " 10 0.1758 23.2% 2.6m\n", + " 20 0.1389 23.2% 2.6m\n", + " 30 0.2108 23.2% 2.6m\n", + " 40 0.1770 23.2% 2.5m\n", + " 50 0.1876 23.2% 2.5m\n", + " 60 0.1781 23.2% 2.5m\n", + " 70 0.1865 23.2% 2.5m\n", + " 80 0.1601 23.2% 2.4m\n", + " 90 0.1646 23.2% 2.4m\n", + " 100 0.1675 23.2% 2.4m\n", + " 110 0.1749 23.2% 2.3m\n", + " 120 0.1913 23.2% 2.3m\n", + " 130 0.1889 23.2% 2.3m\n", + " 140 0.1783 23.2% 2.3m\n", + " 150 0.1379 23.2% 2.2m\n", + " 160 0.1472 23.2% 2.2m\n", + " 170 0.1486 23.2% 2.2m\n", + " 180 0.1969 23.2% 2.1m\n", + " 190 0.1665 23.2% 2.1m\n", + " 200 0.1775 23.2% 2.1m\n", + " 210 0.2318 23.2% 2.0m\n", + " 220 0.1575 23.2% 2.0m\n", + " 230 0.1428 23.2% 2.0m\n", + " 240 0.1995 23.2% 2.0m\n", + " 250 0.1332 23.2% 1.9m\n", + " 260 0.1623 23.2% 1.9m\n", + " 270 0.1559 23.2% 1.9m\n", + " 280 0.1688 23.2% 1.8m\n", + " 290 0.1171 23.2% 1.8m\n", + " 300 0.1886 23.2% 1.8m\n", + " 310 0.1705 23.2% 1.7m\n", + " 320 0.1761 23.2% 1.7m\n", + " 330 0.1562 23.2% 1.7m\n", + " 340 0.1681 23.2% 1.6m\n", + " 350 0.1727 23.2% 1.6m\n", + " 360 0.1507 23.2% 1.6m\n", + " 370 0.1649 23.2% 1.5m\n", + " 380 0.1648 23.2% 1.5m\n", + " 390 0.1664 23.2% 1.5m\n", + " 400 0.1583 23.2% 1.4m\n", + " 410 0.1784 23.2% 1.4m\n", + " 420 0.1735 23.2% 1.4m\n", + " 430 0.1471 23.2% 1.4m\n", + " 440 0.1695 23.2% 1.3m\n", + " 450 0.1707 23.2% 1.3m\n", + " 460 0.1540 23.2% 1.3m\n", + " 470 0.1441 23.2% 1.2m\n", + " 480 0.1601 23.2% 1.2m\n", + " 490 0.1837 23.2% 1.2m\n", + " 500 0.1637 23.2% 1.1m\n", + " 510 0.1357 23.2% 1.1m\n", + " 520 0.1835 23.2% 1.1m\n", + " 530 0.1472 23.2% 1.0m\n", + " 540 0.1429 23.2% 1.0m\n", + " 550 0.1336 23.2% 1.0m\n", + " 560 0.1770 23.2% 0.9m\n", + " 570 0.1220 23.2% 0.9m\n", + " 580 0.1749 23.2% 0.9m\n", + " 590 0.1978 23.2% 0.9m\n", + " 600 0.1613 23.2% 0.8m\n", + " 610 0.1473 23.2% 0.8m\n", + " 620 0.2704 23.2% 0.8m\n", + " 630 0.1798 23.2% 0.7m\n", + " 640 0.1548 23.2% 0.7m\n", + " 650 0.1341 23.2% 0.7m\n", + " 660 0.1921 23.2% 0.6m\n", + " 670 0.1383 23.2% 0.6m\n", + " 680 0.2126 23.2% 0.6m\n", + " 690 0.1554 23.2% 0.5m\n", + " 700 0.1461 23.2% 0.5m\n", + " 710 0.1979 23.2% 0.5m\n", + " 720 0.1930 23.2% 0.4m\n", + " 730 0.1674 23.2% 0.4m\n", + " 740 0.1532 23.2% 0.4m\n", + " 750 0.1669 23.2% 0.4m\n", + " 760 0.1363 23.2% 0.3m\n", + " 770 0.1627 23.2% 0.3m\n", + " 780 0.1594 23.2% 0.3m\n", + " 790 0.1426 23.2% 0.2m\n", + " 800 0.1865 23.2% 0.2m\n", + " 810 0.2133 23.2% 0.2m\n", + " 820 0.1452 23.2% 0.1m\n", + " 830 0.1660 23.2% 0.1m\n", + " 840 0.1772 23.2% 0.1m\n", + " 850 0.1710 23.2% 0.0m\n", + " 860 0.1762 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 36 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 36 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16702 โ†’ Val=0.16421\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 181m\n", + "\n", + "๐Ÿ”„ Epoch 37/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1706 23.2% 2.7m\n", + " 10 0.1645 23.2% 2.6m\n", + " 20 0.1520 23.2% 2.6m\n", + " 30 0.1665 23.2% 2.6m\n", + " 40 0.1825 23.2% 2.6m\n", + " 50 0.1990 23.2% 2.5m\n", + " 60 0.1491 23.2% 2.5m\n", + " 70 0.1962 23.2% 2.5m\n", + " 80 0.1559 23.2% 2.4m\n", + " 90 0.1875 23.2% 2.4m\n", + " 100 0.1821 23.2% 2.4m\n", + " 110 0.1841 23.2% 2.3m\n", + " 120 0.1521 23.2% 2.3m\n", + " 130 0.1565 23.2% 2.3m\n", + " 140 0.1701 23.2% 2.2m\n", + " 150 0.1283 23.2% 2.2m\n", + " 160 0.1390 23.2% 2.2m\n", + " 170 0.1865 23.2% 2.2m\n", + " 180 0.1596 23.2% 2.1m\n", + " 190 0.1405 23.2% 2.1m\n", + " 200 0.1445 23.2% 2.1m\n", + " 210 0.1641 23.2% 2.0m\n", + " 220 0.2201 23.2% 2.0m\n", + " 230 0.1568 23.2% 2.0m\n", + " 240 0.1572 23.2% 1.9m\n", + " 250 0.1796 23.2% 1.9m\n", + " 260 0.1842 23.2% 1.9m\n", + " 270 0.1649 23.2% 1.8m\n", + " 280 0.2124 23.2% 1.8m\n", + " 290 0.1835 23.2% 1.8m\n", + " 300 0.1804 23.2% 1.8m\n", + " 310 0.2051 23.2% 1.7m\n", + " 320 0.2081 23.2% 1.7m\n", + " 330 0.1695 23.2% 1.7m\n", + " 340 0.1616 23.2% 1.6m\n", + " 350 0.1713 23.2% 1.6m\n", + " 360 0.2001 23.2% 1.6m\n", + " 370 0.1607 23.2% 1.5m\n", + " 380 0.1801 23.2% 1.5m\n", + " 390 0.1573 23.2% 1.5m\n", + " 400 0.1241 23.2% 1.4m\n", + " 410 0.1523 23.2% 1.4m\n", + " 420 0.1474 23.2% 1.4m\n", + " 430 0.1557 23.2% 1.3m\n", + " 440 0.1524 23.2% 1.3m\n", + " 450 0.1584 23.2% 1.3m\n", + " 460 0.1814 23.2% 1.3m\n", + " 470 0.1960 23.2% 1.2m\n", + " 480 0.1608 23.2% 1.2m\n", + " 490 0.1603 23.2% 1.2m\n", + " 500 0.1914 23.2% 1.1m\n", + " 510 0.1668 23.2% 1.1m\n", + " 520 0.1557 23.2% 1.1m\n", + " 530 0.1495 23.2% 1.0m\n", + " 540 0.1681 23.2% 1.0m\n", + " 550 0.1525 23.2% 1.0m\n", + " 560 0.1179 23.2% 0.9m\n", + " 570 0.1781 23.2% 0.9m\n", + " 580 0.1900 23.2% 0.9m\n", + " 590 0.1336 23.2% 0.8m\n", + " 600 0.1740 23.2% 0.8m\n", + " 610 0.1306 23.2% 0.8m\n", + " 620 0.1636 23.2% 0.8m\n", + " 630 0.1578 23.2% 0.7m\n", + " 640 0.1809 23.2% 0.7m\n", + " 650 0.1802 23.2% 0.7m\n", + " 660 0.1700 23.2% 0.6m\n", + " 670 0.1651 23.2% 0.6m\n", + " 680 0.1644 23.2% 0.6m\n", + " 690 0.1129 23.2% 0.5m\n", + " 700 0.1515 23.2% 0.5m\n", + " 710 0.1913 23.2% 0.5m\n", + " 720 0.1681 23.2% 0.4m\n", + " 730 0.1638 23.2% 0.4m\n", + " 740 0.1408 23.2% 0.4m\n", + " 750 0.1558 23.2% 0.4m\n", + " 760 0.1494 23.2% 0.3m\n", + " 770 0.1619 23.2% 0.3m\n", + " 780 0.1487 23.2% 0.3m\n", + " 790 0.1684 23.2% 0.2m\n", + " 800 0.1584 23.2% 0.2m\n", + " 810 0.1923 23.2% 0.2m\n", + " 820 0.1372 23.2% 0.1m\n", + " 830 0.1515 23.2% 0.1m\n", + " 840 0.1317 23.2% 0.1m\n", + " 850 0.2047 23.2% 0.0m\n", + " 860 0.1861 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 37 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 37 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16692 โ†’ Val=0.16431\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โš ๏ธ No improvement for 10/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 177m\n", + "\n", + "๐Ÿ”„ Epoch 38/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1610 23.2% 2.6m\n", + " 10 0.1731 23.2% 2.6m\n", + " 20 0.1689 23.2% 2.6m\n", + " 30 0.1612 23.2% 2.6m\n", + " 40 0.1084 23.2% 2.6m\n", + " 50 0.1877 23.2% 2.5m\n", + " 60 0.1289 23.2% 2.5m\n", + " 70 0.1813 23.2% 2.5m\n", + " 80 0.1665 23.2% 2.4m\n", + " 90 0.1562 23.2% 2.4m\n", + " 100 0.1460 23.2% 2.4m\n", + " 110 0.1509 23.2% 2.4m\n", + " 120 0.1649 23.2% 2.3m\n", + " 130 0.1951 23.2% 2.3m\n", + " 140 0.1546 23.2% 2.3m\n", + " 150 0.1636 23.2% 2.2m\n", + " 160 0.1670 23.2% 2.2m\n", + " 170 0.1738 23.2% 2.2m\n", + " 180 0.1931 23.2% 2.1m\n", + " 190 0.1546 23.2% 2.1m\n", + " 200 0.1626 23.2% 2.1m\n", + " 210 0.1630 23.2% 2.0m\n", + " 220 0.1675 23.2% 2.0m\n", + " 230 0.1555 23.2% 2.0m\n", + " 240 0.1249 23.2% 2.0m\n", + " 250 0.1239 23.2% 1.9m\n", + " 260 0.1627 23.2% 1.9m\n", + " 270 0.1737 23.2% 1.9m\n", + " 280 0.1546 23.2% 1.8m\n", + " 290 0.1708 23.2% 1.8m\n", + " 300 0.1316 23.2% 1.8m\n", + " 310 0.1936 23.2% 1.7m\n", + " 320 0.1823 23.2% 1.7m\n", + " 330 0.1694 23.2% 1.7m\n", + " 340 0.1374 23.2% 1.6m\n", + " 350 0.1551 23.2% 1.6m\n", + " 360 0.1770 23.2% 1.6m\n", + " 370 0.1803 23.2% 1.5m\n", + " 380 0.1637 23.2% 1.5m\n", + " 390 0.1543 23.2% 1.5m\n", + " 400 0.1821 23.2% 1.4m\n", + " 410 0.1233 23.2% 1.4m\n", + " 420 0.2193 23.2% 1.4m\n", + " 430 0.1827 23.2% 1.4m\n", + " 440 0.1886 23.2% 1.3m\n", + " 450 0.1644 23.2% 1.3m\n", + " 460 0.1462 23.2% 1.3m\n", + " 470 0.1625 23.2% 1.2m\n", + " 480 0.1679 23.2% 1.2m\n", + " 490 0.1396 23.2% 1.2m\n", + " 500 0.1619 23.2% 1.1m\n", + " 510 0.1391 23.2% 1.1m\n", + " 520 0.1885 23.2% 1.1m\n", + " 530 0.1656 23.2% 1.0m\n", + " 540 0.1684 23.2% 1.0m\n", + " 550 0.1763 23.2% 1.0m\n", + " 560 0.1450 23.2% 0.9m\n", + " 570 0.1745 23.2% 0.9m\n", + " 580 0.1856 23.2% 0.9m\n", + " 590 0.1711 23.2% 0.9m\n", + " 600 0.1538 23.2% 0.8m\n", + " 610 0.1923 23.2% 0.8m\n", + " 620 0.1830 23.2% 0.8m\n", + " 630 0.1398 23.2% 0.7m\n", + " 640 0.1681 23.2% 0.7m\n", + " 650 0.1357 23.2% 0.7m\n", + " 660 0.1942 23.2% 0.6m\n", + " 670 0.1669 23.2% 0.6m\n", + " 680 0.1381 23.2% 0.6m\n", + " 690 0.1593 23.2% 0.5m\n", + " 700 0.1911 23.2% 0.5m\n", + " 710 0.1694 23.2% 0.5m\n", + " 720 0.1697 23.2% 0.4m\n", + " 730 0.1632 23.2% 0.4m\n", + " 740 0.1721 23.2% 0.4m\n", + " 750 0.1641 23.2% 0.4m\n", + " 760 0.1688 23.2% 0.3m\n", + " 770 0.1876 23.2% 0.3m\n", + " 780 0.1679 23.2% 0.3m\n", + " 790 0.1840 23.2% 0.2m\n", + " 800 0.1961 23.2% 0.2m\n", + " 810 0.1589 23.2% 0.2m\n", + " 820 0.1970 23.2% 0.1m\n", + " 830 0.1510 23.2% 0.1m\n", + " 840 0.1870 23.2% 0.1m\n", + " 850 0.1719 23.2% 0.0m\n", + " 860 0.1533 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 38 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 38 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16682 โ†’ Val=0.16402\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.8% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 175m\n", + "\n", + "๐Ÿ”„ Epoch 39/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1468 23.2% 3.2m\n", + " 10 0.1856 23.2% 2.8m\n", + " 20 0.1361 23.2% 2.7m\n", + " 30 0.1883 23.2% 2.6m\n", + " 40 0.1568 23.2% 2.6m\n", + " 50 0.1638 23.2% 2.5m\n", + " 60 0.1224 23.2% 2.5m\n", + " 70 0.1709 23.2% 2.5m\n", + " 80 0.2056 23.2% 2.4m\n", + " 90 0.1162 23.2% 2.4m\n", + " 100 0.1849 23.2% 2.4m\n", + " 110 0.1438 23.2% 2.3m\n", + " 120 0.1624 23.2% 2.3m\n", + " 130 0.2199 23.2% 2.3m\n", + " 140 0.1825 23.2% 2.3m\n", + " 150 0.1438 23.2% 2.2m\n", + " 160 0.1859 23.2% 2.2m\n", + " 170 0.1695 23.2% 2.2m\n", + " 180 0.1807 23.2% 2.1m\n", + " 190 0.1581 23.2% 2.1m\n", + " 200 0.1397 23.2% 2.1m\n", + " 210 0.1466 23.2% 2.0m\n", + " 220 0.1769 23.2% 2.0m\n", + " 230 0.1521 23.2% 2.0m\n", + " 240 0.1831 23.2% 1.9m\n", + " 250 0.1131 23.2% 1.9m\n", + " 260 0.1622 23.2% 1.9m\n", + " 270 0.1378 23.2% 1.9m\n", + " 280 0.1651 23.2% 1.8m\n", + " 290 0.1702 23.2% 1.8m\n", + " 300 0.2007 23.2% 1.8m\n", + " 310 0.1663 23.2% 1.7m\n", + " 320 0.1831 23.2% 1.7m\n", + " 330 0.1994 23.2% 1.7m\n", + " 340 0.1221 23.2% 1.6m\n", + " 350 0.1536 23.2% 1.6m\n", + " 360 0.1460 23.2% 1.6m\n", + " 370 0.1540 23.2% 1.5m\n", + " 380 0.2180 23.2% 1.5m\n", + " 390 0.1860 23.2% 1.5m\n", + " 400 0.1590 23.2% 1.4m\n", + " 410 0.1601 23.2% 1.4m\n", + " 420 0.1886 23.2% 1.4m\n", + " 430 0.1572 23.2% 1.4m\n", + " 440 0.1710 23.2% 1.3m\n", + " 450 0.1729 23.2% 1.3m\n", + " 460 0.1673 23.2% 1.3m\n", + " 470 0.1369 23.2% 1.2m\n", + " 480 0.1674 23.2% 1.2m\n", + " 490 0.1749 23.2% 1.2m\n", + " 500 0.1687 23.2% 1.1m\n", + " 510 0.1468 23.2% 1.1m\n", + " 520 0.1674 23.2% 1.1m\n", + " 530 0.1952 23.2% 1.0m\n", + " 540 0.1784 23.2% 1.0m\n", + " 550 0.2042 23.2% 1.0m\n", + " 560 0.1700 23.2% 0.9m\n", + " 570 0.1989 23.2% 0.9m\n", + " 580 0.2158 23.2% 0.9m\n", + " 590 0.2061 23.2% 0.9m\n", + " 600 0.1719 23.2% 0.8m\n", + " 610 0.1523 23.2% 0.8m\n", + " 620 0.1857 23.2% 0.8m\n", + " 630 0.1742 23.2% 0.7m\n", + " 640 0.1452 23.2% 0.7m\n", + " 650 0.1815 23.2% 0.7m\n", + " 660 0.1966 23.2% 0.6m\n", + " 670 0.1876 23.2% 0.6m\n", + " 680 0.1428 23.2% 0.6m\n", + " 690 0.1370 23.2% 0.5m\n", + " 700 0.1350 23.2% 0.5m\n", + " 710 0.1782 23.2% 0.5m\n", + " 720 0.1707 23.2% 0.4m\n", + " 730 0.1765 23.2% 0.4m\n", + " 740 0.1533 23.2% 0.4m\n", + " 750 0.1931 23.2% 0.4m\n", + " 760 0.1832 23.2% 0.3m\n", + " 770 0.1797 23.2% 0.3m\n", + " 780 0.1925 23.2% 0.3m\n", + " 790 0.1681 23.2% 0.2m\n", + " 800 0.1729 23.2% 0.2m\n", + " 810 0.1612 23.2% 0.2m\n", + " 820 0.1684 23.2% 0.1m\n", + " 830 0.1488 23.2% 0.1m\n", + " 840 0.1174 23.2% 0.1m\n", + " 850 0.1946 23.2% 0.0m\n", + " 860 0.1723 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 39 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 39 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16677 โ†’ Val=0.16444\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 173m\n", + "\n", + "๐Ÿ”„ Epoch 40/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1648 23.2% 2.7m\n", + " 10 0.1650 23.2% 2.7m\n", + " 20 0.1469 23.2% 2.7m\n", + " 30 0.1505 23.2% 2.7m\n", + " 40 0.1371 23.2% 2.6m\n", + " 50 0.1731 23.2% 2.6m\n", + " 60 0.1252 23.2% 2.6m\n", + " 70 0.1745 23.2% 2.5m\n", + " 80 0.1404 23.2% 2.5m\n", + " 90 0.1827 23.2% 2.4m\n", + " 100 0.1932 23.2% 2.4m\n", + " 110 0.1426 23.2% 2.4m\n", + " 120 0.1643 23.2% 2.3m\n", + " 130 0.2053 23.2% 2.3m\n", + " 140 0.1729 23.2% 2.3m\n", + " 150 0.1834 23.2% 2.2m\n", + " 160 0.1736 23.2% 2.2m\n", + " 170 0.1357 23.2% 2.2m\n", + " 180 0.1564 23.2% 2.1m\n", + " 190 0.1615 23.2% 2.1m\n", + " 200 0.1961 23.2% 2.1m\n", + " 210 0.2009 23.2% 2.1m\n", + " 220 0.1319 23.2% 2.0m\n", + " 230 0.1593 23.2% 2.0m\n", + " 240 0.1426 23.2% 2.0m\n", + " 250 0.1546 23.2% 1.9m\n", + " 260 0.1718 23.2% 1.9m\n", + " 270 0.1422 23.2% 1.9m\n", + " 280 0.1678 23.2% 1.8m\n", + " 290 0.1966 23.2% 1.8m\n", + " 300 0.1761 23.2% 1.8m\n", + " 310 0.1464 23.2% 1.7m\n", + " 320 0.1388 23.2% 1.7m\n", + " 330 0.1516 23.2% 1.7m\n", + " 340 0.1792 23.2% 1.6m\n", + " 350 0.1747 23.2% 1.6m\n", + " 360 0.1901 23.2% 1.6m\n", + " 370 0.1968 23.2% 1.5m\n", + " 380 0.1757 23.2% 1.5m\n", + " 390 0.1735 23.2% 1.5m\n", + " 400 0.1602 23.2% 1.5m\n", + " 410 0.1836 23.2% 1.4m\n", + " 420 0.1591 23.2% 1.4m\n", + " 430 0.1954 23.2% 1.4m\n", + " 440 0.1476 23.2% 1.3m\n", + " 450 0.1558 23.2% 1.3m\n", + " 460 0.1661 23.2% 1.3m\n", + " 470 0.1587 23.2% 1.2m\n", + " 480 0.1865 23.2% 1.2m\n", + " 490 0.1165 23.2% 1.2m\n", + " 500 0.2042 23.2% 1.1m\n", + " 510 0.1683 23.2% 1.1m\n", + " 520 0.1638 23.2% 1.1m\n", + " 530 0.1926 23.2% 1.0m\n", + " 540 0.1992 23.2% 1.0m\n", + " 550 0.1782 23.2% 1.0m\n", + " 560 0.1856 23.2% 1.0m\n", + " 570 0.1972 23.2% 0.9m\n", + " 580 0.1756 23.2% 0.9m\n", + " 590 0.1665 23.2% 0.9m\n", + " 600 0.1505 23.2% 0.8m\n", + " 610 0.1343 23.2% 0.8m\n", + " 620 0.1516 23.2% 0.8m\n", + " 630 0.1914 23.2% 0.7m\n", + " 640 0.1346 23.2% 0.7m\n", + " 650 0.1594 23.2% 0.7m\n", + " 660 0.1463 23.2% 0.6m\n", + " 670 0.1614 23.2% 0.6m\n", + " 680 0.1772 23.2% 0.6m\n", + " 690 0.1977 23.2% 0.5m\n", + " 700 0.1681 23.2% 0.5m\n", + " 710 0.1790 23.2% 0.5m\n", + " 720 0.1281 23.2% 0.4m\n", + " 730 0.1834 23.2% 0.4m\n", + " 740 0.1872 23.2% 0.4m\n", + " 750 0.1744 23.2% 0.4m\n", + " 760 0.1663 23.2% 0.3m\n", + " 770 0.1267 23.2% 0.3m\n", + " 780 0.1806 23.2% 0.3m\n", + " 790 0.1799 23.2% 0.2m\n", + " 800 0.1750 23.2% 0.2m\n", + " 810 0.1865 23.2% 0.2m\n", + " 820 0.1704 23.2% 0.1m\n", + " 830 0.2173 23.2% 0.1m\n", + " 840 0.1551 23.2% 0.1m\n", + " 850 0.1358 23.2% 0.0m\n", + " 860 0.1361 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 40 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 40 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16691 โ†’ Val=0.16423\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 171m\n", + "\n", + "๐Ÿ”„ Epoch 41/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1245 23.2% 2.8m\n", + " 10 0.1597 23.2% 2.7m\n", + " 20 0.1572 23.2% 2.7m\n", + " 30 0.1726 23.2% 2.6m\n", + " 40 0.1440 23.2% 2.6m\n", + " 50 0.1656 23.2% 2.6m\n", + " 60 0.1773 23.2% 2.5m\n", + " 70 0.1552 23.2% 2.5m\n", + " 80 0.1472 23.2% 2.5m\n", + " 90 0.1701 23.2% 2.4m\n", + " 100 0.1590 23.2% 2.4m\n", + " 110 0.2005 23.2% 2.4m\n", + " 120 0.1670 23.2% 2.3m\n", + " 130 0.2091 23.2% 2.3m\n", + " 140 0.1552 23.2% 2.3m\n", + " 150 0.1826 23.2% 2.3m\n", + " 160 0.1691 23.2% 2.2m\n", + " 170 0.1942 23.2% 2.2m\n", + " 180 0.1430 23.2% 2.2m\n", + " 190 0.1677 23.2% 2.1m\n", + " 200 0.1446 23.2% 2.1m\n", + " 210 0.1682 23.2% 2.1m\n", + " 220 0.1911 23.2% 2.0m\n", + " 230 0.1861 23.2% 2.0m\n", + " 240 0.1569 23.2% 2.0m\n", + " 250 0.1540 23.2% 1.9m\n", + " 260 0.1258 23.2% 1.9m\n", + " 270 0.1690 23.2% 1.9m\n", + " 280 0.1837 23.2% 1.8m\n", + " 290 0.1479 23.2% 1.8m\n", + " 300 0.2147 23.2% 1.8m\n", + " 310 0.1657 23.2% 1.7m\n", + " 320 0.2066 23.2% 1.7m\n", + " 330 0.1436 23.2% 1.7m\n", + " 340 0.1789 23.2% 1.6m\n", + " 350 0.1691 23.2% 1.6m\n", + " 360 0.1602 23.2% 1.6m\n", + " 370 0.1887 23.2% 1.6m\n", + " 380 0.1561 23.2% 1.5m\n", + " 390 0.1846 23.2% 1.5m\n", + " 400 0.1880 23.2% 1.5m\n", + " 410 0.1791 23.2% 1.4m\n", + " 420 0.1569 23.2% 1.4m\n", + " 430 0.1901 23.2% 1.4m\n", + " 440 0.1992 23.2% 1.3m\n", + " 450 0.1828 23.2% 1.3m\n", + " 460 0.1745 23.2% 1.3m\n", + " 470 0.1832 23.2% 1.2m\n", + " 480 0.1930 23.2% 1.2m\n", + " 490 0.1610 23.2% 1.2m\n", + " 500 0.2084 23.2% 1.1m\n", + " 510 0.1788 23.2% 1.1m\n", + " 520 0.1541 23.2% 1.1m\n", + " 530 0.1925 23.2% 1.0m\n", + " 540 0.1361 23.2% 1.0m\n", + " 550 0.1828 23.2% 1.0m\n", + " 560 0.1495 23.2% 1.0m\n", + " 570 0.1278 23.2% 0.9m\n", + " 580 0.1650 23.2% 0.9m\n", + " 590 0.1580 23.2% 0.9m\n", + " 600 0.1770 23.2% 0.8m\n", + " 610 0.1635 23.2% 0.8m\n", + " 620 0.1806 23.2% 0.8m\n", + " 630 0.1533 23.2% 0.7m\n", + " 640 0.1363 23.2% 0.7m\n", + " 650 0.1517 23.2% 0.7m\n", + " 660 0.1982 23.2% 0.6m\n", + " 670 0.1789 23.2% 0.6m\n", + " 680 0.1786 23.2% 0.6m\n", + " 690 0.1898 23.2% 0.5m\n", + " 700 0.2053 23.2% 0.5m\n", + " 710 0.1730 23.2% 0.5m\n", + " 720 0.1654 23.2% 0.5m\n", + " 730 0.1686 23.2% 0.4m\n", + " 740 0.1626 23.2% 0.4m\n", + " 750 0.1723 23.2% 0.4m\n", + " 760 0.1467 23.2% 0.3m\n", + " 770 0.1835 23.2% 0.3m\n", + " 780 0.1135 23.2% 0.3m\n", + " 790 0.1701 23.2% 0.2m\n", + " 800 0.1864 23.2% 0.2m\n", + " 810 0.1312 23.2% 0.2m\n", + " 820 0.1640 23.2% 0.1m\n", + " 830 0.1618 23.2% 0.1m\n", + " 840 0.2036 23.2% 0.1m\n", + " 850 0.1900 23.2% 0.0m\n", + " 860 0.1385 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 41 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 41 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16696 โ†’ Val=0.16421\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 168m\n", + "\n", + "๐Ÿ”„ Epoch 42/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1692 23.2% 2.8m\n", + " 10 0.1681 23.2% 2.7m\n", + " 20 0.1644 23.2% 2.7m\n", + " 30 0.1210 23.2% 2.6m\n", + " 40 0.1972 23.2% 2.6m\n", + " 50 0.1769 23.2% 2.6m\n", + " 60 0.1355 23.2% 2.5m\n", + " 70 0.1693 23.2% 2.5m\n", + " 80 0.1827 23.2% 2.5m\n", + " 90 0.1844 23.2% 2.4m\n", + " 100 0.1781 23.2% 2.4m\n", + " 110 0.1794 23.2% 2.4m\n", + " 120 0.1803 23.2% 2.3m\n", + " 130 0.1733 23.2% 2.3m\n", + " 140 0.1765 23.2% 2.3m\n", + " 150 0.1523 23.2% 2.2m\n", + " 160 0.1449 23.2% 2.2m\n", + " 170 0.1799 23.2% 2.2m\n", + " 180 0.1441 23.2% 2.1m\n", + " 190 0.1941 23.2% 2.1m\n", + " 200 0.1648 23.2% 2.1m\n", + " 210 0.1615 23.2% 2.0m\n", + " 220 0.1859 23.2% 2.0m\n", + " 230 0.1332 23.2% 2.0m\n", + " 240 0.1487 23.2% 1.9m\n", + " 250 0.1362 23.2% 1.9m\n", + " 260 0.1708 23.2% 1.9m\n", + " 270 0.1744 23.2% 1.8m\n", + " 280 0.1820 23.2% 1.8m\n", + " 290 0.1350 23.2% 1.8m\n", + " 300 0.1725 23.2% 1.8m\n", + " 310 0.1882 23.2% 1.7m\n", + " 320 0.1723 23.2% 1.7m\n", + " 330 0.1493 23.2% 1.7m\n", + " 340 0.1738 23.2% 1.6m\n", + " 350 0.1313 23.2% 1.6m\n", + " 360 0.1833 23.2% 1.6m\n", + " 370 0.1440 23.2% 1.5m\n", + " 380 0.1536 23.2% 1.5m\n", + " 390 0.1826 23.2% 1.5m\n", + " 400 0.1402 23.2% 1.4m\n", + " 410 0.1874 23.2% 1.4m\n", + " 420 0.2039 23.2% 1.4m\n", + " 430 0.1655 23.2% 1.4m\n", + " 440 0.1430 23.2% 1.3m\n", + " 450 0.1744 23.2% 1.3m\n", + " 460 0.1773 23.2% 1.3m\n", + " 470 0.1586 23.2% 1.2m\n", + " 480 0.1774 23.2% 1.2m\n", + " 490 0.1609 23.2% 1.2m\n", + " 500 0.2073 23.2% 1.1m\n", + " 510 0.1273 23.2% 1.1m\n", + " 520 0.1694 23.2% 1.1m\n", + " 530 0.1879 23.2% 1.0m\n", + " 540 0.1314 23.2% 1.0m\n", + " 550 0.1549 23.2% 1.0m\n", + " 560 0.1433 23.2% 0.9m\n", + " 570 0.1906 23.2% 0.9m\n", + " 580 0.1846 23.2% 0.9m\n", + " 590 0.1823 23.2% 0.9m\n", + " 600 0.1594 23.2% 0.8m\n", + " 610 0.1732 23.2% 0.8m\n", + " 620 0.1766 23.2% 0.8m\n", + " 630 0.1711 23.2% 0.7m\n", + " 640 0.1438 23.2% 0.7m\n", + " 650 0.1642 23.2% 0.7m\n", + " 660 0.1697 23.2% 0.6m\n", + " 670 0.1620 23.2% 0.6m\n", + " 680 0.1564 23.2% 0.6m\n", + " 690 0.1709 23.2% 0.5m\n", + " 700 0.1805 23.2% 0.5m\n", + " 710 0.1813 23.2% 0.5m\n", + " 720 0.1855 23.2% 0.4m\n", + " 730 0.2039 23.2% 0.4m\n", + " 740 0.1759 23.2% 0.4m\n", + " 750 0.1312 23.2% 0.4m\n", + " 760 0.1674 23.2% 0.3m\n", + " 770 0.1422 23.2% 0.3m\n", + " 780 0.1878 23.2% 0.3m\n", + " 790 0.1726 23.2% 0.2m\n", + " 800 0.1938 23.2% 0.2m\n", + " 810 0.1532 23.2% 0.2m\n", + " 820 0.1755 23.2% 0.1m\n", + " 830 0.1692 23.2% 0.1m\n", + " 840 0.1706 23.2% 0.1m\n", + " 850 0.1595 23.2% 0.0m\n", + " 860 0.1691 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 42 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 42 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16688 โ†’ Val=0.16426\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โš ๏ธ No improvement for 15/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 164m\n", + "\n", + "๐Ÿ”„ Epoch 43/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1739 23.2% 2.8m\n", + " 10 0.1641 23.2% 2.6m\n", + " 20 0.1649 23.2% 2.6m\n", + " 30 0.1805 23.2% 2.6m\n", + " 40 0.1427 23.2% 2.6m\n", + " 50 0.1830 23.2% 2.5m\n", + " 60 0.1719 23.2% 2.5m\n", + " 70 0.1411 23.2% 2.5m\n", + " 80 0.1611 23.2% 2.5m\n", + " 90 0.1701 23.2% 2.4m\n", + " 100 0.1508 23.2% 2.4m\n", + " 110 0.1775 23.2% 2.4m\n", + " 120 0.1615 23.2% 2.3m\n", + " 130 0.1482 23.2% 2.3m\n", + " 140 0.1524 23.2% 2.3m\n", + " 150 0.1654 23.2% 2.2m\n", + " 160 0.1544 23.2% 2.2m\n", + " 170 0.2314 23.2% 2.2m\n", + " 180 0.1476 23.2% 2.2m\n", + " 190 0.1584 23.2% 2.1m\n", + " 200 0.1698 23.2% 2.1m\n", + " 210 0.1514 23.2% 2.1m\n", + " 220 0.1775 23.2% 2.0m\n", + " 230 0.1620 23.2% 2.0m\n", + " 240 0.1631 23.2% 2.0m\n", + " 250 0.1603 23.2% 1.9m\n", + " 260 0.2000 23.2% 1.9m\n", + " 270 0.1390 23.2% 1.9m\n", + " 280 0.1540 23.2% 1.8m\n", + " 290 0.2018 23.2% 1.8m\n", + " 300 0.1582 23.2% 1.8m\n", + " 310 0.1632 23.2% 1.7m\n", + " 320 0.1706 23.2% 1.7m\n", + " 330 0.1777 23.2% 1.7m\n", + " 340 0.1490 23.2% 1.6m\n", + " 350 0.1536 23.2% 1.6m\n", + " 360 0.1801 23.2% 1.6m\n", + " 370 0.1614 23.2% 1.5m\n", + " 380 0.1082 23.2% 1.5m\n", + " 390 0.1832 23.2% 1.5m\n", + " 400 0.1497 23.2% 1.5m\n", + " 410 0.1563 23.2% 1.4m\n", + " 420 0.2119 23.2% 1.4m\n", + " 430 0.1633 23.2% 1.4m\n", + " 440 0.1938 23.2% 1.3m\n", + " 450 0.1635 23.2% 1.3m\n", + " 460 0.1394 23.2% 1.3m\n", + " 470 0.1419 23.2% 1.2m\n", + " 480 0.1523 23.2% 1.2m\n", + " 490 0.1801 23.2% 1.2m\n", + " 500 0.1797 23.2% 1.1m\n", + " 510 0.1601 23.2% 1.1m\n", + " 520 0.1475 23.2% 1.1m\n", + " 530 0.1556 23.2% 1.0m\n", + " 540 0.1316 23.2% 1.0m\n", + " 550 0.1541 23.2% 1.0m\n", + " 560 0.1748 23.2% 1.0m\n", + " 570 0.1180 23.2% 0.9m\n", + " 580 0.2223 23.2% 0.9m\n", + " 590 0.2015 23.2% 0.9m\n", + " 600 0.1409 23.2% 0.8m\n", + " 610 0.1549 23.2% 0.8m\n", + " 620 0.1572 23.2% 0.8m\n", + " 630 0.1955 23.2% 0.7m\n", + " 640 0.1611 23.2% 0.7m\n", + " 650 0.1494 23.2% 0.7m\n", + " 660 0.1729 23.2% 0.6m\n", + " 670 0.1823 23.2% 0.6m\n", + " 680 0.1424 23.2% 0.6m\n", + " 690 0.1949 23.2% 0.5m\n", + " 700 0.2105 23.2% 0.5m\n", + " 710 0.1817 23.2% 0.5m\n", + " 720 0.1381 23.2% 0.4m\n", + " 730 0.1678 23.2% 0.4m\n", + " 740 0.1284 23.2% 0.4m\n", + " 750 0.1766 23.2% 0.4m\n", + " 760 0.2068 23.2% 0.3m\n", + " 770 0.1806 23.2% 0.3m\n", + " 780 0.1773 23.2% 0.3m\n", + " 790 0.1678 23.2% 0.2m\n", + " 800 0.1927 23.2% 0.2m\n", + " 810 0.1807 23.2% 0.2m\n", + " 820 0.1320 23.2% 0.1m\n", + " 830 0.1499 23.2% 0.1m\n", + " 840 0.1883 23.2% 0.1m\n", + " 850 0.1587 23.2% 0.0m\n", + " 860 0.1556 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 43 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 43 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16678 โ†’ Val=0.16424\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 162m\n", + "\n", + "๐Ÿ”„ Epoch 44/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1480 23.2% 2.6m\n", + " 10 0.1603 23.2% 2.6m\n", + " 20 0.1662 23.2% 2.6m\n", + " 30 0.1517 23.2% 2.6m\n", + " 40 0.1760 23.2% 2.6m\n", + " 50 0.1919 23.2% 2.5m\n", + " 60 0.1615 23.2% 2.5m\n", + " 70 0.2222 23.2% 2.5m\n", + " 80 0.1815 23.2% 2.5m\n", + " 90 0.2182 23.2% 2.4m\n", + " 100 0.1501 23.2% 2.4m\n", + " 110 0.2107 23.2% 2.4m\n", + " 120 0.1732 23.2% 2.3m\n", + " 130 0.1297 23.2% 2.3m\n", + " 140 0.1664 23.2% 2.3m\n", + " 150 0.1818 23.2% 2.2m\n", + " 160 0.1887 23.2% 2.2m\n", + " 170 0.1489 23.2% 2.2m\n", + " 180 0.1832 23.2% 2.1m\n", + " 190 0.1635 23.2% 2.1m\n", + " 200 0.1343 23.2% 2.1m\n", + " 210 0.1769 23.2% 2.0m\n", + " 220 0.1578 23.2% 2.0m\n", + " 230 0.1395 23.2% 2.0m\n", + " 240 0.1665 23.2% 2.0m\n", + " 250 0.1752 23.2% 1.9m\n", + " 260 0.1547 23.2% 1.9m\n", + " 270 0.1784 23.2% 1.9m\n", + " 280 0.1522 23.2% 1.8m\n", + " 290 0.1666 23.2% 1.8m\n", + " 300 0.1419 23.2% 1.8m\n", + " 310 0.1839 23.2% 1.7m\n", + " 320 0.1547 23.2% 1.7m\n", + " 330 0.1442 23.2% 1.7m\n", + " 340 0.1604 23.2% 1.6m\n", + " 350 0.1765 23.2% 1.6m\n", + " 360 0.1743 23.2% 1.6m\n", + " 370 0.1711 23.2% 1.5m\n", + " 380 0.1624 23.2% 1.5m\n", + " 390 0.1647 23.2% 1.5m\n", + " 400 0.1478 23.2% 1.5m\n", + " 410 0.1914 23.2% 1.4m\n", + " 420 0.1552 23.2% 1.4m\n", + " 430 0.1808 23.2% 1.4m\n", + " 440 0.1322 23.2% 1.3m\n", + " 450 0.1225 23.2% 1.3m\n", + " 460 0.1867 23.2% 1.3m\n", + " 470 0.2174 23.2% 1.2m\n", + " 480 0.1659 23.2% 1.2m\n", + " 490 0.1899 23.2% 1.2m\n", + " 500 0.1620 23.2% 1.1m\n", + " 510 0.1690 23.2% 1.1m\n", + " 520 0.1443 23.2% 1.1m\n", + " 530 0.1806 23.2% 1.0m\n", + " 540 0.1719 23.2% 1.0m\n", + " 550 0.1829 23.2% 1.0m\n", + " 560 0.1928 23.2% 0.9m\n", + " 570 0.1734 23.2% 0.9m\n", + " 580 0.1783 23.2% 0.9m\n", + " 590 0.1527 23.2% 0.9m\n", + " 600 0.1626 23.2% 0.8m\n", + " 610 0.1498 23.2% 0.8m\n", + " 620 0.1896 23.2% 0.8m\n", + " 630 0.1724 23.2% 0.7m\n", + " 640 0.1491 23.2% 0.7m\n", + " 650 0.1760 23.2% 0.7m\n", + " 660 0.1725 23.2% 0.6m\n", + " 670 0.1529 23.2% 0.6m\n", + " 680 0.1579 23.2% 0.6m\n", + " 690 0.1445 23.2% 0.5m\n", + " 700 0.1955 23.2% 0.5m\n", + " 710 0.1823 23.2% 0.5m\n", + " 720 0.1710 23.2% 0.4m\n", + " 730 0.1396 23.2% 0.4m\n", + " 740 0.1271 23.2% 0.4m\n", + " 750 0.1645 23.2% 0.4m\n", + " 760 0.1373 23.2% 0.3m\n", + " 770 0.1920 23.2% 0.3m\n", + " 780 0.1680 23.2% 0.3m\n", + " 790 0.1814 23.2% 0.2m\n", + " 800 0.1820 23.2% 0.2m\n", + " 810 0.1398 23.2% 0.2m\n", + " 820 0.1760 23.2% 0.1m\n", + " 830 0.1657 23.2% 0.1m\n", + " 840 0.1824 23.2% 0.1m\n", + " 850 0.1617 23.2% 0.0m\n", + " 860 0.1258 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 44 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 44 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16683 โ†’ Val=0.16410\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 159m\n", + "\n", + "๐Ÿ”„ Epoch 45/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1606 23.2% 2.7m\n", + " 10 0.1780 23.2% 2.7m\n", + " 20 0.1808 23.2% 2.7m\n", + " 30 0.1660 23.2% 2.6m\n", + " 40 0.1104 23.2% 2.6m\n", + " 50 0.1540 23.2% 2.5m\n", + " 60 0.2006 23.2% 2.5m\n", + " 70 0.1639 23.2% 2.5m\n", + " 80 0.1508 23.2% 2.5m\n", + " 90 0.1630 23.2% 2.4m\n", + " 100 0.1539 23.2% 2.4m\n", + " 110 0.1285 23.2% 2.4m\n", + " 120 0.1797 23.2% 2.3m\n", + " 130 0.1853 23.2% 2.3m\n", + " 140 0.1571 23.2% 2.3m\n", + " 150 0.1749 23.2% 2.2m\n", + " 160 0.1364 23.2% 2.2m\n", + " 170 0.1311 23.2% 2.2m\n", + " 180 0.1808 23.2% 2.1m\n", + " 190 0.1454 23.2% 2.1m\n", + " 200 0.1618 23.2% 2.1m\n", + " 210 0.1787 23.2% 2.0m\n", + " 220 0.1770 23.2% 2.0m\n", + " 230 0.1779 23.2% 2.0m\n", + " 240 0.1493 23.2% 2.0m\n", + " 250 0.1909 23.2% 1.9m\n", + " 260 0.1844 23.2% 1.9m\n", + " 270 0.1844 23.2% 1.9m\n", + " 280 0.1781 23.2% 1.8m\n", + " 290 0.1630 23.2% 1.8m\n", + " 300 0.1722 23.2% 1.8m\n", + " 310 0.1906 23.2% 1.7m\n", + " 320 0.1644 23.2% 1.7m\n", + " 330 0.1693 23.2% 1.7m\n", + " 340 0.1911 23.2% 1.6m\n", + " 350 0.1821 23.2% 1.6m\n", + " 360 0.1821 23.2% 1.6m\n", + " 370 0.1558 23.2% 1.5m\n", + " 380 0.1721 23.2% 1.5m\n", + " 390 0.1461 23.2% 1.5m\n", + " 400 0.1815 23.2% 1.5m\n", + " 410 0.1803 23.2% 1.4m\n", + " 420 0.2168 23.2% 1.4m\n", + " 430 0.1660 23.2% 1.4m\n", + " 440 0.1816 23.2% 1.3m\n", + " 450 0.1616 23.2% 1.3m\n", + " 460 0.1717 23.2% 1.3m\n", + " 470 0.1890 23.2% 1.2m\n", + " 480 0.1605 23.2% 1.2m\n", + " 490 0.1389 23.2% 1.2m\n", + " 500 0.1694 23.2% 1.1m\n", + " 510 0.2012 23.2% 1.1m\n", + " 520 0.1612 23.2% 1.1m\n", + " 530 0.1577 23.2% 1.0m\n", + " 540 0.1660 23.2% 1.0m\n", + " 550 0.1381 23.2% 1.0m\n", + " 560 0.1715 23.2% 1.0m\n", + " 570 0.1617 23.2% 0.9m\n", + " 580 0.1573 23.2% 0.9m\n", + " 590 0.1454 23.2% 0.9m\n", + " 600 0.1302 23.2% 0.8m\n", + " 610 0.1572 23.2% 0.8m\n", + " 620 0.1586 23.2% 0.8m\n", + " 630 0.1524 23.2% 0.7m\n", + " 640 0.1635 23.2% 0.7m\n", + " 650 0.1927 23.2% 0.7m\n", + " 660 0.1207 23.2% 0.6m\n", + " 670 0.1526 23.2% 0.6m\n", + " 680 0.1818 23.2% 0.6m\n", + " 690 0.2094 23.2% 0.5m\n", + " 700 0.1597 23.2% 0.5m\n", + " 710 0.1703 23.2% 0.5m\n", + " 720 0.1584 23.2% 0.4m\n", + " 730 0.1688 23.2% 0.4m\n", + " 740 0.1546 23.2% 0.4m\n", + " 750 0.1916 23.2% 0.4m\n", + " 760 0.2073 23.2% 0.3m\n", + " 770 0.1735 23.2% 0.3m\n", + " 780 0.1537 23.2% 0.3m\n", + " 790 0.1896 23.2% 0.2m\n", + " 800 0.1815 23.2% 0.2m\n", + " 810 0.1464 23.2% 0.2m\n", + " 820 0.1750 23.2% 0.1m\n", + " 830 0.1918 23.2% 0.1m\n", + " 840 0.1668 23.2% 0.1m\n", + " 850 0.1618 23.2% 0.0m\n", + " 860 0.1417 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 45 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 45 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16683 โ†’ Val=0.16424\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 157m\n", + "\n", + "๐Ÿ”„ Epoch 46/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1733 23.2% 2.7m\n", + " 10 0.1726 23.2% 2.7m\n", + " 20 0.1437 23.2% 2.7m\n", + " 30 0.1695 23.2% 2.6m\n", + " 40 0.1579 23.2% 2.6m\n", + " 50 0.1666 23.2% 2.6m\n", + " 60 0.1327 23.2% 2.5m\n", + " 70 0.1747 23.2% 2.5m\n", + " 80 0.1443 23.2% 2.5m\n", + " 90 0.1769 23.2% 2.4m\n", + " 100 0.1754 23.2% 2.4m\n", + " 110 0.1536 23.2% 2.4m\n", + " 120 0.1510 23.2% 2.3m\n", + " 130 0.1585 23.2% 2.3m\n", + " 140 0.1418 23.2% 2.3m\n", + " 150 0.1559 23.2% 2.2m\n", + " 160 0.1963 23.2% 2.2m\n", + " 170 0.1509 23.2% 2.2m\n", + " 180 0.1986 23.2% 2.1m\n", + " 190 0.1716 23.2% 2.1m\n", + " 200 0.2109 23.2% 2.1m\n", + " 210 0.1733 23.2% 2.0m\n", + " 220 0.1894 23.2% 2.0m\n", + " 230 0.1970 23.2% 2.0m\n", + " 240 0.1647 23.2% 2.0m\n", + " 250 0.1462 23.2% 1.9m\n", + " 260 0.1548 23.2% 1.9m\n", + " 270 0.1524 23.2% 1.9m\n", + " 280 0.1573 23.2% 1.8m\n", + " 290 0.1844 23.2% 1.8m\n", + " 300 0.1867 23.2% 1.8m\n", + " 310 0.1804 23.2% 1.7m\n", + " 320 0.1729 23.2% 1.7m\n", + " 330 0.1484 23.2% 1.7m\n", + " 340 0.1574 23.2% 1.6m\n", + " 350 0.1604 23.2% 1.6m\n", + " 360 0.1807 23.2% 1.6m\n", + " 370 0.1723 23.2% 1.5m\n", + " 380 0.2117 23.2% 1.5m\n", + " 390 0.1589 23.2% 1.5m\n", + " 400 0.1772 23.2% 1.4m\n", + " 410 0.1444 23.2% 1.4m\n", + " 420 0.1621 23.2% 1.4m\n", + " 430 0.1964 23.2% 1.4m\n", + " 440 0.1552 23.2% 1.3m\n", + " 450 0.1623 23.2% 1.3m\n", + " 460 0.2137 23.2% 1.3m\n", + " 470 0.1306 23.2% 1.2m\n", + " 480 0.1172 23.2% 1.2m\n", + " 490 0.2077 23.2% 1.2m\n", + " 500 0.1816 23.2% 1.1m\n", + " 510 0.1851 23.2% 1.1m\n", + " 520 0.1756 23.2% 1.1m\n", + " 530 0.1763 23.2% 1.0m\n", + " 540 0.1785 23.2% 1.0m\n", + " 550 0.1697 23.2% 1.0m\n", + " 560 0.1958 23.2% 0.9m\n", + " 570 0.1616 23.2% 0.9m\n", + " 580 0.1731 23.2% 0.9m\n", + " 590 0.1604 23.2% 0.9m\n", + " 600 0.1645 23.2% 0.8m\n", + " 610 0.1257 23.2% 0.8m\n", + " 620 0.1745 23.2% 0.8m\n", + " 630 0.1452 23.2% 0.7m\n", + " 640 0.1850 23.2% 0.7m\n", + " 650 0.1550 23.2% 0.7m\n", + " 660 0.1820 23.2% 0.6m\n", + " 670 0.1605 23.2% 0.6m\n", + " 680 0.1755 23.2% 0.6m\n", + " 690 0.1706 23.2% 0.5m\n", + " 700 0.1707 23.2% 0.5m\n", + " 710 0.1580 23.2% 0.5m\n", + " 720 0.1910 23.2% 0.4m\n", + " 730 0.1789 23.2% 0.4m\n", + " 740 0.1227 23.2% 0.4m\n", + " 750 0.1293 23.2% 0.4m\n", + " 760 0.1718 23.2% 0.3m\n", + " 770 0.1474 23.2% 0.3m\n", + " 780 0.1910 23.2% 0.3m\n", + " 790 0.1459 23.2% 0.2m\n", + " 800 0.1549 23.2% 0.2m\n", + " 810 0.1245 23.2% 0.2m\n", + " 820 0.1780 23.2% 0.1m\n", + " 830 0.1346 23.2% 0.1m\n", + " 840 0.1598 23.2% 0.1m\n", + " 850 0.2060 23.2% 0.0m\n", + " 860 0.1606 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 46 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 46 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16686 โ†’ Val=0.16452\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 153m\n", + "\n", + "๐Ÿ”„ Epoch 47/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2202 23.2% 2.7m\n", + " 10 0.1547 23.2% 2.6m\n", + " 20 0.1801 23.2% 2.6m\n", + " 30 0.1675 23.2% 2.6m\n", + " 40 0.1539 23.2% 2.6m\n", + " 50 0.1607 23.2% 2.5m\n", + " 60 0.1615 23.2% 2.5m\n", + " 70 0.1521 23.2% 2.5m\n", + " 80 0.1629 23.2% 2.4m\n", + " 90 0.1203 23.2% 2.4m\n", + " 100 0.1775 23.2% 2.4m\n", + " 110 0.1737 23.2% 2.4m\n", + " 120 0.1402 23.2% 2.3m\n", + " 130 0.1739 23.2% 2.3m\n", + " 140 0.1603 23.2% 2.3m\n", + " 150 0.1635 23.2% 2.2m\n", + " 160 0.1732 23.2% 2.2m\n", + " 170 0.1769 23.2% 2.2m\n", + " 180 0.2079 23.2% 2.1m\n", + " 190 0.1484 23.2% 2.1m\n", + " 200 0.1449 23.2% 2.1m\n", + " 210 0.1240 23.2% 2.0m\n", + " 220 0.1493 23.2% 2.0m\n", + " 230 0.1935 23.2% 2.0m\n", + " 240 0.1219 23.2% 1.9m\n", + " 250 0.1600 23.2% 1.9m\n", + " 260 0.1370 23.2% 1.9m\n", + " 270 0.1520 23.2% 1.9m\n", + " 280 0.1519 23.2% 1.8m\n", + " 290 0.1418 23.2% 1.8m\n", + " 300 0.1661 23.2% 1.8m\n", + " 310 0.1564 23.2% 1.7m\n", + " 320 0.1862 23.2% 1.7m\n", + " 330 0.1609 23.2% 1.7m\n", + " 340 0.1363 23.2% 1.6m\n", + " 350 0.1944 23.2% 1.6m\n", + " 360 0.1450 23.2% 1.6m\n", + " 370 0.1617 23.2% 1.5m\n", + " 380 0.1685 23.2% 1.5m\n", + " 390 0.1714 23.2% 1.5m\n", + " 400 0.1449 23.2% 1.4m\n", + " 410 0.1197 23.2% 1.4m\n", + " 420 0.1826 23.2% 1.4m\n", + " 430 0.1786 23.2% 1.4m\n", + " 440 0.1759 23.2% 1.3m\n", + " 450 0.1760 23.2% 1.3m\n", + " 460 0.1633 23.2% 1.3m\n", + " 470 0.1868 23.2% 1.2m\n", + " 480 0.1829 23.2% 1.2m\n", + " 490 0.1623 23.2% 1.2m\n", + " 500 0.1463 23.2% 1.1m\n", + " 510 0.1669 23.2% 1.1m\n", + " 520 0.1405 23.2% 1.1m\n", + " 530 0.1412 23.2% 1.0m\n", + " 540 0.1316 23.2% 1.0m\n", + " 550 0.1455 23.2% 1.0m\n", + " 560 0.1340 23.2% 0.9m\n", + " 570 0.2020 23.2% 0.9m\n", + " 580 0.1514 23.2% 0.9m\n", + " 590 0.1919 23.2% 0.9m\n", + " 600 0.1545 23.2% 0.8m\n", + " 610 0.1876 23.2% 0.8m\n", + " 620 0.1940 23.2% 0.8m\n", + " 630 0.2004 23.2% 0.7m\n", + " 640 0.1513 23.2% 0.7m\n", + " 650 0.1385 23.2% 0.7m\n", + " 660 0.1868 23.2% 0.6m\n", + " 670 0.1478 23.2% 0.6m\n", + " 680 0.1630 23.2% 0.6m\n", + " 690 0.1344 23.2% 0.5m\n", + " 700 0.2115 23.2% 0.5m\n", + " 710 0.2033 23.2% 0.5m\n", + " 720 0.1842 23.2% 0.4m\n", + " 730 0.1576 23.2% 0.4m\n", + " 740 0.1438 23.2% 0.4m\n", + " 750 0.1679 23.2% 0.4m\n", + " 760 0.1788 23.2% 0.3m\n", + " 770 0.1432 23.2% 0.3m\n", + " 780 0.1775 23.2% 0.3m\n", + " 790 0.1677 23.2% 0.2m\n", + " 800 0.1689 23.2% 0.2m\n", + " 810 0.1838 23.2% 0.2m\n", + " 820 0.1628 23.2% 0.1m\n", + " 830 0.1529 23.2% 0.1m\n", + " 840 0.1619 23.2% 0.1m\n", + " 850 0.1927 23.2% 0.0m\n", + " 860 0.1799 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 47 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 47 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16675 โ†’ Val=0.16421\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โš ๏ธ No improvement for 20/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 150m\n", + "\n", + "๐Ÿ”„ Epoch 48/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1634 23.2% 2.7m\n", + " 10 0.1571 23.2% 2.6m\n", + " 20 0.1670 23.2% 2.7m\n", + " 30 0.1434 23.2% 2.6m\n", + " 40 0.2027 23.2% 2.6m\n", + " 50 0.1908 23.2% 2.6m\n", + " 60 0.1969 23.2% 2.5m\n", + " 70 0.1647 23.2% 2.5m\n", + " 80 0.1632 23.2% 2.4m\n", + " 90 0.1777 23.2% 2.4m\n", + " 100 0.1550 23.2% 2.4m\n", + " 110 0.1801 23.2% 2.4m\n", + " 120 0.1679 23.2% 2.3m\n", + " 130 0.1423 23.2% 2.3m\n", + " 140 0.1975 23.2% 2.3m\n", + " 150 0.1602 23.2% 2.2m\n", + " 160 0.1775 23.2% 2.2m\n", + " 170 0.1636 23.2% 2.2m\n", + " 180 0.1446 23.2% 2.1m\n", + " 190 0.1761 23.2% 2.1m\n", + " 200 0.1385 23.2% 2.1m\n", + " 210 0.1750 23.2% 2.1m\n", + " 220 0.1688 23.2% 2.0m\n", + " 230 0.1621 23.2% 2.0m\n", + " 240 0.1681 23.2% 2.0m\n", + " 250 0.1692 23.2% 1.9m\n", + " 260 0.1554 23.2% 1.9m\n", + " 270 0.1893 23.2% 1.9m\n", + " 280 0.1411 23.2% 1.8m\n", + " 290 0.1074 23.2% 1.8m\n", + " 300 0.1476 23.2% 1.8m\n", + " 310 0.1729 23.2% 1.7m\n", + " 320 0.1886 23.2% 1.7m\n", + " 330 0.1970 23.2% 1.7m\n", + " 340 0.2117 23.2% 1.6m\n", + " 350 0.1606 23.2% 1.6m\n", + " 360 0.1861 23.2% 1.6m\n", + " 370 0.1816 23.2% 1.5m\n", + " 380 0.1874 23.2% 1.5m\n", + " 390 0.1559 23.2% 1.5m\n", + " 400 0.1818 23.2% 1.5m\n", + " 410 0.1638 23.2% 1.4m\n", + " 420 0.2009 23.2% 1.4m\n", + " 430 0.1671 23.2% 1.4m\n", + " 440 0.1635 23.2% 1.3m\n", + " 450 0.1647 23.2% 1.3m\n", + " 460 0.1953 23.2% 1.3m\n", + " 470 0.1452 23.2% 1.2m\n", + " 480 0.1560 23.2% 1.2m\n", + " 490 0.2376 23.2% 1.2m\n", + " 500 0.1871 23.2% 1.1m\n", + " 510 0.1445 23.2% 1.1m\n", + " 520 0.1689 23.2% 1.1m\n", + " 530 0.1560 23.2% 1.0m\n", + " 540 0.1567 23.2% 1.0m\n", + " 550 0.1644 23.2% 1.0m\n", + " 560 0.1631 23.2% 0.9m\n", + " 570 0.1907 23.2% 0.9m\n", + " 580 0.1410 23.2% 0.9m\n", + " 590 0.1531 23.2% 0.9m\n", + " 600 0.1546 23.2% 0.8m\n", + " 610 0.1652 23.2% 0.8m\n", + " 620 0.1429 23.2% 0.8m\n", + " 630 0.1623 23.2% 0.7m\n", + " 640 0.1951 23.2% 0.7m\n", + " 650 0.1813 23.2% 0.7m\n", + " 660 0.1718 23.2% 0.6m\n", + " 670 0.1677 23.2% 0.6m\n", + " 680 0.1375 23.2% 0.6m\n", + " 690 0.1263 23.2% 0.5m\n", + " 700 0.1154 23.2% 0.5m\n", + " 710 0.1531 23.2% 0.5m\n", + " 720 0.1589 23.2% 0.4m\n", + " 730 0.1562 23.2% 0.4m\n", + " 740 0.1796 23.2% 0.4m\n", + " 750 0.1866 23.2% 0.4m\n", + " 760 0.1987 23.2% 0.3m\n", + " 770 0.1832 23.2% 0.3m\n", + " 780 0.1344 23.2% 0.3m\n", + " 790 0.1486 23.2% 0.2m\n", + " 800 0.1531 23.2% 0.2m\n", + " 810 0.1411 23.2% 0.2m\n", + " 820 0.1398 23.2% 0.1m\n", + " 830 0.1718 23.2% 0.1m\n", + " 840 0.1515 23.2% 0.1m\n", + " 850 0.1652 23.2% 0.0m\n", + " 860 0.1706 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 48 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 48 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16666 โ†’ Val=0.16419\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 148m\n", + "\n", + "๐Ÿ”„ Epoch 49/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1861 23.2% 2.7m\n", + " 10 0.1468 23.2% 2.7m\n", + " 20 0.1236 23.2% 2.6m\n", + " 30 0.1559 23.2% 2.6m\n", + " 40 0.1083 23.2% 2.6m\n", + " 50 0.1703 23.2% 2.5m\n", + " 60 0.1804 23.2% 2.5m\n", + " 70 0.1582 23.2% 2.5m\n", + " 80 0.1642 23.2% 2.5m\n", + " 90 0.1702 23.2% 2.4m\n", + " 100 0.1616 23.2% 2.4m\n", + " 110 0.1767 23.2% 2.4m\n", + " 120 0.1902 23.2% 2.3m\n", + " 130 0.1989 23.2% 2.3m\n", + " 140 0.1462 23.2% 2.3m\n", + " 150 0.1462 23.2% 2.2m\n", + " 160 0.1634 23.2% 2.2m\n", + " 170 0.2023 23.2% 2.2m\n", + " 180 0.1885 23.2% 2.1m\n", + " 190 0.1514 23.2% 2.1m\n", + " 200 0.1293 23.2% 2.1m\n", + " 210 0.1416 23.2% 2.0m\n", + " 220 0.1952 23.2% 2.0m\n", + " 230 0.1928 23.2% 2.0m\n", + " 240 0.1591 23.2% 2.0m\n", + " 250 0.1599 23.2% 1.9m\n", + " 260 0.1454 23.2% 1.9m\n", + " 270 0.1467 23.2% 1.9m\n", + " 280 0.1435 23.2% 1.8m\n", + " 290 0.1627 23.2% 1.8m\n", + " 300 0.1430 23.2% 1.8m\n", + " 310 0.1415 23.2% 1.7m\n", + " 320 0.1667 23.2% 1.7m\n", + " 330 0.1453 23.2% 1.7m\n", + " 340 0.1569 23.2% 1.6m\n", + " 350 0.1906 23.2% 1.6m\n", + " 360 0.1690 23.2% 1.6m\n", + " 370 0.1528 23.2% 1.5m\n", + " 380 0.1802 23.2% 1.5m\n", + " 390 0.1824 23.2% 1.5m\n", + " 400 0.1525 23.2% 1.4m\n", + " 410 0.1702 23.2% 1.4m\n", + " 420 0.1829 23.2% 1.4m\n", + " 430 0.1327 23.2% 1.4m\n", + " 440 0.1801 23.2% 1.3m\n", + " 450 0.1872 23.2% 1.3m\n", + " 460 0.1571 23.2% 1.3m\n", + " 470 0.1486 23.2% 1.2m\n", + " 480 0.1623 23.2% 1.2m\n", + " 490 0.1815 23.2% 1.2m\n", + " 500 0.1687 23.2% 1.1m\n", + " 510 0.1588 23.2% 1.1m\n", + " 520 0.1871 23.2% 1.1m\n", + " 530 0.1844 23.2% 1.0m\n", + " 540 0.1747 23.2% 1.0m\n", + " 550 0.1905 23.2% 1.0m\n", + " 560 0.1745 23.2% 0.9m\n", + " 570 0.1401 23.2% 0.9m\n", + " 580 0.1465 23.2% 0.9m\n", + " 590 0.1847 23.2% 0.9m\n", + " 600 0.1389 23.2% 0.8m\n", + " 610 0.1636 23.2% 0.8m\n", + " 620 0.1771 23.2% 0.8m\n", + " 630 0.1481 23.2% 0.7m\n", + " 640 0.1688 23.2% 0.7m\n", + " 650 0.1686 23.2% 0.7m\n", + " 660 0.1502 23.2% 0.6m\n", + " 670 0.1550 23.2% 0.6m\n", + " 680 0.1877 23.2% 0.6m\n", + " 690 0.1984 23.2% 0.5m\n", + " 700 0.1707 23.2% 0.5m\n", + " 710 0.1337 23.2% 0.5m\n", + " 720 0.1886 23.2% 0.4m\n", + " 730 0.1658 23.2% 0.4m\n", + " 740 0.2102 23.2% 0.4m\n", + " 750 0.1419 23.2% 0.4m\n", + " 760 0.1489 23.2% 0.3m\n", + " 770 0.1627 23.2% 0.3m\n", + " 780 0.1764 23.2% 0.3m\n", + " 790 0.1511 23.2% 0.2m\n", + " 800 0.1628 23.2% 0.2m\n", + " 810 0.1241 23.2% 0.2m\n", + " 820 0.1659 23.2% 0.1m\n", + " 830 0.1711 23.2% 0.1m\n", + " 840 0.1491 23.2% 0.1m\n", + " 850 0.1910 23.2% 0.0m\n", + " 860 0.1544 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 49 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 49 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16682 โ†’ Val=0.16438\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 145m\n", + "\n", + "๐Ÿ”„ Epoch 50/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2047 23.2% 2.7m\n", + " 10 0.1427 23.2% 2.7m\n", + " 20 0.1596 23.2% 2.6m\n", + " 30 0.1963 23.2% 2.6m\n", + " 40 0.1631 23.2% 2.6m\n", + " 50 0.1339 23.2% 2.6m\n", + " 60 0.1807 23.2% 2.5m\n", + " 70 0.1812 23.2% 2.5m\n", + " 80 0.1530 23.2% 2.5m\n", + " 90 0.1871 23.2% 2.4m\n", + " 100 0.1634 23.2% 2.4m\n", + " 110 0.1790 23.2% 2.4m\n", + " 120 0.1951 23.2% 2.3m\n", + " 130 0.1597 23.2% 2.3m\n", + " 140 0.1427 23.2% 2.3m\n", + " 150 0.1901 23.2% 2.2m\n", + " 160 0.1495 23.2% 2.2m\n", + " 170 0.1620 23.2% 2.2m\n", + " 180 0.1138 23.2% 2.1m\n", + " 190 0.1536 23.2% 2.1m\n", + " 200 0.1620 23.2% 2.1m\n", + " 210 0.1586 23.2% 2.0m\n", + " 220 0.1299 23.2% 2.0m\n", + " 230 0.1323 23.2% 2.0m\n", + " 240 0.1313 23.2% 1.9m\n", + " 250 0.1487 23.2% 1.9m\n", + " 260 0.1742 23.2% 1.9m\n", + " 270 0.1631 23.2% 1.9m\n", + " 280 0.1343 23.2% 1.8m\n", + " 290 0.1469 23.2% 1.8m\n", + " 300 0.1522 23.2% 1.8m\n", + " 310 0.2166 23.2% 1.7m\n", + " 320 0.1569 23.2% 1.7m\n", + " 330 0.1646 23.2% 1.7m\n", + " 340 0.1628 23.2% 1.6m\n", + " 350 0.1984 23.2% 1.6m\n", + " 360 0.1791 23.2% 1.6m\n", + " 370 0.1797 23.2% 1.5m\n", + " 380 0.1665 23.2% 1.5m\n", + " 390 0.1837 23.2% 1.5m\n", + " 400 0.1593 23.2% 1.5m\n", + " 410 0.1468 23.2% 1.4m\n", + " 420 0.1657 23.2% 1.4m\n", + " 430 0.1769 23.2% 1.4m\n", + " 440 0.1540 23.2% 1.3m\n", + " 450 0.1928 23.2% 1.3m\n", + " 460 0.2016 23.2% 1.3m\n", + " 470 0.1476 23.2% 1.2m\n", + " 480 0.2153 23.2% 1.2m\n", + " 490 0.1352 23.2% 1.2m\n", + " 500 0.1430 23.2% 1.1m\n", + " 510 0.1935 23.2% 1.1m\n", + " 520 0.1696 23.2% 1.1m\n", + " 530 0.1895 23.2% 1.0m\n", + " 540 0.1581 23.2% 1.0m\n", + " 550 0.1690 23.2% 1.0m\n", + " 560 0.1682 23.2% 0.9m\n", + " 570 0.1682 23.2% 0.9m\n", + " 580 0.2040 23.2% 0.9m\n", + " 590 0.1432 23.2% 0.9m\n", + " 600 0.1931 23.2% 0.8m\n", + " 610 0.1790 23.2% 0.8m\n", + " 620 0.1591 23.2% 0.8m\n", + " 630 0.1362 23.2% 0.7m\n", + " 640 0.1661 23.2% 0.7m\n", + " 650 0.1650 23.2% 0.7m\n", + " 660 0.2035 23.2% 0.6m\n", + " 670 0.1470 23.2% 0.6m\n", + " 680 0.1682 23.2% 0.6m\n", + " 690 0.1458 23.2% 0.5m\n", + " 700 0.1498 23.2% 0.5m\n", + " 710 0.1864 23.2% 0.5m\n", + " 720 0.1814 23.2% 0.4m\n", + " 730 0.1419 23.2% 0.4m\n", + " 740 0.1714 23.2% 0.4m\n", + " 750 0.1582 23.2% 0.4m\n", + " 760 0.1762 23.2% 0.3m\n", + " 770 0.2078 23.2% 0.3m\n", + " 780 0.1521 23.2% 0.3m\n", + " 790 0.1450 23.2% 0.2m\n", + " 800 0.2110 23.2% 0.2m\n", + " 810 0.2000 23.2% 0.2m\n", + " 820 0.1935 23.2% 0.1m\n", + " 830 0.1777 23.2% 0.1m\n", + " 840 0.1913 23.2% 0.1m\n", + " 850 0.1640 23.2% 0.0m\n", + " 860 0.1429 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 50 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 50 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16672 โ†’ Val=0.16416\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 142m\n", + "\n", + "๐Ÿ”„ Epoch 51/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1297 23.2% 2.6m\n", + " 10 0.1714 23.2% 2.7m\n", + " 20 0.1958 23.2% 2.6m\n", + " 30 0.1414 23.2% 2.6m\n", + " 40 0.1519 23.2% 2.6m\n", + " 50 0.1641 23.2% 2.5m\n", + " 60 0.1623 23.2% 2.5m\n", + " 70 0.1468 23.2% 2.5m\n", + " 80 0.1464 23.2% 2.4m\n", + " 90 0.1449 23.2% 2.4m\n", + " 100 0.1751 23.2% 2.4m\n", + " 110 0.1535 23.2% 2.3m\n", + " 120 0.1571 23.2% 2.3m\n", + " 130 0.1724 23.2% 2.3m\n", + " 140 0.1360 23.2% 2.3m\n", + " 150 0.1524 23.2% 2.2m\n", + " 160 0.1787 23.2% 2.2m\n", + " 170 0.1673 23.2% 2.2m\n", + " 180 0.1592 23.2% 2.1m\n", + " 190 0.1441 23.2% 2.1m\n", + " 200 0.1499 23.2% 2.1m\n", + " 210 0.1558 23.2% 2.0m\n", + " 220 0.1473 23.2% 2.0m\n", + " 230 0.1530 23.2% 2.0m\n", + " 240 0.1437 23.2% 1.9m\n", + " 250 0.1640 23.2% 1.9m\n", + " 260 0.1914 23.2% 1.9m\n", + " 270 0.1544 23.2% 1.8m\n", + " 280 0.1702 23.2% 1.8m\n", + " 290 0.1504 23.2% 1.8m\n", + " 300 0.1822 23.2% 1.8m\n", + " 310 0.1517 23.2% 1.7m\n", + " 320 0.1450 23.2% 1.7m\n", + " 330 0.1617 23.2% 1.7m\n", + " 340 0.1290 23.2% 1.6m\n", + " 350 0.1621 23.2% 1.6m\n", + " 360 0.1640 23.2% 1.6m\n", + " 370 0.1745 23.2% 1.5m\n", + " 380 0.1913 23.2% 1.5m\n", + " 390 0.1354 23.2% 1.5m\n", + " 400 0.1680 23.2% 1.4m\n", + " 410 0.1542 23.2% 1.4m\n", + " 420 0.1723 23.2% 1.4m\n", + " 430 0.1661 23.2% 1.3m\n", + " 440 0.1842 23.2% 1.3m\n", + " 450 0.1703 23.2% 1.3m\n", + " 460 0.1656 23.2% 1.3m\n", + " 470 0.1792 23.2% 1.2m\n", + " 480 0.1587 23.2% 1.2m\n", + " 490 0.1573 23.2% 1.2m\n", + " 500 0.1672 23.2% 1.1m\n", + " 510 0.1569 23.2% 1.1m\n", + " 520 0.1626 23.2% 1.1m\n", + " 530 0.1575 23.2% 1.0m\n", + " 540 0.1903 23.2% 1.0m\n", + " 550 0.1931 23.2% 1.0m\n", + " 560 0.1650 23.2% 0.9m\n", + " 570 0.1592 23.2% 0.9m\n", + " 580 0.1984 23.2% 0.9m\n", + " 590 0.1628 23.2% 0.9m\n", + " 600 0.1759 23.2% 0.8m\n", + " 610 0.1745 23.2% 0.8m\n", + " 620 0.1180 23.2% 0.8m\n", + " 630 0.1597 23.2% 0.7m\n", + " 640 0.1698 23.2% 0.7m\n", + " 650 0.1386 23.2% 0.7m\n", + " 660 0.1621 23.2% 0.6m\n", + " 670 0.1892 23.2% 0.6m\n", + " 680 0.1608 23.2% 0.6m\n", + " 690 0.1682 23.2% 0.5m\n", + " 700 0.1622 23.2% 0.5m\n", + " 710 0.1528 23.2% 0.5m\n", + " 720 0.1938 23.2% 0.4m\n", + " 730 0.1805 23.2% 0.4m\n", + " 740 0.1533 23.2% 0.4m\n", + " 750 0.1667 23.2% 0.4m\n", + " 760 0.1637 23.2% 0.3m\n", + " 770 0.1839 23.2% 0.3m\n", + " 780 0.1809 23.2% 0.3m\n", + " 790 0.2384 23.2% 0.2m\n", + " 800 0.1690 23.2% 0.2m\n", + " 810 0.1483 23.2% 0.2m\n", + " 820 0.1959 23.2% 0.1m\n", + " 830 0.1913 23.2% 0.1m\n", + " 840 0.1891 23.2% 0.1m\n", + " 850 0.1525 23.2% 0.0m\n", + " 860 0.1840 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 51 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 51 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16685 โ†’ Val=0.16413\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 138m\n", + "\n", + "๐Ÿ”„ Epoch 52/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1527 23.2% 2.6m\n", + " 10 0.1858 23.2% 2.7m\n", + " 20 0.1898 23.2% 2.6m\n", + " 30 0.1805 23.2% 2.6m\n", + " 40 0.1713 23.2% 2.6m\n", + " 50 0.1488 23.2% 2.6m\n", + " 60 0.1192 23.2% 2.5m\n", + " 70 0.1517 23.2% 2.5m\n", + " 80 0.1878 23.2% 2.5m\n", + " 90 0.1432 23.2% 2.4m\n", + " 100 0.1843 23.2% 2.4m\n", + " 110 0.1939 23.2% 2.4m\n", + " 120 0.1668 23.2% 2.3m\n", + " 130 0.1816 23.2% 2.3m\n", + " 140 0.1563 23.2% 2.3m\n", + " 150 0.1904 23.2% 2.2m\n", + " 160 0.1836 23.2% 2.2m\n", + " 170 0.1410 23.2% 2.2m\n", + " 180 0.1601 23.2% 2.1m\n", + " 190 0.1784 23.2% 2.1m\n", + " 200 0.1455 23.2% 2.1m\n", + " 210 0.1606 23.2% 2.0m\n", + " 220 0.2166 23.2% 2.0m\n", + " 230 0.1754 23.2% 2.0m\n", + " 240 0.1695 23.2% 1.9m\n", + " 250 0.1649 23.2% 1.9m\n", + " 260 0.1702 23.2% 1.9m\n", + " 270 0.1893 23.2% 1.8m\n", + " 280 0.1658 23.2% 1.8m\n", + " 290 0.1874 23.2% 1.8m\n", + " 300 0.1650 23.2% 1.8m\n", + " 310 0.1779 23.2% 1.7m\n", + " 320 0.1574 23.2% 1.7m\n", + " 330 0.1416 23.2% 1.7m\n", + " 340 0.1286 23.2% 1.6m\n", + " 350 0.1657 23.2% 1.6m\n", + " 360 0.1667 23.2% 1.6m\n", + " 370 0.2033 23.2% 1.5m\n", + " 380 0.1563 23.2% 1.5m\n", + " 390 0.1806 23.2% 1.5m\n", + " 400 0.1765 23.2% 1.4m\n", + " 410 0.1835 23.2% 1.4m\n", + " 420 0.2048 23.2% 1.4m\n", + " 430 0.1751 23.2% 1.3m\n", + " 440 0.1682 23.2% 1.3m\n", + " 450 0.1684 23.2% 1.3m\n", + " 460 0.1652 23.2% 1.3m\n", + " 470 0.1405 23.2% 1.2m\n", + " 480 0.1713 23.2% 1.2m\n", + " 490 0.1712 23.2% 1.2m\n", + " 500 0.1639 23.2% 1.1m\n", + " 510 0.1725 23.2% 1.1m\n", + " 520 0.1519 23.2% 1.1m\n", + " 530 0.1722 23.2% 1.0m\n", + " 540 0.1640 23.2% 1.0m\n", + " 550 0.1417 23.2% 1.0m\n", + " 560 0.1354 23.2% 0.9m\n", + " 570 0.1944 23.2% 0.9m\n", + " 580 0.1330 23.2% 0.9m\n", + " 590 0.1550 23.2% 0.9m\n", + " 600 0.1490 23.2% 0.8m\n", + " 610 0.1791 23.2% 0.8m\n", + " 620 0.2019 23.2% 0.8m\n", + " 630 0.1855 23.2% 0.7m\n", + " 640 0.1681 23.2% 0.7m\n", + " 650 0.1829 23.2% 0.7m\n", + " 660 0.1529 23.2% 0.6m\n", + " 670 0.1969 23.2% 0.6m\n", + " 680 0.1709 23.2% 0.6m\n", + " 690 0.1610 23.2% 0.5m\n", + " 700 0.1682 23.2% 0.5m\n", + " 710 0.1413 23.2% 0.5m\n", + " 720 0.1596 23.2% 0.4m\n", + " 730 0.1605 23.2% 0.4m\n", + " 740 0.1875 23.2% 0.4m\n", + " 750 0.1878 23.2% 0.4m\n", + " 760 0.1688 23.2% 0.3m\n", + " 770 0.1775 23.2% 0.3m\n", + " 780 0.1707 23.2% 0.3m\n", + " 790 0.1602 23.2% 0.2m\n", + " 800 0.1627 23.2% 0.2m\n", + " 810 0.1934 23.2% 0.2m\n", + " 820 0.1438 23.2% 0.1m\n", + " 830 0.1469 23.2% 0.1m\n", + " 840 0.1433 23.2% 0.1m\n", + " 850 0.1620 23.2% 0.0m\n", + " 860 0.1743 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 52 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 52 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16664 โ†’ Val=0.16436\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โš ๏ธ No improvement for 25/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 136m\n", + "\n", + "๐Ÿ”„ Epoch 53/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1303 23.2% 2.7m\n", + " 10 0.1677 23.2% 2.7m\n", + " 20 0.1742 23.2% 2.6m\n", + " 30 0.1611 23.2% 2.6m\n", + " 40 0.1661 23.2% 2.6m\n", + " 50 0.2062 23.2% 2.5m\n", + " 60 0.2194 23.2% 2.5m\n", + " 70 0.1598 23.2% 2.5m\n", + " 80 0.1682 23.2% 2.4m\n", + " 90 0.1472 23.2% 2.4m\n", + " 100 0.2113 23.2% 2.4m\n", + " 110 0.1748 23.2% 2.4m\n", + " 120 0.1660 23.2% 2.3m\n", + " 130 0.2183 23.2% 2.3m\n", + " 140 0.1896 23.2% 2.3m\n", + " 150 0.1670 23.2% 2.2m\n", + " 160 0.1531 23.2% 2.2m\n", + " 170 0.1589 23.2% 2.2m\n", + " 180 0.1795 23.2% 2.1m\n", + " 190 0.2099 23.2% 2.1m\n", + " 200 0.1666 23.2% 2.1m\n", + " 210 0.1633 23.2% 2.0m\n", + " 220 0.1712 23.2% 2.0m\n", + " 230 0.1644 23.2% 2.0m\n", + " 240 0.1740 23.2% 1.9m\n", + " 250 0.1842 23.2% 1.9m\n", + " 260 0.1630 23.2% 1.9m\n", + " 270 0.1846 23.2% 1.8m\n", + " 280 0.1099 23.2% 1.8m\n", + " 290 0.1720 23.2% 1.8m\n", + " 300 0.2259 23.2% 1.8m\n", + " 310 0.1749 23.2% 1.7m\n", + " 320 0.1515 23.2% 1.7m\n", + " 330 0.1602 23.2% 1.7m\n", + " 340 0.1712 23.2% 1.6m\n", + " 350 0.1503 23.2% 1.6m\n", + " 360 0.1533 23.2% 1.6m\n", + " 370 0.1603 23.2% 1.5m\n", + " 380 0.1647 23.2% 1.5m\n", + " 390 0.2030 23.2% 1.5m\n", + " 400 0.1770 23.2% 1.4m\n", + " 410 0.1768 23.2% 1.4m\n", + " 420 0.1852 23.2% 1.4m\n", + " 430 0.1801 23.2% 1.4m\n", + " 440 0.1701 23.2% 1.3m\n", + " 450 0.1601 23.2% 1.3m\n", + " 460 0.1723 23.2% 1.3m\n", + " 470 0.1586 23.2% 1.2m\n", + " 480 0.1767 23.2% 1.2m\n", + " 490 0.1611 23.2% 1.2m\n", + " 500 0.1977 23.2% 1.1m\n", + " 510 0.1652 23.2% 1.1m\n", + " 520 0.1770 23.2% 1.1m\n", + " 530 0.1768 23.2% 1.0m\n", + " 540 0.2250 23.2% 1.0m\n", + " 550 0.2119 23.2% 1.0m\n", + " 560 0.1602 23.2% 0.9m\n", + " 570 0.1453 23.2% 0.9m\n", + " 580 0.1382 23.2% 0.9m\n", + " 590 0.1556 23.2% 0.9m\n", + " 600 0.1603 23.2% 0.8m\n", + " 610 0.1626 23.2% 0.8m\n", + " 620 0.2032 23.2% 0.8m\n", + " 630 0.1620 23.2% 0.7m\n", + " 640 0.1650 23.2% 0.7m\n", + " 650 0.1685 23.2% 0.7m\n", + " 660 0.1349 23.2% 0.6m\n", + " 670 0.1703 23.2% 0.6m\n", + " 680 0.1839 23.2% 0.6m\n", + " 690 0.1718 23.2% 0.5m\n", + " 700 0.1804 23.2% 0.5m\n", + " 710 0.1293 23.2% 0.5m\n", + " 720 0.1738 23.2% 0.4m\n", + " 730 0.1369 23.2% 0.4m\n", + " 740 0.2137 23.2% 0.4m\n", + " 750 0.1783 23.2% 0.4m\n", + " 760 0.1971 23.2% 0.3m\n", + " 770 0.1896 23.2% 0.3m\n", + " 780 0.1789 23.2% 0.3m\n", + " 790 0.1502 23.2% 0.2m\n", + " 800 0.2096 23.2% 0.2m\n", + " 810 0.1542 23.2% 0.2m\n", + " 820 0.1560 23.2% 0.1m\n", + " 830 0.1510 23.2% 0.1m\n", + " 840 0.1567 23.2% 0.1m\n", + " 850 0.1493 23.2% 0.0m\n", + " 860 0.1303 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 53 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 53 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16673 โ†’ Val=0.16406\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 133m\n", + "\n", + "๐Ÿ”„ Epoch 54/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1719 23.2% 2.9m\n", + " 10 0.1752 23.2% 2.6m\n", + " 20 0.1501 23.2% 2.6m\n", + " 30 0.2110 23.2% 2.6m\n", + " 40 0.1647 23.2% 2.6m\n", + " 50 0.2085 23.2% 2.6m\n", + " 60 0.1486 23.2% 2.5m\n", + " 70 0.1738 23.2% 2.5m\n", + " 80 0.1782 23.2% 2.5m\n", + " 90 0.1509 23.2% 2.4m\n", + " 100 0.1870 23.2% 2.4m\n", + " 110 0.1678 23.2% 2.4m\n", + " 120 0.1323 23.2% 2.3m\n", + " 130 0.1745 23.2% 2.3m\n", + " 140 0.1707 23.2% 2.3m\n", + " 150 0.1687 23.2% 2.2m\n", + " 160 0.1347 23.2% 2.2m\n", + " 170 0.1338 23.2% 2.2m\n", + " 180 0.1478 23.2% 2.1m\n", + " 190 0.2004 23.2% 2.1m\n", + " 200 0.1585 23.2% 2.1m\n", + " 210 0.1715 23.2% 2.1m\n", + " 220 0.1808 23.2% 2.0m\n", + " 230 0.1867 23.2% 2.0m\n", + " 240 0.1719 23.2% 2.0m\n", + " 250 0.1444 23.2% 1.9m\n", + " 260 0.1935 23.2% 1.9m\n", + " 270 0.1608 23.2% 1.9m\n", + " 280 0.1435 23.2% 1.8m\n", + " 290 0.1765 23.2% 1.8m\n", + " 300 0.1872 23.2% 1.8m\n", + " 310 0.1704 23.2% 1.7m\n", + " 320 0.1210 23.2% 1.7m\n", + " 330 0.1440 23.2% 1.7m\n", + " 340 0.1608 23.2% 1.6m\n", + " 350 0.1664 23.2% 1.6m\n", + " 360 0.1742 23.2% 1.6m\n", + " 370 0.1998 23.2% 1.5m\n", + " 380 0.1581 23.2% 1.5m\n", + " 390 0.1607 23.2% 1.5m\n", + " 400 0.1477 23.2% 1.5m\n", + " 410 0.1571 23.2% 1.4m\n", + " 420 0.1718 23.2% 1.4m\n", + " 430 0.1408 23.2% 1.4m\n", + " 440 0.1727 23.2% 1.3m\n", + " 450 0.1624 23.2% 1.3m\n", + " 460 0.1555 23.2% 1.3m\n", + " 470 0.1431 23.2% 1.2m\n", + " 480 0.1598 23.2% 1.2m\n", + " 490 0.1527 23.2% 1.2m\n", + " 500 0.1508 23.2% 1.1m\n", + " 510 0.1562 23.2% 1.1m\n", + " 520 0.1538 23.2% 1.1m\n", + " 530 0.1634 23.2% 1.0m\n", + " 540 0.1420 23.2% 1.0m\n", + " 550 0.1594 23.2% 1.0m\n", + " 560 0.1925 23.2% 0.9m\n", + " 570 0.2026 23.2% 0.9m\n", + " 580 0.1784 23.2% 0.9m\n", + " 590 0.1453 23.2% 0.9m\n", + " 600 0.1720 23.2% 0.8m\n", + " 610 0.1929 23.2% 0.8m\n", + " 620 0.1485 23.2% 0.8m\n", + " 630 0.1621 23.2% 0.7m\n", + " 640 0.2041 23.2% 0.7m\n", + " 650 0.1379 23.2% 0.7m\n", + " 660 0.1739 23.2% 0.6m\n", + " 670 0.1457 23.2% 0.6m\n", + " 680 0.1509 23.2% 0.6m\n", + " 690 0.1701 23.2% 0.5m\n", + " 700 0.1735 23.2% 0.5m\n", + " 710 0.2307 23.2% 0.5m\n", + " 720 0.2016 23.2% 0.4m\n", + " 730 0.1782 23.2% 0.4m\n", + " 740 0.1693 23.2% 0.4m\n", + " 750 0.1714 23.2% 0.4m\n", + " 760 0.1430 23.2% 0.3m\n", + " 770 0.1595 23.2% 0.3m\n", + " 780 0.1286 23.2% 0.3m\n", + " 790 0.1728 23.2% 0.2m\n", + " 800 0.2263 23.2% 0.2m\n", + " 810 0.1371 23.2% 0.2m\n", + " 820 0.1764 23.2% 0.1m\n", + " 830 0.1953 23.2% 0.1m\n", + " 840 0.1409 23.2% 0.1m\n", + " 850 0.2235 23.2% 0.0m\n", + " 860 0.1004 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 54 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 54 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16667 โ†’ Val=0.16404\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 130m\n", + "\n", + "๐Ÿ”„ Epoch 55/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1304 23.2% 2.7m\n", + " 10 0.1609 23.2% 2.6m\n", + " 20 0.1322 23.2% 2.6m\n", + " 30 0.1739 23.2% 2.6m\n", + " 40 0.1760 23.2% 2.5m\n", + " 50 0.1643 23.2% 2.5m\n", + " 60 0.1547 23.2% 2.5m\n", + " 70 0.1660 23.2% 2.5m\n", + " 80 0.1478 23.2% 2.4m\n", + " 90 0.1548 23.2% 2.4m\n", + " 100 0.1784 23.2% 2.4m\n", + " 110 0.1468 23.2% 2.3m\n", + " 120 0.1750 23.2% 2.3m\n", + " 130 0.1953 23.2% 2.3m\n", + " 140 0.1599 23.2% 2.3m\n", + " 150 0.1710 23.2% 2.2m\n", + " 160 0.1930 23.2% 2.2m\n", + " 170 0.1710 23.2% 2.2m\n", + " 180 0.1596 23.2% 2.1m\n", + " 190 0.1511 23.2% 2.1m\n", + " 200 0.1706 23.2% 2.1m\n", + " 210 0.1497 23.2% 2.0m\n", + " 220 0.1528 23.2% 2.0m\n", + " 230 0.1534 23.2% 2.0m\n", + " 240 0.1067 23.2% 1.9m\n", + " 250 0.1519 23.2% 1.9m\n", + " 260 0.1621 23.2% 1.9m\n", + " 270 0.1283 23.2% 1.9m\n", + " 280 0.1603 23.2% 1.8m\n", + " 290 0.1499 23.2% 1.8m\n", + " 300 0.1334 23.2% 1.8m\n", + " 310 0.1628 23.2% 1.7m\n", + " 320 0.1822 23.2% 1.7m\n", + " 330 0.1663 23.2% 1.7m\n", + " 340 0.1652 23.2% 1.6m\n", + " 350 0.1572 23.2% 1.6m\n", + " 360 0.1732 23.2% 1.6m\n", + " 370 0.1553 23.2% 1.5m\n", + " 380 0.1779 23.2% 1.5m\n", + " 390 0.1871 23.2% 1.5m\n", + " 400 0.1575 23.2% 1.4m\n", + " 410 0.1504 23.2% 1.4m\n", + " 420 0.1538 23.2% 1.4m\n", + " 430 0.1132 23.2% 1.4m\n", + " 440 0.1411 23.2% 1.3m\n", + " 450 0.1717 23.2% 1.3m\n", + " 460 0.1422 23.2% 1.3m\n", + " 470 0.1306 23.2% 1.2m\n", + " 480 0.1772 23.2% 1.2m\n", + " 490 0.1530 23.2% 1.2m\n", + " 500 0.1755 23.2% 1.1m\n", + " 510 0.1355 23.2% 1.1m\n", + " 520 0.1379 23.2% 1.1m\n", + " 530 0.1653 23.2% 1.0m\n", + " 540 0.1736 23.2% 1.0m\n", + " 550 0.1647 23.2% 1.0m\n", + " 560 0.1871 23.2% 0.9m\n", + " 570 0.1968 23.2% 0.9m\n", + " 580 0.1855 23.2% 0.9m\n", + " 590 0.1695 23.2% 0.9m\n", + " 600 0.1735 23.2% 0.8m\n", + " 610 0.1587 23.2% 0.8m\n", + " 620 0.1804 23.2% 0.8m\n", + " 630 0.1303 23.2% 0.7m\n", + " 640 0.1726 23.2% 0.7m\n", + " 650 0.1439 23.2% 0.7m\n", + " 660 0.1679 23.2% 0.6m\n", + " 670 0.1394 23.2% 0.6m\n", + " 680 0.1735 23.2% 0.6m\n", + " 690 0.1739 23.2% 0.5m\n", + " 700 0.1590 23.2% 0.5m\n", + " 710 0.1813 23.2% 0.5m\n", + " 720 0.1617 23.2% 0.4m\n", + " 730 0.1580 23.2% 0.4m\n", + " 740 0.1621 23.2% 0.4m\n", + " 750 0.1526 23.2% 0.4m\n", + " 760 0.1675 23.2% 0.3m\n", + " 770 0.1462 23.2% 0.3m\n", + " 780 0.1977 23.2% 0.3m\n", + " 790 0.1657 23.2% 0.2m\n", + " 800 0.1711 23.2% 0.2m\n", + " 810 0.1860 23.2% 0.2m\n", + " 820 0.1413 23.2% 0.1m\n", + " 830 0.1772 23.2% 0.1m\n", + " 840 0.1412 23.2% 0.1m\n", + " 850 0.1755 23.2% 0.0m\n", + " 860 0.1405 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 55 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 55 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16669 โ†’ Val=0.16407\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 127m\n", + "\n", + "๐Ÿ”„ Epoch 56/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1600 23.2% 2.7m\n", + " 10 0.1490 23.2% 2.6m\n", + " 20 0.2142 23.2% 2.6m\n", + " 30 0.1677 23.2% 2.6m\n", + " 40 0.1587 23.2% 2.6m\n", + " 50 0.2010 23.2% 2.5m\n", + " 60 0.1885 23.2% 2.5m\n", + " 70 0.1760 23.2% 2.5m\n", + " 80 0.1811 23.2% 2.4m\n", + " 90 0.1616 23.2% 2.4m\n", + " 100 0.1821 23.2% 2.4m\n", + " 110 0.1537 23.2% 2.4m\n", + " 120 0.1337 23.2% 2.3m\n", + " 130 0.1557 23.2% 2.3m\n", + " 140 0.1626 23.2% 2.3m\n", + " 150 0.1730 23.2% 2.2m\n", + " 160 0.1984 23.2% 2.2m\n", + " 170 0.1809 23.2% 2.2m\n", + " 180 0.1845 23.2% 2.1m\n", + " 190 0.2071 23.2% 2.1m\n", + " 200 0.1600 23.2% 2.1m\n", + " 210 0.1765 23.2% 2.0m\n", + " 220 0.1867 23.2% 2.0m\n", + " 230 0.2149 23.2% 2.0m\n", + " 240 0.1071 23.2% 1.9m\n", + " 250 0.1585 23.2% 1.9m\n", + " 260 0.1550 23.2% 1.9m\n", + " 270 0.1732 23.2% 1.8m\n", + " 280 0.1252 23.2% 1.8m\n", + " 290 0.1232 23.2% 1.8m\n", + " 300 0.1610 23.2% 1.8m\n", + " 310 0.1589 23.2% 1.7m\n", + " 320 0.1858 23.2% 1.7m\n", + " 330 0.1759 23.2% 1.7m\n", + " 340 0.1268 23.2% 1.6m\n", + " 350 0.1756 23.2% 1.6m\n", + " 360 0.1878 23.2% 1.6m\n", + " 370 0.1668 23.2% 1.5m\n", + " 380 0.1859 23.2% 1.5m\n", + " 390 0.1730 23.2% 1.5m\n", + " 400 0.1470 23.2% 1.4m\n", + " 410 0.1743 23.2% 1.4m\n", + " 420 0.1706 23.2% 1.4m\n", + " 430 0.2061 23.2% 1.3m\n", + " 440 0.1829 23.2% 1.3m\n", + " 450 0.1418 23.2% 1.3m\n", + " 460 0.1562 23.2% 1.3m\n", + " 470 0.1866 23.2% 1.2m\n", + " 480 0.1668 23.2% 1.2m\n", + " 490 0.1950 23.2% 1.2m\n", + " 500 0.1495 23.2% 1.1m\n", + " 510 0.1639 23.2% 1.1m\n", + " 520 0.1886 23.2% 1.1m\n", + " 530 0.1538 23.2% 1.0m\n", + " 540 0.1841 23.2% 1.0m\n", + " 550 0.1560 23.2% 1.0m\n", + " 560 0.1700 23.2% 0.9m\n", + " 570 0.1604 23.2% 0.9m\n", + " 580 0.1562 23.2% 0.9m\n", + " 590 0.1841 23.2% 0.8m\n", + " 600 0.1599 23.2% 0.8m\n", + " 610 0.1427 23.2% 0.8m\n", + " 620 0.1639 23.2% 0.8m\n", + " 630 0.1683 23.2% 0.7m\n", + " 640 0.1835 23.2% 0.7m\n", + " 650 0.1591 23.2% 0.7m\n", + " 660 0.1690 23.2% 0.6m\n", + " 670 0.1809 23.2% 0.6m\n", + " 680 0.1832 23.2% 0.6m\n", + " 690 0.1693 23.2% 0.5m\n", + " 700 0.1701 23.2% 0.5m\n", + " 710 0.1679 23.2% 0.5m\n", + " 720 0.2075 23.2% 0.4m\n", + " 730 0.1630 23.2% 0.4m\n", + " 740 0.1691 23.2% 0.4m\n", + " 750 0.1250 23.2% 0.4m\n", + " 760 0.1623 23.2% 0.3m\n", + " 770 0.1800 23.2% 0.3m\n", + " 780 0.1777 23.2% 0.3m\n", + " 790 0.1630 23.2% 0.2m\n", + " 800 0.1773 23.2% 0.2m\n", + " 810 0.1638 23.2% 0.2m\n", + " 820 0.1213 23.2% 0.1m\n", + " 830 0.2062 23.2% 0.1m\n", + " 840 0.1583 23.2% 0.1m\n", + " 850 0.2057 23.2% 0.0m\n", + " 860 0.1501 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 56 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 56 SUMMARY\n", + " โฑ๏ธ Time: 169s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16665 โ†’ Val=0.16408\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 124m\n", + "\n", + "๐Ÿ”„ Epoch 57/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1816 23.2% 2.7m\n", + " 10 0.1606 23.2% 2.7m\n", + " 20 0.1607 23.2% 2.7m\n", + " 30 0.1511 23.2% 2.6m\n", + " 40 0.2000 23.2% 2.6m\n", + " 50 0.1401 23.2% 2.6m\n", + " 60 0.2117 23.2% 2.5m\n", + " 70 0.1622 23.2% 2.5m\n", + " 80 0.1644 23.2% 2.5m\n", + " 90 0.1510 23.2% 2.4m\n", + " 100 0.2003 23.2% 2.4m\n", + " 110 0.1572 23.2% 2.4m\n", + " 120 0.2166 23.2% 2.3m\n", + " 130 0.1439 23.2% 2.3m\n", + " 140 0.1651 23.2% 2.3m\n", + " 150 0.1702 23.2% 2.2m\n", + " 160 0.1323 23.2% 2.2m\n", + " 170 0.1507 23.2% 2.2m\n", + " 180 0.1652 23.2% 2.2m\n", + " 190 0.1570 23.2% 2.1m\n", + " 200 0.1839 23.2% 2.1m\n", + " 210 0.1583 23.2% 2.1m\n", + " 220 0.1617 23.2% 2.0m\n", + " 230 0.1381 23.2% 2.0m\n", + " 240 0.1410 23.2% 2.0m\n", + " 250 0.1951 23.2% 1.9m\n", + " 260 0.1670 23.2% 1.9m\n", + " 270 0.1851 23.2% 1.9m\n", + " 280 0.2040 23.2% 1.8m\n", + " 290 0.1825 23.2% 1.8m\n", + " 300 0.1727 23.2% 1.8m\n", + " 310 0.1202 23.2% 1.7m\n", + " 320 0.1682 23.2% 1.7m\n", + " 330 0.1485 23.2% 1.7m\n", + " 340 0.1660 23.2% 1.6m\n", + " 350 0.2154 23.2% 1.6m\n", + " 360 0.1614 23.2% 1.6m\n", + " 370 0.1783 23.2% 1.6m\n", + " 380 0.1544 23.2% 1.5m\n", + " 390 0.1463 23.2% 1.5m\n", + " 400 0.1352 23.2% 1.5m\n", + " 410 0.1842 23.2% 1.4m\n", + " 420 0.1945 23.2% 1.4m\n", + " 430 0.1672 23.2% 1.4m\n", + " 440 0.1700 23.2% 1.3m\n", + " 450 0.1462 23.2% 1.3m\n", + " 460 0.1861 23.2% 1.3m\n", + " 470 0.1630 23.2% 1.2m\n", + " 480 0.1615 23.2% 1.2m\n", + " 490 0.2109 23.2% 1.2m\n", + " 500 0.1573 23.2% 1.1m\n", + " 510 0.1770 23.2% 1.1m\n", + " 520 0.1934 23.2% 1.1m\n", + " 530 0.2170 23.2% 1.0m\n", + " 540 0.1648 23.2% 1.0m\n", + " 550 0.2179 23.2% 1.0m\n", + " 560 0.1799 23.2% 1.0m\n", + " 570 0.1585 23.2% 0.9m\n", + " 580 0.1768 23.2% 0.9m\n", + " 590 0.1307 23.2% 0.9m\n", + " 600 0.1854 23.2% 0.8m\n", + " 610 0.1826 23.2% 0.8m\n", + " 620 0.1898 23.2% 0.8m\n", + " 630 0.1756 23.2% 0.7m\n", + " 640 0.1894 23.2% 0.7m\n", + " 650 0.1530 23.2% 0.7m\n", + " 660 0.1608 23.2% 0.6m\n", + " 670 0.1618 23.2% 0.6m\n", + " 680 0.1387 23.2% 0.6m\n", + " 690 0.1349 23.2% 0.5m\n", + " 700 0.1754 23.2% 0.5m\n", + " 710 0.1707 23.2% 0.5m\n", + " 720 0.1706 23.2% 0.4m\n", + " 730 0.1822 23.2% 0.4m\n", + " 740 0.1314 23.2% 0.4m\n", + " 750 0.1497 23.2% 0.4m\n", + " 760 0.1821 23.2% 0.3m\n", + " 770 0.1302 23.2% 0.3m\n", + " 780 0.1417 23.2% 0.3m\n", + " 790 0.1683 23.2% 0.2m\n", + " 800 0.1920 23.2% 0.2m\n", + " 810 0.1771 23.2% 0.2m\n", + " 820 0.1429 23.2% 0.1m\n", + " 830 0.1717 23.2% 0.1m\n", + " 840 0.1752 23.2% 0.1m\n", + " 850 0.1386 23.2% 0.0m\n", + " 860 0.1722 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 57 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 57 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16672 โ†’ Val=0.16393\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16393)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 122m\n", + "\n", + "๐Ÿ”„ Epoch 58/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1805 23.2% 2.7m\n", + " 10 0.1533 23.2% 2.7m\n", + " 20 0.1760 23.2% 2.7m\n", + " 30 0.1390 23.2% 2.7m\n", + " 40 0.1423 23.2% 2.7m\n", + " 50 0.2147 23.2% 2.6m\n", + " 60 0.1676 23.2% 2.6m\n", + " 70 0.1635 23.2% 2.5m\n", + " 80 0.1496 23.2% 2.5m\n", + " 90 0.1604 23.2% 2.5m\n", + " 100 0.1827 23.2% 2.4m\n", + " 110 0.1429 23.2% 2.4m\n", + " 120 0.1557 23.2% 2.4m\n", + " 130 0.1418 23.2% 2.3m\n", + " 140 0.1470 23.2% 2.3m\n", + " 150 0.1457 23.2% 2.3m\n", + " 160 0.1664 23.2% 2.2m\n", + " 170 0.1664 23.2% 2.2m\n", + " 180 0.1695 23.2% 2.2m\n", + " 190 0.1630 23.2% 2.1m\n", + " 200 0.1460 23.2% 2.1m\n", + " 210 0.1921 23.2% 2.1m\n", + " 220 0.1789 23.2% 2.0m\n", + " 230 0.1533 23.2% 2.0m\n", + " 240 0.1790 23.2% 2.0m\n", + " 250 0.1624 23.2% 1.9m\n", + " 260 0.1662 23.2% 1.9m\n", + " 270 0.1982 23.2% 1.9m\n", + " 280 0.1499 23.2% 1.8m\n", + " 290 0.1605 23.2% 1.8m\n", + " 300 0.1490 23.2% 1.8m\n", + " 310 0.1861 23.2% 1.7m\n", + " 320 0.2172 23.2% 1.7m\n", + " 330 0.1455 23.2% 1.7m\n", + " 340 0.1662 23.2% 1.6m\n", + " 350 0.1323 23.2% 1.6m\n", + " 360 0.1672 23.2% 1.6m\n", + " 370 0.1848 23.2% 1.6m\n", + " 380 0.1881 23.2% 1.5m\n", + " 390 0.1721 23.2% 1.5m\n", + " 400 0.1486 23.2% 1.5m\n", + " 410 0.1446 23.2% 1.4m\n", + " 420 0.1414 23.2% 1.4m\n", + " 430 0.1904 23.2% 1.4m\n", + " 440 0.1690 23.2% 1.3m\n", + " 450 0.1423 23.2% 1.3m\n", + " 460 0.1767 23.2% 1.3m\n", + " 470 0.1817 23.2% 1.2m\n", + " 480 0.1581 23.2% 1.2m\n", + " 490 0.1196 23.2% 1.2m\n", + " 500 0.1852 23.2% 1.1m\n", + " 510 0.1435 23.2% 1.1m\n", + " 520 0.1607 23.2% 1.1m\n", + " 530 0.1774 23.2% 1.0m\n", + " 540 0.1436 23.2% 1.0m\n", + " 550 0.1807 23.2% 1.0m\n", + " 560 0.1624 23.2% 1.0m\n", + " 570 0.1578 23.2% 0.9m\n", + " 580 0.2074 23.2% 0.9m\n", + " 590 0.1508 23.2% 0.9m\n", + " 600 0.1459 23.2% 0.8m\n", + " 610 0.1473 23.2% 0.8m\n", + " 620 0.1239 23.2% 0.8m\n", + " 630 0.1763 23.2% 0.7m\n", + " 640 0.1703 23.2% 0.7m\n", + " 650 0.1468 23.2% 0.7m\n", + " 660 0.1739 23.2% 0.6m\n", + " 670 0.1605 23.2% 0.6m\n", + " 680 0.1810 23.2% 0.6m\n", + " 690 0.1699 23.2% 0.5m\n", + " 700 0.1816 23.2% 0.5m\n", + " 710 0.1726 23.2% 0.5m\n", + " 720 0.1286 23.2% 0.4m\n", + " 730 0.1286 23.2% 0.4m\n", + " 740 0.1931 23.2% 0.4m\n", + " 750 0.1932 23.2% 0.4m\n", + " 760 0.1655 23.2% 0.3m\n", + " 770 0.1722 23.2% 0.3m\n", + " 780 0.1741 23.2% 0.3m\n", + " 790 0.1230 23.2% 0.2m\n", + " 800 0.1914 23.2% 0.2m\n", + " 810 0.1796 23.2% 0.2m\n", + " 820 0.1488 23.2% 0.1m\n", + " 830 0.1578 23.2% 0.1m\n", + " 840 0.1857 23.2% 0.1m\n", + " 850 0.1528 23.2% 0.0m\n", + " 860 0.1718 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 58 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 58 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16668 โ†’ Val=0.16396\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 119m\n", + "\n", + "๐Ÿ”„ Epoch 59/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1857 23.2% 2.7m\n", + " 10 0.2005 23.2% 2.7m\n", + " 20 0.1689 23.2% 2.6m\n", + " 30 0.1472 23.2% 2.6m\n", + " 40 0.1742 23.2% 2.6m\n", + " 50 0.1909 23.2% 2.5m\n", + " 60 0.2001 23.2% 2.5m\n", + " 70 0.1393 23.2% 2.5m\n", + " 80 0.1696 23.2% 2.4m\n", + " 90 0.1688 23.2% 2.4m\n", + " 100 0.1860 23.2% 2.4m\n", + " 110 0.1501 23.2% 2.4m\n", + " 120 0.1596 23.2% 2.3m\n", + " 130 0.1543 23.2% 2.3m\n", + " 140 0.1775 23.2% 2.3m\n", + " 150 0.1560 23.2% 2.2m\n", + " 160 0.1588 23.2% 2.2m\n", + " 170 0.1627 23.2% 2.2m\n", + " 180 0.1617 23.2% 2.1m\n", + " 190 0.1942 23.2% 2.1m\n", + " 200 0.2044 23.2% 2.1m\n", + " 210 0.1568 23.2% 2.0m\n", + " 220 0.1468 23.2% 2.0m\n", + " 230 0.1753 23.2% 2.0m\n", + " 240 0.1557 23.2% 1.9m\n", + " 250 0.1549 23.2% 1.9m\n", + " 260 0.1796 23.2% 1.9m\n", + " 270 0.1688 23.2% 1.8m\n", + " 280 0.1801 23.2% 1.8m\n", + " 290 0.1473 23.2% 1.8m\n", + " 300 0.2012 23.2% 1.8m\n", + " 310 0.1586 23.2% 1.7m\n", + " 320 0.1724 23.2% 1.7m\n", + " 330 0.1549 23.2% 1.7m\n", + " 340 0.1714 23.2% 1.6m\n", + " 350 0.1766 23.2% 1.6m\n", + " 360 0.1535 23.2% 1.6m\n", + " 370 0.1705 23.2% 1.5m\n", + " 380 0.1806 23.2% 1.5m\n", + " 390 0.1548 23.2% 1.5m\n", + " 400 0.1405 23.2% 1.4m\n", + " 410 0.1674 23.2% 1.4m\n", + " 420 0.1567 23.2% 1.4m\n", + " 430 0.1475 23.2% 1.4m\n", + " 440 0.1709 23.2% 1.3m\n", + " 450 0.1877 23.2% 1.3m\n", + " 460 0.1635 23.2% 1.3m\n", + " 470 0.1513 23.2% 1.2m\n", + " 480 0.1298 23.2% 1.2m\n", + " 490 0.1789 23.2% 1.2m\n", + " 500 0.1323 23.2% 1.1m\n", + " 510 0.1765 23.2% 1.1m\n", + " 520 0.1898 23.2% 1.1m\n", + " 530 0.1221 23.2% 1.0m\n", + " 540 0.1391 23.2% 1.0m\n", + " 550 0.1877 23.2% 1.0m\n", + " 560 0.1565 23.2% 0.9m\n", + " 570 0.1613 23.2% 0.9m\n", + " 580 0.1628 23.2% 0.9m\n", + " 590 0.1717 23.2% 0.9m\n", + " 600 0.1580 23.2% 0.8m\n", + " 610 0.1701 23.2% 0.8m\n", + " 620 0.1386 23.2% 0.8m\n", + " 630 0.1509 23.2% 0.7m\n", + " 640 0.1746 23.2% 0.7m\n", + " 650 0.1763 23.2% 0.7m\n", + " 660 0.1279 23.2% 0.6m\n", + " 670 0.1492 23.2% 0.6m\n", + " 680 0.1627 23.2% 0.6m\n", + " 690 0.1569 23.2% 0.5m\n", + " 700 0.1849 23.2% 0.5m\n", + " 710 0.1933 23.2% 0.5m\n", + " 720 0.1214 23.2% 0.4m\n", + " 730 0.1595 23.2% 0.4m\n", + " 740 0.1827 23.2% 0.4m\n", + " 750 0.2011 23.2% 0.4m\n", + " 760 0.1905 23.2% 0.3m\n", + " 770 0.1497 23.2% 0.3m\n", + " 780 0.1830 23.2% 0.3m\n", + " 790 0.1529 23.2% 0.2m\n", + " 800 0.1657 23.2% 0.2m\n", + " 810 0.1551 23.2% 0.2m\n", + " 820 0.1831 23.2% 0.1m\n", + " 830 0.1800 23.2% 0.1m\n", + " 840 0.2052 23.2% 0.1m\n", + " 850 0.1707 23.2% 0.0m\n", + " 860 0.1808 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 59 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 59 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16659 โ†’ Val=0.16415\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 116m\n", + "\n", + "๐Ÿ”„ Epoch 60/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1700 23.2% 2.7m\n", + " 10 0.1300 23.2% 2.7m\n", + " 20 0.1633 23.2% 2.7m\n", + " 30 0.1520 23.2% 2.6m\n", + " 40 0.1526 23.2% 2.6m\n", + " 50 0.1821 23.2% 2.5m\n", + " 60 0.1514 23.2% 2.5m\n", + " 70 0.1776 23.2% 2.5m\n", + " 80 0.1361 23.2% 2.4m\n", + " 90 0.1424 23.2% 2.4m\n", + " 100 0.1443 23.2% 2.4m\n", + " 110 0.2094 23.2% 2.3m\n", + " 120 0.1589 23.2% 2.3m\n", + " 130 0.1820 23.2% 2.3m\n", + " 140 0.1421 23.2% 2.3m\n", + " 150 0.1814 23.2% 2.2m\n", + " 160 0.1480 23.2% 2.2m\n", + " 170 0.1036 23.2% 2.2m\n", + " 180 0.1336 23.2% 2.1m\n", + " 190 0.1637 23.2% 2.1m\n", + " 200 0.1946 23.2% 2.1m\n", + " 210 0.1452 23.2% 2.0m\n", + " 220 0.1852 23.2% 2.0m\n", + " 230 0.1762 23.2% 2.0m\n", + " 240 0.1445 23.2% 1.9m\n", + " 250 0.1458 23.2% 1.9m\n", + " 260 0.1592 23.2% 1.9m\n", + " 270 0.1712 23.2% 1.9m\n", + " 280 0.1689 23.2% 1.8m\n", + " 290 0.1452 23.2% 1.8m\n", + " 300 0.1481 23.2% 1.8m\n", + " 310 0.1831 23.2% 1.7m\n", + " 320 0.1714 23.2% 1.7m\n", + " 330 0.1911 23.2% 1.7m\n", + " 340 0.1704 23.2% 1.6m\n", + " 350 0.1761 23.2% 1.6m\n", + " 360 0.1593 23.2% 1.6m\n", + " 370 0.1836 23.2% 1.5m\n", + " 380 0.1571 23.2% 1.5m\n", + " 390 0.1380 23.2% 1.5m\n", + " 400 0.2413 23.2% 1.5m\n", + " 410 0.1709 23.2% 1.4m\n", + " 420 0.1738 23.2% 1.4m\n", + " 430 0.1470 23.2% 1.4m\n", + " 440 0.1308 23.2% 1.3m\n", + " 450 0.1856 23.2% 1.3m\n", + " 460 0.1222 23.2% 1.3m\n", + " 470 0.1734 23.2% 1.2m\n", + " 480 0.2228 23.2% 1.2m\n", + " 490 0.1349 23.2% 1.2m\n", + " 500 0.1642 23.2% 1.1m\n", + " 510 0.1732 23.2% 1.1m\n", + " 520 0.1492 23.2% 1.1m\n", + " 530 0.1664 23.2% 1.0m\n", + " 540 0.1679 23.2% 1.0m\n", + " 550 0.1575 23.2% 1.0m\n", + " 560 0.1765 23.2% 0.9m\n", + " 570 0.1482 23.2% 0.9m\n", + " 580 0.1583 23.2% 0.9m\n", + " 590 0.1968 23.2% 0.9m\n", + " 600 0.1839 23.2% 0.8m\n", + " 610 0.1750 23.2% 0.8m\n", + " 620 0.1915 23.2% 0.8m\n", + " 630 0.1829 23.2% 0.7m\n", + " 640 0.1779 23.2% 0.7m\n", + " 650 0.1903 23.2% 0.7m\n", + " 660 0.1582 23.2% 0.6m\n", + " 670 0.1651 23.2% 0.6m\n", + " 680 0.1631 23.2% 0.6m\n", + " 690 0.1663 23.2% 0.5m\n", + " 700 0.2024 23.2% 0.5m\n", + " 710 0.1568 23.2% 0.5m\n", + " 720 0.1638 23.2% 0.4m\n", + " 730 0.1774 23.2% 0.4m\n", + " 740 0.1762 23.2% 0.4m\n", + " 750 0.1815 23.2% 0.4m\n", + " 760 0.1857 23.2% 0.3m\n", + " 770 0.1511 23.2% 0.3m\n", + " 780 0.1329 23.2% 0.3m\n", + " 790 0.2023 23.2% 0.2m\n", + " 800 0.1790 23.2% 0.2m\n", + " 810 0.1434 23.2% 0.2m\n", + " 820 0.1847 23.2% 0.1m\n", + " 830 0.1608 23.2% 0.1m\n", + " 840 0.1548 23.2% 0.1m\n", + " 850 0.1361 23.2% 0.0m\n", + " 860 0.1536 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 60 COMPLETE in 161s\n", + "\n", + "โœ… EPOCH 60 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16666 โ†’ Val=0.16479\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 113m\n", + "\n", + "๐Ÿ”„ Epoch 61/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1917 23.2% 2.6m\n", + " 10 0.1465 23.2% 2.6m\n", + " 20 0.1421 23.2% 2.6m\n", + " 30 0.1211 23.2% 2.6m\n", + " 40 0.1644 23.2% 2.6m\n", + " 50 0.1882 23.2% 2.5m\n", + " 60 0.1439 23.2% 2.5m\n", + " 70 0.2124 23.2% 2.5m\n", + " 80 0.1227 23.2% 2.4m\n", + " 90 0.1368 23.2% 2.4m\n", + " 100 0.1683 23.2% 2.4m\n", + " 110 0.1607 23.2% 2.4m\n", + " 120 0.1807 23.2% 2.3m\n", + " 130 0.1608 23.2% 2.3m\n", + " 140 0.2018 23.2% 2.3m\n", + " 150 0.1703 23.2% 2.2m\n", + " 160 0.1639 23.2% 2.2m\n", + " 170 0.1735 23.2% 2.2m\n", + " 180 0.1432 23.2% 2.1m\n", + " 190 0.1481 23.2% 2.1m\n", + " 200 0.1816 23.2% 2.1m\n", + " 210 0.1888 23.2% 2.0m\n", + " 220 0.1689 23.2% 2.0m\n", + " 230 0.1634 23.2% 2.0m\n", + " 240 0.1818 23.2% 1.9m\n", + " 250 0.1354 23.2% 1.9m\n", + " 260 0.1891 23.2% 1.9m\n", + " 270 0.2137 23.2% 1.9m\n", + " 280 0.1592 23.2% 1.8m\n", + " 290 0.1656 23.2% 1.8m\n", + " 300 0.1681 23.2% 1.8m\n", + " 310 0.1781 23.2% 1.7m\n", + " 320 0.1363 23.2% 1.7m\n", + " 330 0.1632 23.2% 1.7m\n", + " 340 0.1440 23.2% 1.6m\n", + " 350 0.2139 23.2% 1.6m\n", + " 360 0.1604 23.2% 1.6m\n", + " 370 0.1651 23.2% 1.5m\n", + " 380 0.1411 23.2% 1.5m\n", + " 390 0.1484 23.2% 1.5m\n", + " 400 0.1162 23.2% 1.4m\n", + " 410 0.1491 23.2% 1.4m\n", + " 420 0.1257 23.2% 1.4m\n", + " 430 0.1451 23.2% 1.4m\n", + " 440 0.1977 23.2% 1.3m\n", + " 450 0.1590 23.2% 1.3m\n", + " 460 0.1267 23.2% 1.3m\n", + " 470 0.1766 23.2% 1.2m\n", + " 480 0.1840 23.2% 1.2m\n", + " 490 0.1906 23.2% 1.2m\n", + " 500 0.1732 23.2% 1.1m\n", + " 510 0.1422 23.2% 1.1m\n", + " 520 0.1704 23.2% 1.1m\n", + " 530 0.2108 23.2% 1.0m\n", + " 540 0.1768 23.2% 1.0m\n", + " 550 0.1669 23.2% 1.0m\n", + " 560 0.1598 23.2% 0.9m\n", + " 570 0.1750 23.2% 0.9m\n", + " 580 0.1383 23.2% 0.9m\n", + " 590 0.1786 23.2% 0.9m\n", + " 600 0.1887 23.2% 0.8m\n", + " 610 0.1361 23.2% 0.8m\n", + " 620 0.1457 23.2% 0.8m\n", + " 630 0.1857 23.2% 0.7m\n", + " 640 0.1987 23.2% 0.7m\n", + " 650 0.1432 23.2% 0.7m\n", + " 660 0.1399 23.2% 0.6m\n", + " 670 0.2035 23.2% 0.6m\n", + " 680 0.2007 23.2% 0.6m\n", + " 690 0.2008 23.2% 0.5m\n", + " 700 0.1693 23.2% 0.5m\n", + " 710 0.1439 23.2% 0.5m\n", + " 720 0.1550 23.2% 0.4m\n", + " 730 0.1651 23.2% 0.4m\n", + " 740 0.1985 23.2% 0.4m\n", + " 750 0.1735 23.2% 0.4m\n", + " 760 0.1626 23.2% 0.3m\n", + " 770 0.1656 23.2% 0.3m\n", + " 780 0.1697 23.2% 0.3m\n", + " 790 0.1729 23.2% 0.2m\n", + " 800 0.1951 23.2% 0.2m\n", + " 810 0.1622 23.2% 0.2m\n", + " 820 0.2239 23.2% 0.1m\n", + " 830 0.1393 23.2% 0.1m\n", + " 840 0.1483 23.2% 0.1m\n", + " 850 0.1384 23.2% 0.0m\n", + " 860 0.1995 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 61 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 61 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16664 โ†’ Val=0.16383\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16383)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 111m\n", + "\n", + "๐Ÿ”„ Epoch 62/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1784 23.2% 2.7m\n", + " 10 0.1716 23.2% 2.7m\n", + " 20 0.1580 23.2% 2.7m\n", + " 30 0.1623 23.2% 2.6m\n", + " 40 0.1605 23.2% 2.6m\n", + " 50 0.1758 23.2% 2.6m\n", + " 60 0.1813 23.2% 2.5m\n", + " 70 0.1720 23.2% 2.5m\n", + " 80 0.1528 23.2% 2.5m\n", + " 90 0.1459 23.2% 2.4m\n", + " 100 0.1602 23.2% 2.4m\n", + " 110 0.1466 23.2% 2.4m\n", + " 120 0.1164 23.2% 2.4m\n", + " 130 0.1943 23.2% 2.3m\n", + " 140 0.1475 23.2% 2.3m\n", + " 150 0.1384 23.2% 2.3m\n", + " 160 0.1885 23.2% 2.2m\n", + " 170 0.1945 23.2% 2.2m\n", + " 180 0.2034 23.2% 2.2m\n", + " 190 0.1835 23.2% 2.1m\n", + " 200 0.1872 23.2% 2.1m\n", + " 210 0.1817 23.2% 2.1m\n", + " 220 0.1860 23.2% 2.0m\n", + " 230 0.1697 23.2% 2.0m\n", + " 240 0.1999 23.2% 2.0m\n", + " 250 0.1818 23.2% 1.9m\n", + " 260 0.1691 23.2% 1.9m\n", + " 270 0.1565 23.2% 1.9m\n", + " 280 0.1915 23.2% 1.8m\n", + " 290 0.2024 23.2% 1.8m\n", + " 300 0.1694 23.2% 1.8m\n", + " 310 0.1534 23.2% 1.7m\n", + " 320 0.1809 23.2% 1.7m\n", + " 330 0.1649 23.2% 1.7m\n", + " 340 0.1469 23.2% 1.6m\n", + " 350 0.1698 23.2% 1.6m\n", + " 360 0.1610 23.2% 1.6m\n", + " 370 0.1771 23.2% 1.6m\n", + " 380 0.1594 23.2% 1.5m\n", + " 390 0.1938 23.2% 1.5m\n", + " 400 0.1858 23.2% 1.5m\n", + " 410 0.1513 23.2% 1.4m\n", + " 420 0.1482 23.2% 1.4m\n", + " 430 0.1768 23.2% 1.4m\n", + " 440 0.1600 23.2% 1.3m\n", + " 450 0.1723 23.2% 1.3m\n", + " 460 0.1560 23.2% 1.3m\n", + " 470 0.2107 23.2% 1.2m\n", + " 480 0.1556 23.2% 1.2m\n", + " 490 0.1607 23.2% 1.2m\n", + " 500 0.1495 23.2% 1.1m\n", + " 510 0.1382 23.2% 1.1m\n", + " 520 0.1750 23.2% 1.1m\n", + " 530 0.1503 23.2% 1.0m\n", + " 540 0.1718 23.2% 1.0m\n", + " 550 0.1429 23.2% 1.0m\n", + " 560 0.2241 23.2% 1.0m\n", + " 570 0.1659 23.2% 0.9m\n", + " 580 0.1700 23.2% 0.9m\n", + " 590 0.1895 23.2% 0.9m\n", + " 600 0.1780 23.2% 0.8m\n", + " 610 0.1693 23.2% 0.8m\n", + " 620 0.1710 23.2% 0.8m\n", + " 630 0.1830 23.2% 0.7m\n", + " 640 0.1587 23.2% 0.7m\n", + " 650 0.1551 23.2% 0.7m\n", + " 660 0.1462 23.2% 0.6m\n", + " 670 0.1504 23.2% 0.6m\n", + " 680 0.1468 23.2% 0.6m\n", + " 690 0.1332 23.2% 0.5m\n", + " 700 0.1260 23.2% 0.5m\n", + " 710 0.1187 23.2% 0.5m\n", + " 720 0.1556 23.2% 0.5m\n", + " 730 0.1573 23.2% 0.4m\n", + " 740 0.1933 23.2% 0.4m\n", + " 750 0.1728 23.2% 0.4m\n", + " 760 0.2077 23.2% 0.3m\n", + " 770 0.1570 23.2% 0.3m\n", + " 780 0.1974 23.2% 0.3m\n", + " 790 0.1227 23.2% 0.2m\n", + " 800 0.1823 23.2% 0.2m\n", + " 810 0.1781 23.2% 0.2m\n", + " 820 0.1654 23.2% 0.1m\n", + " 830 0.1354 23.2% 0.1m\n", + " 840 0.1989 23.2% 0.1m\n", + " 850 0.1620 23.2% 0.0m\n", + " 860 0.1752 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 62 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 62 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16654 โ†’ Val=0.16476\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.4% EJ=86.6% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 108m\n", + "\n", + "๐Ÿ”„ Epoch 63/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1730 23.2% 3.0m\n", + " 10 0.1466 23.2% 2.7m\n", + " 20 0.1656 23.2% 2.7m\n", + " 30 0.1974 23.2% 2.6m\n", + " 40 0.1929 23.2% 2.6m\n", + " 50 0.1635 23.2% 2.5m\n", + " 60 0.1577 23.2% 2.5m\n", + " 70 0.1678 23.2% 2.5m\n", + " 80 0.1884 23.2% 2.5m\n", + " 90 0.1683 23.2% 2.4m\n", + " 100 0.1340 23.2% 2.4m\n", + " 110 0.1227 23.2% 2.4m\n", + " 120 0.1685 23.2% 2.3m\n", + " 130 0.1191 23.2% 2.3m\n", + " 140 0.1490 23.2% 2.3m\n", + " 150 0.1743 23.2% 2.2m\n", + " 160 0.1493 23.2% 2.2m\n", + " 170 0.1604 23.2% 2.2m\n", + " 180 0.1930 23.2% 2.1m\n", + " 190 0.1677 23.2% 2.1m\n", + " 200 0.1669 23.2% 2.1m\n", + " 210 0.1527 23.2% 2.0m\n", + " 220 0.1412 23.2% 2.0m\n", + " 230 0.1895 23.2% 2.0m\n", + " 240 0.1646 23.2% 1.9m\n", + " 250 0.1773 23.2% 1.9m\n", + " 260 0.1683 23.2% 1.9m\n", + " 270 0.1860 23.2% 1.9m\n", + " 280 0.1783 23.2% 1.8m\n", + " 290 0.1639 23.2% 1.8m\n", + " 300 0.2027 23.2% 1.8m\n", + " 310 0.1786 23.2% 1.7m\n", + " 320 0.1964 23.2% 1.7m\n", + " 330 0.1590 23.2% 1.7m\n", + " 340 0.1969 23.2% 1.6m\n", + " 350 0.1663 23.2% 1.6m\n", + " 360 0.1773 23.2% 1.6m\n", + " 370 0.1135 23.2% 1.5m\n", + " 380 0.1421 23.2% 1.5m\n", + " 390 0.1503 23.2% 1.5m\n", + " 400 0.1906 23.2% 1.4m\n", + " 410 0.1349 23.2% 1.4m\n", + " 420 0.1657 23.2% 1.4m\n", + " 430 0.1349 23.2% 1.4m\n", + " 440 0.1756 23.2% 1.3m\n", + " 450 0.1501 23.2% 1.3m\n", + " 460 0.1482 23.2% 1.3m\n", + " 470 0.1577 23.2% 1.2m\n", + " 480 0.1499 23.2% 1.2m\n", + " 490 0.1251 23.2% 1.2m\n", + " 500 0.1327 23.2% 1.1m\n", + " 510 0.1828 23.2% 1.1m\n", + " 520 0.1536 23.2% 1.1m\n", + " 530 0.1974 23.2% 1.0m\n", + " 540 0.1741 23.2% 1.0m\n", + " 550 0.1732 23.2% 1.0m\n", + " 560 0.1803 23.2% 0.9m\n", + " 570 0.1586 23.2% 0.9m\n", + " 580 0.1682 23.2% 0.9m\n", + " 590 0.1858 23.2% 0.9m\n", + " 600 0.1951 23.2% 0.8m\n", + " 610 0.1938 23.2% 0.8m\n", + " 620 0.1511 23.2% 0.8m\n", + " 630 0.1774 23.2% 0.7m\n", + " 640 0.1920 23.2% 0.7m\n", + " 650 0.1600 23.2% 0.7m\n", + " 660 0.1894 23.2% 0.6m\n", + " 670 0.1417 23.2% 0.6m\n", + " 680 0.1553 23.2% 0.6m\n", + " 690 0.1631 23.2% 0.5m\n", + " 700 0.2005 23.2% 0.5m\n", + " 710 0.1414 23.2% 0.5m\n", + " 720 0.1252 23.2% 0.4m\n", + " 730 0.1787 23.2% 0.4m\n", + " 740 0.1928 23.2% 0.4m\n", + " 750 0.1531 23.2% 0.4m\n", + " 760 0.1421 23.2% 0.3m\n", + " 770 0.1688 23.2% 0.3m\n", + " 780 0.1894 23.2% 0.3m\n", + " 790 0.2329 23.2% 0.2m\n", + " 800 0.1660 23.2% 0.2m\n", + " 810 0.1463 23.2% 0.2m\n", + " 820 0.1815 23.2% 0.1m\n", + " 830 0.1795 23.2% 0.1m\n", + " 840 0.1751 23.2% 0.1m\n", + " 850 0.1766 23.2% 0.0m\n", + " 860 0.1618 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 63 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 63 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16670 โ†’ Val=0.16395\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 105m\n", + "\n", + "๐Ÿ”„ Epoch 64/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1844 23.2% 2.7m\n", + " 10 0.1650 23.2% 2.7m\n", + " 20 0.2116 23.2% 2.7m\n", + " 30 0.1603 23.2% 2.7m\n", + " 40 0.2244 23.2% 2.6m\n", + " 50 0.1894 23.2% 2.6m\n", + " 60 0.1731 23.2% 2.5m\n", + " 70 0.1483 23.2% 2.5m\n", + " 80 0.2023 23.2% 2.5m\n", + " 90 0.1732 23.2% 2.4m\n", + " 100 0.1690 23.2% 2.4m\n", + " 110 0.1857 23.2% 2.4m\n", + " 120 0.1718 23.2% 2.3m\n", + " 130 0.1274 23.2% 2.3m\n", + " 140 0.2050 23.2% 2.3m\n", + " 150 0.1678 23.2% 2.2m\n", + " 160 0.1844 23.2% 2.2m\n", + " 170 0.1793 23.2% 2.2m\n", + " 180 0.1568 23.2% 2.1m\n", + " 190 0.1683 23.2% 2.1m\n", + " 200 0.1537 23.2% 2.1m\n", + " 210 0.1712 23.2% 2.1m\n", + " 220 0.2206 23.2% 2.0m\n", + " 230 0.1758 23.2% 2.0m\n", + " 240 0.1590 23.2% 2.0m\n", + " 250 0.1608 23.2% 1.9m\n", + " 260 0.1446 23.2% 1.9m\n", + " 270 0.1851 23.2% 1.9m\n", + " 280 0.1640 23.2% 1.8m\n", + " 290 0.1589 23.2% 1.8m\n", + " 300 0.1994 23.2% 1.8m\n", + " 310 0.1334 23.2% 1.7m\n", + " 320 0.1635 23.2% 1.7m\n", + " 330 0.1459 23.2% 1.7m\n", + " 340 0.1694 23.2% 1.6m\n", + " 350 0.1541 23.2% 1.6m\n", + " 360 0.2121 23.2% 1.6m\n", + " 370 0.1709 23.2% 1.6m\n", + " 380 0.1787 23.2% 1.5m\n", + " 390 0.1678 23.2% 1.5m\n", + " 400 0.1944 23.2% 1.5m\n", + " 410 0.1758 23.2% 1.4m\n", + " 420 0.2177 23.2% 1.4m\n", + " 430 0.1940 23.2% 1.4m\n", + " 440 0.1594 23.2% 1.3m\n", + " 450 0.1503 23.2% 1.3m\n", + " 460 0.1750 23.2% 1.3m\n", + " 470 0.1846 23.2% 1.2m\n", + " 480 0.1284 23.2% 1.2m\n", + " 490 0.1328 23.2% 1.2m\n", + " 500 0.1970 23.2% 1.1m\n", + " 510 0.1446 23.2% 1.1m\n", + " 520 0.1471 23.2% 1.1m\n", + " 530 0.1642 23.2% 1.0m\n", + " 540 0.1936 23.2% 1.0m\n", + " 550 0.1791 23.2% 1.0m\n", + " 560 0.1296 23.2% 1.0m\n", + " 570 0.1761 23.2% 0.9m\n", + " 580 0.1138 23.2% 0.9m\n", + " 590 0.1745 23.2% 0.9m\n", + " 600 0.2255 23.2% 0.8m\n", + " 610 0.1779 23.2% 0.8m\n", + " 620 0.1353 23.2% 0.8m\n", + " 630 0.1636 23.2% 0.7m\n", + " 640 0.1731 23.2% 0.7m\n", + " 650 0.1528 23.2% 0.7m\n", + " 660 0.1753 23.2% 0.6m\n", + " 670 0.1798 23.2% 0.6m\n", + " 680 0.1708 23.2% 0.6m\n", + " 690 0.1957 23.2% 0.5m\n", + " 700 0.1326 23.2% 0.5m\n", + " 710 0.1631 23.2% 0.5m\n", + " 720 0.1160 23.2% 0.4m\n", + " 730 0.1866 23.2% 0.4m\n", + " 740 0.1702 23.2% 0.4m\n", + " 750 0.2170 23.2% 0.4m\n", + " 760 0.1860 23.2% 0.3m\n", + " 770 0.1764 23.2% 0.3m\n", + " 780 0.1234 23.2% 0.3m\n", + " 790 0.1529 23.2% 0.2m\n", + " 800 0.1106 23.2% 0.2m\n", + " 810 0.1924 23.2% 0.2m\n", + " 820 0.1505 23.2% 0.1m\n", + " 830 0.1521 23.2% 0.1m\n", + " 840 0.1826 23.2% 0.1m\n", + " 850 0.1971 23.2% 0.0m\n", + " 860 0.1815 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 64 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 64 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16671 โ†’ Val=0.16392\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 102m\n", + "\n", + "๐Ÿ”„ Epoch 65/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1760 23.2% 2.7m\n", + " 10 0.1809 23.2% 2.7m\n", + " 20 0.1602 23.2% 2.6m\n", + " 30 0.1323 23.2% 2.6m\n", + " 40 0.1644 23.2% 2.6m\n", + " 50 0.1950 23.2% 2.5m\n", + " 60 0.1875 23.2% 2.5m\n", + " 70 0.2017 23.2% 2.5m\n", + " 80 0.1611 23.2% 2.4m\n", + " 90 0.1864 23.2% 2.4m\n", + " 100 0.1694 23.2% 2.4m\n", + " 110 0.1565 23.2% 2.4m\n", + " 120 0.1444 23.2% 2.3m\n", + " 130 0.1796 23.2% 2.3m\n", + " 140 0.1800 23.2% 2.3m\n", + " 150 0.1661 23.2% 2.2m\n", + " 160 0.1659 23.2% 2.2m\n", + " 170 0.1963 23.2% 2.2m\n", + " 180 0.1222 23.2% 2.1m\n", + " 190 0.1497 23.2% 2.1m\n", + " 200 0.1364 23.2% 2.1m\n", + " 210 0.1915 23.2% 2.0m\n", + " 220 0.1890 23.2% 2.0m\n", + " 230 0.1740 23.2% 2.0m\n", + " 240 0.1841 23.2% 1.9m\n", + " 250 0.2019 23.2% 1.9m\n", + " 260 0.1251 23.2% 1.9m\n", + " 270 0.1861 23.2% 1.8m\n", + " 280 0.1465 23.2% 1.8m\n", + " 290 0.1408 23.2% 1.8m\n", + " 300 0.1745 23.2% 1.8m\n", + " 310 0.1589 23.2% 1.7m\n", + " 320 0.1879 23.2% 1.7m\n", + " 330 0.1745 23.2% 1.7m\n", + " 340 0.1554 23.2% 1.6m\n", + " 350 0.1604 23.2% 1.6m\n", + " 360 0.1485 23.2% 1.6m\n", + " 370 0.1851 23.2% 1.5m\n", + " 380 0.1650 23.2% 1.5m\n", + " 390 0.1736 23.2% 1.5m\n", + " 400 0.2103 23.2% 1.4m\n", + " 410 0.1731 23.2% 1.4m\n", + " 420 0.1484 23.2% 1.4m\n", + " 430 0.1406 23.2% 1.4m\n", + " 440 0.1525 23.2% 1.3m\n", + " 450 0.1756 23.2% 1.3m\n", + " 460 0.1971 23.2% 1.3m\n", + " 470 0.1666 23.2% 1.2m\n", + " 480 0.1541 23.2% 1.2m\n", + " 490 0.1631 23.2% 1.2m\n", + " 500 0.1943 23.2% 1.1m\n", + " 510 0.1673 23.2% 1.1m\n", + " 520 0.1595 23.2% 1.1m\n", + " 530 0.1650 23.2% 1.0m\n", + " 540 0.1922 23.2% 1.0m\n", + " 550 0.1454 23.2% 1.0m\n", + " 560 0.1940 23.2% 0.9m\n", + " 570 0.1414 23.2% 0.9m\n", + " 580 0.1972 23.2% 0.9m\n", + " 590 0.1349 23.2% 0.9m\n", + " 600 0.1622 23.2% 0.8m\n", + " 610 0.1982 23.2% 0.8m\n", + " 620 0.1445 23.2% 0.8m\n", + " 630 0.1803 23.2% 0.7m\n", + " 640 0.2003 23.2% 0.7m\n", + " 650 0.1265 23.2% 0.7m\n", + " 660 0.1694 23.2% 0.6m\n", + " 670 0.1782 23.2% 0.6m\n", + " 680 0.1953 23.2% 0.6m\n", + " 690 0.1742 23.2% 0.5m\n", + " 700 0.1847 23.2% 0.5m\n", + " 710 0.1795 23.2% 0.5m\n", + " 720 0.1930 23.2% 0.4m\n", + " 730 0.1955 23.2% 0.4m\n", + " 740 0.1904 23.2% 0.4m\n", + " 750 0.1527 23.2% 0.4m\n", + " 760 0.1938 23.2% 0.3m\n", + " 770 0.1711 23.2% 0.3m\n", + " 780 0.1658 23.2% 0.3m\n", + " 790 0.1207 23.2% 0.2m\n", + " 800 0.1667 23.2% 0.2m\n", + " 810 0.1369 23.2% 0.2m\n", + " 820 0.1863 23.2% 0.1m\n", + " 830 0.1404 23.2% 0.1m\n", + " 840 0.1844 23.2% 0.1m\n", + " 850 0.1572 23.2% 0.0m\n", + " 860 0.1914 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 65 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 65 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16659 โ†’ Val=0.16393\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 99m\n", + "\n", + "๐Ÿ”„ Epoch 66/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1368 23.2% 2.7m\n", + " 10 0.1750 23.2% 2.6m\n", + " 20 0.1832 23.2% 2.6m\n", + " 30 0.1355 23.2% 2.6m\n", + " 40 0.1555 23.2% 2.6m\n", + " 50 0.2031 23.2% 2.5m\n", + " 60 0.2294 23.2% 2.5m\n", + " 70 0.1442 23.2% 2.5m\n", + " 80 0.2189 23.2% 2.4m\n", + " 90 0.1749 23.2% 2.4m\n", + " 100 0.1777 23.2% 2.4m\n", + " 110 0.1316 23.2% 2.4m\n", + " 120 0.1652 23.2% 2.3m\n", + " 130 0.1739 23.2% 2.3m\n", + " 140 0.1423 23.2% 2.3m\n", + " 150 0.1802 23.2% 2.2m\n", + " 160 0.1618 23.2% 2.2m\n", + " 170 0.1564 23.2% 2.2m\n", + " 180 0.1307 23.2% 2.1m\n", + " 190 0.1709 23.2% 2.1m\n", + " 200 0.1593 23.2% 2.1m\n", + " 210 0.1434 23.2% 2.1m\n", + " 220 0.1608 23.2% 2.0m\n", + " 230 0.2078 23.2% 2.0m\n", + " 240 0.1963 23.2% 2.0m\n", + " 250 0.1895 23.2% 1.9m\n", + " 260 0.1642 23.2% 1.9m\n", + " 270 0.1460 23.2% 1.9m\n", + " 280 0.1693 23.2% 1.8m\n", + " 290 0.2244 23.2% 1.8m\n", + " 300 0.1368 23.2% 1.8m\n", + " 310 0.1817 23.2% 1.7m\n", + " 320 0.2000 23.2% 1.7m\n", + " 330 0.1644 23.2% 1.7m\n", + " 340 0.1931 23.2% 1.6m\n", + " 350 0.1235 23.2% 1.6m\n", + " 360 0.1888 23.2% 1.6m\n", + " 370 0.0987 23.2% 1.5m\n", + " 380 0.1647 23.2% 1.5m\n", + " 390 0.1720 23.2% 1.5m\n", + " 400 0.1471 23.2% 1.5m\n", + " 410 0.1565 23.2% 1.4m\n", + " 420 0.1996 23.2% 1.4m\n", + " 430 0.1902 23.2% 1.4m\n", + " 440 0.1481 23.2% 1.3m\n", + " 450 0.1720 23.2% 1.3m\n", + " 460 0.1683 23.2% 1.3m\n", + " 470 0.1589 23.2% 1.2m\n", + " 480 0.1584 23.2% 1.2m\n", + " 490 0.1878 23.2% 1.2m\n", + " 500 0.1567 23.2% 1.1m\n", + " 510 0.1760 23.2% 1.1m\n", + " 520 0.1679 23.2% 1.1m\n", + " 530 0.1678 23.2% 1.0m\n", + " 540 0.1480 23.2% 1.0m\n", + " 550 0.1626 23.2% 1.0m\n", + " 560 0.1491 23.2% 1.0m\n", + " 570 0.1769 23.2% 0.9m\n", + " 580 0.1824 23.2% 0.9m\n", + " 590 0.1706 23.2% 0.9m\n", + " 600 0.1531 23.2% 0.8m\n", + " 610 0.1385 23.2% 0.8m\n", + " 620 0.1680 23.2% 0.8m\n", + " 630 0.2185 23.2% 0.7m\n", + " 640 0.1855 23.2% 0.7m\n", + " 650 0.1920 23.2% 0.7m\n", + " 660 0.1739 23.2% 0.6m\n", + " 670 0.1525 23.2% 0.6m\n", + " 680 0.1442 23.2% 0.6m\n", + " 690 0.1953 23.2% 0.5m\n", + " 700 0.2097 23.2% 0.5m\n", + " 710 0.1734 23.2% 0.5m\n", + " 720 0.2091 23.2% 0.4m\n", + " 730 0.1625 23.2% 0.4m\n", + " 740 0.1857 23.2% 0.4m\n", + " 750 0.1612 23.2% 0.4m\n", + " 760 0.1557 23.2% 0.3m\n", + " 770 0.1849 23.2% 0.3m\n", + " 780 0.1595 23.2% 0.3m\n", + " 790 0.1614 23.2% 0.2m\n", + " 800 0.1761 23.2% 0.2m\n", + " 810 0.1455 23.2% 0.2m\n", + " 820 0.1966 23.2% 0.1m\n", + " 830 0.1708 23.2% 0.1m\n", + " 840 0.1596 23.2% 0.1m\n", + " 850 0.1507 23.2% 0.0m\n", + " 860 0.1868 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 66 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 66 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16658 โ†’ Val=0.16399\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โš ๏ธ No improvement for 5/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 97m\n", + "\n", + "๐Ÿ”„ Epoch 67/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1238 23.2% 2.6m\n", + " 10 0.1678 23.2% 2.6m\n", + " 20 0.1443 23.2% 2.6m\n", + " 30 0.1524 23.2% 2.6m\n", + " 40 0.1927 23.2% 2.6m\n", + " 50 0.1444 23.2% 2.6m\n", + " 60 0.2063 23.2% 2.5m\n", + " 70 0.1850 23.2% 2.5m\n", + " 80 0.1517 23.2% 2.5m\n", + " 90 0.1609 23.2% 2.4m\n", + " 100 0.1594 23.2% 2.4m\n", + " 110 0.1666 23.2% 2.4m\n", + " 120 0.1657 23.2% 2.3m\n", + " 130 0.1294 23.2% 2.3m\n", + " 140 0.1591 23.2% 2.3m\n", + " 150 0.1890 23.2% 2.2m\n", + " 160 0.1669 23.2% 2.2m\n", + " 170 0.1300 23.2% 2.2m\n", + " 180 0.1708 23.2% 2.1m\n", + " 190 0.2005 23.2% 2.1m\n", + " 200 0.1136 23.2% 2.1m\n", + " 210 0.1631 23.2% 2.0m\n", + " 220 0.1848 23.2% 2.0m\n", + " 230 0.1499 23.2% 2.0m\n", + " 240 0.1667 23.2% 2.0m\n", + " 250 0.1506 23.2% 1.9m\n", + " 260 0.2018 23.2% 1.9m\n", + " 270 0.1963 23.2% 1.9m\n", + " 280 0.1511 23.2% 1.8m\n", + " 290 0.1904 23.2% 1.8m\n", + " 300 0.1321 23.2% 1.8m\n", + " 310 0.1351 23.2% 1.7m\n", + " 320 0.1914 23.2% 1.7m\n", + " 330 0.1439 23.2% 1.7m\n", + " 340 0.1602 23.2% 1.6m\n", + " 350 0.1500 23.2% 1.6m\n", + " 360 0.1683 23.2% 1.6m\n", + " 370 0.1153 23.2% 1.5m\n", + " 380 0.1782 23.2% 1.5m\n", + " 390 0.2033 23.2% 1.5m\n", + " 400 0.2078 23.2% 1.4m\n", + " 410 0.1588 23.2% 1.4m\n", + " 420 0.1411 23.2% 1.4m\n", + " 430 0.2008 23.2% 1.4m\n", + " 440 0.1800 23.2% 1.3m\n", + " 450 0.1968 23.2% 1.3m\n", + " 460 0.1716 23.2% 1.3m\n", + " 470 0.1466 23.2% 1.2m\n", + " 480 0.2054 23.2% 1.2m\n", + " 490 0.1805 23.2% 1.2m\n", + " 500 0.1465 23.2% 1.1m\n", + " 510 0.1619 23.2% 1.1m\n", + " 520 0.1344 23.2% 1.1m\n", + " 530 0.1629 23.2% 1.0m\n", + " 540 0.1608 23.2% 1.0m\n", + " 550 0.1558 23.2% 1.0m\n", + " 560 0.1827 23.2% 0.9m\n", + " 570 0.1411 23.2% 0.9m\n", + " 580 0.1693 23.2% 0.9m\n", + " 590 0.1866 23.2% 0.9m\n", + " 600 0.1952 23.2% 0.8m\n", + " 610 0.1910 23.2% 0.8m\n", + " 620 0.1470 23.2% 0.8m\n", + " 630 0.1344 23.2% 0.7m\n", + " 640 0.1626 23.2% 0.7m\n", + " 650 0.1942 23.2% 0.7m\n", + " 660 0.1196 23.2% 0.6m\n", + " 670 0.2022 23.2% 0.6m\n", + " 680 0.1745 23.2% 0.6m\n", + " 690 0.1587 23.2% 0.5m\n", + " 700 0.1597 23.2% 0.5m\n", + " 710 0.1773 23.2% 0.5m\n", + " 720 0.1435 23.2% 0.4m\n", + " 730 0.1544 23.2% 0.4m\n", + " 740 0.1920 23.2% 0.4m\n", + " 750 0.1435 23.2% 0.4m\n", + " 760 0.1786 23.2% 0.3m\n", + " 770 0.1807 23.2% 0.3m\n", + " 780 0.2178 23.2% 0.3m\n", + " 790 0.1759 23.2% 0.2m\n", + " 800 0.1708 23.2% 0.2m\n", + " 810 0.1245 23.2% 0.2m\n", + " 820 0.1499 23.2% 0.1m\n", + " 830 0.1457 23.2% 0.1m\n", + " 840 0.1482 23.2% 0.1m\n", + " 850 0.1627 23.2% 0.0m\n", + " 860 0.1474 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 67 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 67 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16659 โ†’ Val=0.16395\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.8% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 94m\n", + "\n", + "๐Ÿ”„ Epoch 68/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1739 23.2% 2.8m\n", + " 10 0.1590 23.2% 2.7m\n", + " 20 0.1825 23.2% 2.6m\n", + " 30 0.1629 23.2% 2.6m\n", + " 40 0.1621 23.2% 2.6m\n", + " 50 0.1796 23.2% 2.5m\n", + " 60 0.1753 23.2% 2.5m\n", + " 70 0.1834 23.2% 2.5m\n", + " 80 0.1643 23.2% 2.5m\n", + " 90 0.1957 23.2% 2.4m\n", + " 100 0.1989 23.2% 2.4m\n", + " 110 0.1706 23.2% 2.4m\n", + " 120 0.1634 23.2% 2.3m\n", + " 130 0.1566 23.2% 2.3m\n", + " 140 0.2042 23.2% 2.3m\n", + " 150 0.1727 23.2% 2.2m\n", + " 160 0.1677 23.2% 2.2m\n", + " 170 0.1652 23.2% 2.2m\n", + " 180 0.1414 23.2% 2.1m\n", + " 190 0.1766 23.2% 2.1m\n", + " 200 0.1330 23.2% 2.1m\n", + " 210 0.1687 23.2% 2.1m\n", + " 220 0.1562 23.2% 2.0m\n", + " 230 0.1599 23.2% 2.0m\n", + " 240 0.1644 23.2% 2.0m\n", + " 250 0.1759 23.2% 1.9m\n", + " 260 0.1961 23.2% 1.9m\n", + " 270 0.1575 23.2% 1.9m\n", + " 280 0.1622 23.2% 1.8m\n", + " 290 0.1607 23.2% 1.8m\n", + " 300 0.1846 23.2% 1.8m\n", + " 310 0.1674 23.2% 1.7m\n", + " 320 0.1544 23.2% 1.7m\n", + " 330 0.1647 23.2% 1.7m\n", + " 340 0.1548 23.2% 1.6m\n", + " 350 0.1900 23.2% 1.6m\n", + " 360 0.1782 23.2% 1.6m\n", + " 370 0.1633 23.2% 1.6m\n", + " 380 0.1545 23.2% 1.5m\n", + " 390 0.1581 23.2% 1.5m\n", + " 400 0.1513 23.2% 1.5m\n", + " 410 0.1541 23.2% 1.4m\n", + " 420 0.1562 23.2% 1.4m\n", + " 430 0.1706 23.2% 1.4m\n", + " 440 0.1530 23.2% 1.3m\n", + " 450 0.1786 23.2% 1.3m\n", + " 460 0.1833 23.2% 1.3m\n", + " 470 0.1557 23.2% 1.2m\n", + " 480 0.2081 23.2% 1.2m\n", + " 490 0.1649 23.2% 1.2m\n", + " 500 0.2003 23.2% 1.1m\n", + " 510 0.1527 23.2% 1.1m\n", + " 520 0.1598 23.2% 1.1m\n", + " 530 0.1505 23.2% 1.1m\n", + " 540 0.1625 23.2% 1.0m\n", + " 550 0.1413 23.2% 1.0m\n", + " 560 0.1439 23.2% 1.0m\n", + " 570 0.1816 23.2% 0.9m\n", + " 580 0.1683 23.2% 0.9m\n", + " 590 0.1608 23.2% 0.9m\n", + " 600 0.2359 23.2% 0.8m\n", + " 610 0.1737 23.2% 0.8m\n", + " 620 0.1609 23.2% 0.8m\n", + " 630 0.1479 23.2% 0.7m\n", + " 640 0.1877 23.2% 0.7m\n", + " 650 0.1641 23.2% 0.7m\n", + " 660 0.1718 23.2% 0.6m\n", + " 670 0.1678 23.2% 0.6m\n", + " 680 0.1739 23.2% 0.6m\n", + " 690 0.1750 23.2% 0.5m\n", + " 700 0.2218 23.2% 0.5m\n", + " 710 0.1402 23.2% 0.5m\n", + " 720 0.1670 23.2% 0.5m\n", + " 730 0.1894 23.2% 0.4m\n", + " 740 0.1716 23.2% 0.4m\n", + " 750 0.1340 23.2% 0.4m\n", + " 760 0.1680 23.2% 0.3m\n", + " 770 0.1287 23.2% 0.3m\n", + " 780 0.1926 23.2% 0.3m\n", + " 790 0.1758 23.2% 0.2m\n", + " 800 0.1423 23.2% 0.2m\n", + " 810 0.1608 23.2% 0.2m\n", + " 820 0.1709 23.2% 0.1m\n", + " 830 0.1245 23.2% 0.1m\n", + " 840 0.1850 23.2% 0.1m\n", + " 850 0.1638 23.2% 0.0m\n", + " 860 0.1992 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 68 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 68 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16660 โ†’ Val=0.16411\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 91m\n", + "\n", + "๐Ÿ”„ Epoch 69/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1917 23.2% 2.6m\n", + " 10 0.1844 23.2% 2.7m\n", + " 20 0.1528 23.2% 2.7m\n", + " 30 0.1822 23.2% 2.6m\n", + " 40 0.1625 23.2% 2.6m\n", + " 50 0.1505 23.2% 2.6m\n", + " 60 0.1666 23.2% 2.5m\n", + " 70 0.2017 23.2% 2.5m\n", + " 80 0.1858 23.2% 2.5m\n", + " 90 0.1384 23.2% 2.4m\n", + " 100 0.1755 23.2% 2.4m\n", + " 110 0.1880 23.2% 2.4m\n", + " 120 0.2013 23.2% 2.3m\n", + " 130 0.1884 23.2% 2.3m\n", + " 140 0.1698 23.2% 2.3m\n", + " 150 0.1798 23.2% 2.2m\n", + " 160 0.1451 23.2% 2.2m\n", + " 170 0.1435 23.2% 2.2m\n", + " 180 0.1752 23.2% 2.1m\n", + " 190 0.1997 23.2% 2.1m\n", + " 200 0.1442 23.2% 2.1m\n", + " 210 0.1278 23.2% 2.1m\n", + " 220 0.2023 23.2% 2.0m\n", + " 230 0.1309 23.2% 2.0m\n", + " 240 0.1553 23.2% 2.0m\n", + " 250 0.1656 23.2% 1.9m\n", + " 260 0.1381 23.2% 1.9m\n", + " 270 0.1286 23.2% 1.9m\n", + " 280 0.1696 23.2% 1.8m\n", + " 290 0.1670 23.2% 1.8m\n", + " 300 0.1629 23.2% 1.8m\n", + " 310 0.1580 23.2% 1.7m\n", + " 320 0.1678 23.2% 1.7m\n", + " 330 0.1275 23.2% 1.7m\n", + " 340 0.1356 23.2% 1.7m\n", + " 350 0.1779 23.2% 1.6m\n", + " 360 0.1597 23.2% 1.6m\n", + " 370 0.1637 23.2% 1.6m\n", + " 380 0.1932 23.2% 1.5m\n", + " 390 0.1596 23.2% 1.5m\n", + " 400 0.1625 23.2% 1.5m\n", + " 410 0.1507 23.2% 1.4m\n", + " 420 0.2231 23.2% 1.4m\n", + " 430 0.2116 23.2% 1.4m\n", + " 440 0.1815 23.2% 1.3m\n", + " 450 0.1631 23.2% 1.3m\n", + " 460 0.1293 23.2% 1.3m\n", + " 470 0.1737 23.2% 1.2m\n", + " 480 0.1807 23.2% 1.2m\n", + " 490 0.1463 23.2% 1.2m\n", + " 500 0.1538 23.2% 1.1m\n", + " 510 0.1412 23.2% 1.1m\n", + " 520 0.1785 23.2% 1.1m\n", + " 530 0.1822 23.2% 1.0m\n", + " 540 0.1778 23.2% 1.0m\n", + " 550 0.1588 23.2% 1.0m\n", + " 560 0.1402 23.2% 1.0m\n", + " 570 0.1598 23.2% 0.9m\n", + " 580 0.1919 23.2% 0.9m\n", + " 590 0.1759 23.2% 0.9m\n", + " 600 0.1662 23.2% 0.8m\n", + " 610 0.1757 23.2% 0.8m\n", + " 620 0.1586 23.2% 0.8m\n", + " 630 0.1572 23.2% 0.7m\n", + " 640 0.1682 23.2% 0.7m\n", + " 650 0.1601 23.2% 0.7m\n", + " 660 0.1304 23.2% 0.6m\n", + " 670 0.1806 23.2% 0.6m\n", + " 680 0.1610 23.2% 0.6m\n", + " 690 0.1455 23.2% 0.5m\n", + " 700 0.1971 23.2% 0.5m\n", + " 710 0.1647 23.2% 0.5m\n", + " 720 0.1346 23.2% 0.5m\n", + " 730 0.1708 23.2% 0.4m\n", + " 740 0.1622 23.2% 0.4m\n", + " 750 0.1894 23.2% 0.4m\n", + " 760 0.2017 23.2% 0.3m\n", + " 770 0.1629 23.2% 0.3m\n", + " 780 0.1676 23.2% 0.3m\n", + " 790 0.1697 23.2% 0.2m\n", + " 800 0.1447 23.2% 0.2m\n", + " 810 0.1626 23.2% 0.2m\n", + " 820 0.1593 23.2% 0.1m\n", + " 830 0.1707 23.2% 0.1m\n", + " 840 0.1193 23.2% 0.1m\n", + " 850 0.1420 23.2% 0.0m\n", + " 860 0.1883 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 69 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 69 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16666 โ†’ Val=0.16400\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.8% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 88m\n", + "\n", + "๐Ÿ”„ Epoch 70/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1772 23.2% 2.7m\n", + " 10 0.1727 23.2% 2.7m\n", + " 20 0.2127 23.2% 2.6m\n", + " 30 0.1893 23.2% 2.6m\n", + " 40 0.1864 23.2% 2.6m\n", + " 50 0.1829 23.2% 2.6m\n", + " 60 0.1516 23.2% 2.5m\n", + " 70 0.2114 23.2% 2.5m\n", + " 80 0.1676 23.2% 2.5m\n", + " 90 0.1705 23.2% 2.4m\n", + " 100 0.1600 23.2% 2.4m\n", + " 110 0.1754 23.2% 2.4m\n", + " 120 0.1496 23.2% 2.3m\n", + " 130 0.1826 23.2% 2.3m\n", + " 140 0.1750 23.2% 2.3m\n", + " 150 0.1690 23.2% 2.2m\n", + " 160 0.1395 23.2% 2.2m\n", + " 170 0.1547 23.2% 2.2m\n", + " 180 0.1554 23.2% 2.1m\n", + " 190 0.1321 23.2% 2.1m\n", + " 200 0.1394 23.2% 2.1m\n", + " 210 0.1790 23.2% 2.0m\n", + " 220 0.1312 23.2% 2.0m\n", + " 230 0.2099 23.2% 2.0m\n", + " 240 0.1468 23.2% 2.0m\n", + " 250 0.1931 23.2% 1.9m\n", + " 260 0.1533 23.2% 1.9m\n", + " 270 0.1478 23.2% 1.9m\n", + " 280 0.1645 23.2% 1.8m\n", + " 290 0.1651 23.2% 1.8m\n", + " 300 0.1175 23.2% 1.8m\n", + " 310 0.1908 23.2% 1.7m\n", + " 320 0.1796 23.2% 1.7m\n", + " 330 0.1892 23.2% 1.7m\n", + " 340 0.1592 23.2% 1.6m\n", + " 350 0.1908 23.2% 1.6m\n", + " 360 0.1507 23.2% 1.6m\n", + " 370 0.1371 23.2% 1.5m\n", + " 380 0.1464 23.2% 1.5m\n", + " 390 0.1449 23.2% 1.5m\n", + " 400 0.1826 23.2% 1.4m\n", + " 410 0.1636 23.2% 1.4m\n", + " 420 0.1505 23.2% 1.4m\n", + " 430 0.1808 23.2% 1.4m\n", + " 440 0.1465 23.2% 1.3m\n", + " 450 0.1740 23.2% 1.3m\n", + " 460 0.2131 23.2% 1.3m\n", + " 470 0.1766 23.2% 1.2m\n", + " 480 0.1891 23.2% 1.2m\n", + " 490 0.1359 23.2% 1.2m\n", + " 500 0.1424 23.2% 1.1m\n", + " 510 0.1754 23.2% 1.1m\n", + " 520 0.1692 23.2% 1.1m\n", + " 530 0.1675 23.2% 1.0m\n", + " 540 0.1909 23.2% 1.0m\n", + " 550 0.1602 23.2% 1.0m\n", + " 560 0.2042 23.2% 0.9m\n", + " 570 0.1363 23.2% 0.9m\n", + " 580 0.1750 23.2% 0.9m\n", + " 590 0.1717 23.2% 0.9m\n", + " 600 0.1704 23.2% 0.8m\n", + " 610 0.1351 23.2% 0.8m\n", + " 620 0.1956 23.2% 0.8m\n", + " 630 0.1864 23.2% 0.7m\n", + " 640 0.1582 23.2% 0.7m\n", + " 650 0.1676 23.2% 0.7m\n", + " 660 0.1489 23.2% 0.6m\n", + " 670 0.1487 23.2% 0.6m\n", + " 680 0.1525 23.2% 0.6m\n", + " 690 0.1558 23.2% 0.5m\n", + " 700 0.1555 23.2% 0.5m\n", + " 710 0.1942 23.2% 0.5m\n", + " 720 0.1423 23.2% 0.4m\n", + " 730 0.1758 23.2% 0.4m\n", + " 740 0.1605 23.2% 0.4m\n", + " 750 0.1931 23.2% 0.4m\n", + " 760 0.1722 23.2% 0.3m\n", + " 770 0.1650 23.2% 0.3m\n", + " 780 0.1626 23.2% 0.3m\n", + " 790 0.2032 23.2% 0.2m\n", + " 800 0.1462 23.2% 0.2m\n", + " 810 0.1535 23.2% 0.2m\n", + " 820 0.1722 23.2% 0.1m\n", + " 830 0.1745 23.2% 0.1m\n", + " 840 0.1530 23.2% 0.1m\n", + " 850 0.1818 23.2% 0.0m\n", + " 860 0.1733 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 70 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 70 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16652 โ†’ Val=0.16387\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 85m\n", + "\n", + "๐Ÿ”„ Epoch 71/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1770 23.2% 2.6m\n", + " 10 0.1651 23.2% 2.6m\n", + " 20 0.1470 23.2% 2.7m\n", + " 30 0.1716 23.2% 2.6m\n", + " 40 0.1622 23.2% 2.6m\n", + " 50 0.1921 23.2% 2.6m\n", + " 60 0.1488 23.2% 2.5m\n", + " 70 0.1850 23.2% 2.5m\n", + " 80 0.1413 23.2% 2.5m\n", + " 90 0.1767 23.2% 2.4m\n", + " 100 0.1410 23.2% 2.4m\n", + " 110 0.1748 23.2% 2.4m\n", + " 120 0.1473 23.2% 2.3m\n", + " 130 0.1441 23.2% 2.3m\n", + " 140 0.1386 23.2% 2.3m\n", + " 150 0.1584 23.2% 2.2m\n", + " 160 0.1289 23.2% 2.2m\n", + " 170 0.1383 23.2% 2.2m\n", + " 180 0.1929 23.2% 2.1m\n", + " 190 0.1520 23.2% 2.1m\n", + " 200 0.1417 23.2% 2.1m\n", + " 210 0.1717 23.2% 2.0m\n", + " 220 0.1452 23.2% 2.0m\n", + " 230 0.1795 23.2% 2.0m\n", + " 240 0.1993 23.2% 2.0m\n", + " 250 0.1933 23.2% 1.9m\n", + " 260 0.2023 23.2% 1.9m\n", + " 270 0.1696 23.2% 1.9m\n", + " 280 0.1623 23.2% 1.8m\n", + " 290 0.1796 23.2% 1.8m\n", + " 300 0.1970 23.2% 1.8m\n", + " 310 0.1595 23.2% 1.7m\n", + " 320 0.1862 23.2% 1.7m\n", + " 330 0.1697 23.2% 1.7m\n", + " 340 0.2134 23.2% 1.7m\n", + " 350 0.1801 23.2% 1.6m\n", + " 360 0.1746 23.2% 1.6m\n", + " 370 0.2078 23.2% 1.6m\n", + " 380 0.1512 23.2% 1.5m\n", + " 390 0.1730 23.2% 1.5m\n", + " 400 0.1400 23.2% 1.5m\n", + " 410 0.1850 23.2% 1.4m\n", + " 420 0.1903 23.2% 1.4m\n", + " 430 0.1496 23.2% 1.4m\n", + " 440 0.1723 23.2% 1.3m\n", + " 450 0.2130 23.2% 1.3m\n", + " 460 0.1500 23.2% 1.3m\n", + " 470 0.1317 23.2% 1.2m\n", + " 480 0.1608 23.2% 1.2m\n", + " 490 0.1797 23.2% 1.2m\n", + " 500 0.2055 23.2% 1.1m\n", + " 510 0.1529 23.2% 1.1m\n", + " 520 0.1802 23.2% 1.1m\n", + " 530 0.1704 23.2% 1.1m\n", + " 540 0.1735 23.2% 1.0m\n", + " 550 0.1674 23.2% 1.0m\n", + " 560 0.1982 23.2% 1.0m\n", + " 570 0.1699 23.2% 0.9m\n", + " 580 0.1377 23.2% 0.9m\n", + " 590 0.1854 23.2% 0.9m\n", + " 600 0.1437 23.2% 0.8m\n", + " 610 0.2036 23.2% 0.8m\n", + " 620 0.1487 23.2% 0.8m\n", + " 630 0.1651 23.2% 0.7m\n", + " 640 0.1419 23.2% 0.7m\n", + " 650 0.1774 23.2% 0.7m\n", + " 660 0.1664 23.2% 0.6m\n", + " 670 0.1366 23.2% 0.6m\n", + " 680 0.1462 23.2% 0.6m\n", + " 690 0.1484 23.2% 0.5m\n", + " 700 0.1905 23.2% 0.5m\n", + " 710 0.1560 23.2% 0.5m\n", + " 720 0.1417 23.2% 0.5m\n", + " 730 0.1550 23.2% 0.4m\n", + " 740 0.1843 23.2% 0.4m\n", + " 750 0.1680 23.2% 0.4m\n", + " 760 0.1700 23.2% 0.3m\n", + " 770 0.1999 23.2% 0.3m\n", + " 780 0.1984 23.2% 0.3m\n", + " 790 0.1558 23.2% 0.2m\n", + " 800 0.1708 23.2% 0.2m\n", + " 810 0.1624 23.2% 0.2m\n", + " 820 0.1615 23.2% 0.1m\n", + " 830 0.1293 23.2% 0.1m\n", + " 840 0.1399 23.2% 0.1m\n", + " 850 0.1609 23.2% 0.0m\n", + " 860 0.1773 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 71 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 71 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16663 โ†’ Val=0.16383\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โš ๏ธ No improvement for 10/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 83m\n", + "\n", + "๐Ÿ”„ Epoch 72/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1460 23.2% 2.7m\n", + " 10 0.1374 23.2% 2.7m\n", + " 20 0.1602 23.2% 2.7m\n", + " 30 0.1898 23.2% 2.6m\n", + " 40 0.2084 23.2% 2.6m\n", + " 50 0.1346 23.2% 2.6m\n", + " 60 0.1593 23.2% 2.5m\n", + " 70 0.1589 23.2% 2.5m\n", + " 80 0.1585 23.2% 2.5m\n", + " 90 0.2105 23.2% 2.4m\n", + " 100 0.1581 23.2% 2.4m\n", + " 110 0.1958 23.2% 2.4m\n", + " 120 0.1698 23.2% 2.3m\n", + " 130 0.1708 23.2% 2.3m\n", + " 140 0.1737 23.2% 2.3m\n", + " 150 0.1876 23.2% 2.2m\n", + " 160 0.1612 23.2% 2.2m\n", + " 170 0.1680 23.2% 2.2m\n", + " 180 0.1490 23.2% 2.2m\n", + " 190 0.1500 23.2% 2.1m\n", + " 200 0.1962 23.2% 2.1m\n", + " 210 0.1407 23.2% 2.1m\n", + " 220 0.2218 23.2% 2.0m\n", + " 230 0.1766 23.2% 2.0m\n", + " 240 0.1820 23.2% 2.0m\n", + " 250 0.1661 23.2% 1.9m\n", + " 260 0.1481 23.2% 1.9m\n", + " 270 0.1739 23.2% 1.9m\n", + " 280 0.1648 23.2% 1.8m\n", + " 290 0.1528 23.2% 1.8m\n", + " 300 0.1370 23.2% 1.8m\n", + " 310 0.1865 23.2% 1.7m\n", + " 320 0.1654 23.2% 1.7m\n", + " 330 0.1512 23.2% 1.7m\n", + " 340 0.1600 23.2% 1.7m\n", + " 350 0.1776 23.2% 1.6m\n", + " 360 0.1983 23.2% 1.6m\n", + " 370 0.1520 23.2% 1.6m\n", + " 380 0.1932 23.2% 1.5m\n", + " 390 0.1351 23.2% 1.5m\n", + " 400 0.1469 23.2% 1.5m\n", + " 410 0.1897 23.2% 1.4m\n", + " 420 0.1552 23.2% 1.4m\n", + " 430 0.1816 23.2% 1.4m\n", + " 440 0.1582 23.2% 1.3m\n", + " 450 0.1392 23.2% 1.3m\n", + " 460 0.1631 23.2% 1.3m\n", + " 470 0.2059 23.2% 1.2m\n", + " 480 0.1914 23.2% 1.2m\n", + " 490 0.1350 23.2% 1.2m\n", + " 500 0.1901 23.2% 1.1m\n", + " 510 0.1862 23.2% 1.1m\n", + " 520 0.1740 23.2% 1.1m\n", + " 530 0.1892 23.2% 1.1m\n", + " 540 0.1656 23.2% 1.0m\n", + " 550 0.1644 23.2% 1.0m\n", + " 560 0.1158 23.2% 1.0m\n", + " 570 0.1543 23.2% 0.9m\n", + " 580 0.1824 23.2% 0.9m\n", + " 590 0.1692 23.2% 0.9m\n", + " 600 0.1668 23.2% 0.8m\n", + " 610 0.1778 23.2% 0.8m\n", + " 620 0.2197 23.2% 0.8m\n", + " 630 0.1731 23.2% 0.7m\n", + " 640 0.1330 23.2% 0.7m\n", + " 650 0.1994 23.2% 0.7m\n", + " 660 0.1980 23.2% 0.6m\n", + " 670 0.1741 23.2% 0.6m\n", + " 680 0.1669 23.2% 0.6m\n", + " 690 0.1704 23.2% 0.5m\n", + " 700 0.1280 23.2% 0.5m\n", + " 710 0.1460 23.2% 0.5m\n", + " 720 0.1382 23.2% 0.5m\n", + " 730 0.1461 23.2% 0.4m\n", + " 740 0.1564 23.2% 0.4m\n", + " 750 0.1463 23.2% 0.4m\n", + " 760 0.1633 23.2% 0.3m\n", + " 770 0.2151 23.2% 0.3m\n", + " 780 0.1613 23.2% 0.3m\n", + " 790 0.1315 23.2% 0.2m\n", + " 800 0.1729 23.2% 0.2m\n", + " 810 0.1232 23.2% 0.2m\n", + " 820 0.1823 23.2% 0.1m\n", + " 830 0.1615 23.2% 0.1m\n", + " 840 0.1443 23.2% 0.1m\n", + " 850 0.1931 23.2% 0.0m\n", + " 860 0.1549 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 72 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 72 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16644 โ†’ Val=0.16399\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 80m\n", + "\n", + "๐Ÿ”„ Epoch 73/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1530 23.2% 2.6m\n", + " 10 0.1559 23.2% 2.8m\n", + " 20 0.1547 23.2% 2.7m\n", + " 30 0.1237 23.2% 2.7m\n", + " 40 0.1712 23.2% 2.6m\n", + " 50 0.1759 23.2% 2.6m\n", + " 60 0.1792 23.2% 2.6m\n", + " 70 0.1742 23.2% 2.5m\n", + " 80 0.1645 23.2% 2.5m\n", + " 90 0.1537 23.2% 2.5m\n", + " 100 0.1711 23.2% 2.4m\n", + " 110 0.1771 23.2% 2.4m\n", + " 120 0.1572 23.2% 2.3m\n", + " 130 0.1916 23.2% 2.3m\n", + " 140 0.2037 23.2% 2.3m\n", + " 150 0.1558 23.2% 2.3m\n", + " 160 0.1598 23.2% 2.2m\n", + " 170 0.1565 23.2% 2.2m\n", + " 180 0.1490 23.2% 2.2m\n", + " 190 0.1685 23.2% 2.1m\n", + " 200 0.1625 23.2% 2.1m\n", + " 210 0.2056 23.2% 2.1m\n", + " 220 0.1657 23.2% 2.0m\n", + " 230 0.1371 23.2% 2.0m\n", + " 240 0.1641 23.2% 2.0m\n", + " 250 0.1665 23.2% 1.9m\n", + " 260 0.1719 23.2% 1.9m\n", + " 270 0.2004 23.2% 1.9m\n", + " 280 0.1347 23.2% 1.8m\n", + " 290 0.1733 23.2% 1.8m\n", + " 300 0.1918 23.2% 1.8m\n", + " 310 0.1789 23.2% 1.8m\n", + " 320 0.1862 23.2% 1.7m\n", + " 330 0.1564 23.2% 1.7m\n", + " 340 0.1835 23.2% 1.7m\n", + " 350 0.1505 23.2% 1.6m\n", + " 360 0.1702 23.2% 1.6m\n", + " 370 0.1786 23.2% 1.6m\n", + " 380 0.1897 23.2% 1.5m\n", + " 390 0.1708 23.2% 1.5m\n", + " 400 0.1784 23.2% 1.5m\n", + " 410 0.1777 23.2% 1.4m\n", + " 420 0.1348 23.2% 1.4m\n", + " 430 0.1413 23.2% 1.4m\n", + " 440 0.2023 23.2% 1.3m\n", + " 450 0.1982 23.2% 1.3m\n", + " 460 0.1876 23.2% 1.3m\n", + " 470 0.1432 23.2% 1.2m\n", + " 480 0.1781 23.2% 1.2m\n", + " 490 0.1242 23.2% 1.2m\n", + " 500 0.1469 23.2% 1.1m\n", + " 510 0.1604 23.2% 1.1m\n", + " 520 0.1704 23.2% 1.1m\n", + " 530 0.1165 23.2% 1.1m\n", + " 540 0.1727 23.2% 1.0m\n", + " 550 0.1307 23.2% 1.0m\n", + " 560 0.1442 23.2% 1.0m\n", + " 570 0.1827 23.2% 0.9m\n", + " 580 0.1624 23.2% 0.9m\n", + " 590 0.1583 23.2% 0.9m\n", + " 600 0.1422 23.2% 0.8m\n", + " 610 0.1233 23.2% 0.8m\n", + " 620 0.1532 23.2% 0.8m\n", + " 630 0.1638 23.2% 0.7m\n", + " 640 0.1763 23.2% 0.7m\n", + " 650 0.1746 23.2% 0.7m\n", + " 660 0.1553 23.2% 0.6m\n", + " 670 0.1659 23.2% 0.6m\n", + " 680 0.1887 23.2% 0.6m\n", + " 690 0.2282 23.2% 0.5m\n", + " 700 0.1568 23.2% 0.5m\n", + " 710 0.1874 23.2% 0.5m\n", + " 720 0.1491 23.2% 0.5m\n", + " 730 0.1502 23.2% 0.4m\n", + " 740 0.1637 23.2% 0.4m\n", + " 750 0.1711 23.2% 0.4m\n", + " 760 0.2078 23.2% 0.3m\n", + " 770 0.1662 23.2% 0.3m\n", + " 780 0.1394 23.2% 0.3m\n", + " 790 0.1751 23.2% 0.2m\n", + " 800 0.1679 23.2% 0.2m\n", + " 810 0.1904 23.2% 0.2m\n", + " 820 0.1514 23.2% 0.1m\n", + " 830 0.1584 23.2% 0.1m\n", + " 840 0.1403 23.2% 0.1m\n", + " 850 0.1833 23.2% 0.0m\n", + " 860 0.1472 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 73 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 73 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16660 โ†’ Val=0.16385\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 77m\n", + "\n", + "๐Ÿ”„ Epoch 74/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1854 23.2% 2.7m\n", + " 10 0.1937 23.2% 2.7m\n", + " 20 0.1798 23.2% 2.6m\n", + " 30 0.1679 23.2% 2.6m\n", + " 40 0.1565 23.2% 2.6m\n", + " 50 0.1292 23.2% 2.5m\n", + " 60 0.1664 23.2% 2.5m\n", + " 70 0.1861 23.2% 2.5m\n", + " 80 0.1593 23.2% 2.5m\n", + " 90 0.1319 23.2% 2.4m\n", + " 100 0.1464 23.2% 2.4m\n", + " 110 0.1870 23.2% 2.4m\n", + " 120 0.1893 23.2% 2.3m\n", + " 130 0.1982 23.2% 2.3m\n", + " 140 0.1733 23.2% 2.3m\n", + " 150 0.1295 23.2% 2.3m\n", + " 160 0.1558 23.2% 2.2m\n", + " 170 0.1303 23.2% 2.2m\n", + " 180 0.1430 23.2% 2.2m\n", + " 190 0.2001 23.2% 2.1m\n", + " 200 0.1753 23.2% 2.1m\n", + " 210 0.1839 23.2% 2.1m\n", + " 220 0.1769 23.2% 2.0m\n", + " 230 0.1567 23.2% 2.0m\n", + " 240 0.1864 23.2% 2.0m\n", + " 250 0.1599 23.2% 1.9m\n", + " 260 0.1187 23.2% 1.9m\n", + " 270 0.1408 23.2% 1.9m\n", + " 280 0.1633 23.2% 1.8m\n", + " 290 0.2154 23.2% 1.8m\n", + " 300 0.1733 23.2% 1.8m\n", + " 310 0.1547 23.2% 1.8m\n", + " 320 0.1498 23.2% 1.7m\n", + " 330 0.1402 23.2% 1.7m\n", + " 340 0.2145 23.2% 1.7m\n", + " 350 0.1639 23.2% 1.6m\n", + " 360 0.1718 23.2% 1.6m\n", + " 370 0.2012 23.2% 1.6m\n", + " 380 0.1581 23.2% 1.5m\n", + " 390 0.1540 23.2% 1.5m\n", + " 400 0.1331 23.2% 1.5m\n", + " 410 0.1885 23.2% 1.4m\n", + " 420 0.1698 23.2% 1.4m\n", + " 430 0.1838 23.2% 1.4m\n", + " 440 0.1537 23.2% 1.3m\n", + " 450 0.1739 23.2% 1.3m\n", + " 460 0.1859 23.2% 1.3m\n", + " 470 0.1400 23.2% 1.2m\n", + " 480 0.1172 23.2% 1.2m\n", + " 490 0.1438 23.2% 1.2m\n", + " 500 0.1836 23.2% 1.1m\n", + " 510 0.1814 23.2% 1.1m\n", + " 520 0.1659 23.2% 1.1m\n", + " 530 0.1464 23.2% 1.1m\n", + " 540 0.1705 23.2% 1.0m\n", + " 550 0.1791 23.2% 1.0m\n", + " 560 0.1514 23.2% 1.0m\n", + " 570 0.1653 23.2% 0.9m\n", + " 580 0.1639 23.2% 0.9m\n", + " 590 0.1686 23.2% 0.9m\n", + " 600 0.1560 23.2% 0.8m\n", + " 610 0.1677 23.2% 0.8m\n", + " 620 0.2097 23.2% 0.8m\n", + " 630 0.1249 23.2% 0.7m\n", + " 640 0.1435 23.2% 0.7m\n", + " 650 0.1592 23.2% 0.7m\n", + " 660 0.1681 23.2% 0.6m\n", + " 670 0.1855 23.2% 0.6m\n", + " 680 0.1950 23.2% 0.6m\n", + " 690 0.2005 23.2% 0.5m\n", + " 700 0.1628 23.2% 0.5m\n", + " 710 0.1648 23.2% 0.5m\n", + " 720 0.1939 23.2% 0.5m\n", + " 730 0.1538 23.2% 0.4m\n", + " 740 0.1840 23.2% 0.4m\n", + " 750 0.1282 23.2% 0.4m\n", + " 760 0.1757 23.2% 0.3m\n", + " 770 0.1278 23.2% 0.3m\n", + " 780 0.1672 23.2% 0.3m\n", + " 790 0.1691 23.2% 0.2m\n", + " 800 0.1617 23.2% 0.2m\n", + " 810 0.1839 23.2% 0.2m\n", + " 820 0.1521 23.2% 0.1m\n", + " 830 0.1359 23.2% 0.1m\n", + " 840 0.1768 23.2% 0.1m\n", + " 850 0.1586 23.2% 0.0m\n", + " 860 0.1807 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 74 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 74 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16643 โ†’ Val=0.16398\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 74m\n", + "\n", + "๐Ÿ”„ Epoch 75/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1348 23.2% 2.8m\n", + " 10 0.1661 23.2% 2.6m\n", + " 20 0.1641 23.2% 2.6m\n", + " 30 0.1605 23.2% 2.6m\n", + " 40 0.1486 23.2% 2.6m\n", + " 50 0.1560 23.2% 2.5m\n", + " 60 0.2036 23.2% 2.5m\n", + " 70 0.1772 23.2% 2.5m\n", + " 80 0.1519 23.2% 2.4m\n", + " 90 0.1897 23.2% 2.4m\n", + " 100 0.1337 23.2% 2.4m\n", + " 110 0.1510 23.2% 2.4m\n", + " 120 0.1208 23.2% 2.3m\n", + " 130 0.1406 23.2% 2.3m\n", + " 140 0.1727 23.2% 2.3m\n", + " 150 0.1663 23.2% 2.2m\n", + " 160 0.1448 23.2% 2.2m\n", + " 170 0.1451 23.2% 2.2m\n", + " 180 0.1745 23.2% 2.1m\n", + " 190 0.1352 23.2% 2.1m\n", + " 200 0.1599 23.2% 2.1m\n", + " 210 0.2069 23.2% 2.0m\n", + " 220 0.1652 23.2% 2.0m\n", + " 230 0.1898 23.2% 2.0m\n", + " 240 0.1564 23.2% 1.9m\n", + " 250 0.1997 23.2% 1.9m\n", + " 260 0.1717 23.2% 1.9m\n", + " 270 0.1949 23.2% 1.9m\n", + " 280 0.1976 23.2% 1.8m\n", + " 290 0.1963 23.2% 1.8m\n", + " 300 0.1496 23.2% 1.8m\n", + " 310 0.1628 23.2% 1.7m\n", + " 320 0.2011 23.2% 1.7m\n", + " 330 0.1705 23.2% 1.7m\n", + " 340 0.1339 23.2% 1.6m\n", + " 350 0.1087 23.2% 1.6m\n", + " 360 0.1618 23.2% 1.6m\n", + " 370 0.1727 23.2% 1.6m\n", + " 380 0.2023 23.2% 1.5m\n", + " 390 0.1783 23.2% 1.5m\n", + " 400 0.1583 23.2% 1.5m\n", + " 410 0.1453 23.2% 1.4m\n", + " 420 0.1703 23.2% 1.4m\n", + " 430 0.1885 23.2% 1.4m\n", + " 440 0.1643 23.2% 1.3m\n", + " 450 0.1600 23.2% 1.3m\n", + " 460 0.2062 23.2% 1.3m\n", + " 470 0.1724 23.2% 1.2m\n", + " 480 0.1652 23.2% 1.2m\n", + " 490 0.1844 23.2% 1.2m\n", + " 500 0.1716 23.2% 1.1m\n", + " 510 0.1437 23.2% 1.1m\n", + " 520 0.1692 23.2% 1.1m\n", + " 530 0.1661 23.2% 1.0m\n", + " 540 0.1662 23.2% 1.0m\n", + " 550 0.1950 23.2% 1.0m\n", + " 560 0.1593 23.2% 1.0m\n", + " 570 0.1964 23.2% 0.9m\n", + " 580 0.1304 23.2% 0.9m\n", + " 590 0.1669 23.2% 0.9m\n", + " 600 0.1820 23.2% 0.8m\n", + " 610 0.1414 23.2% 0.8m\n", + " 620 0.1449 23.2% 0.8m\n", + " 630 0.1507 23.2% 0.7m\n", + " 640 0.1519 23.2% 0.7m\n", + " 650 0.1707 23.2% 0.7m\n", + " 660 0.1791 23.2% 0.6m\n", + " 670 0.1912 23.2% 0.6m\n", + " 680 0.1780 23.2% 0.6m\n", + " 690 0.1987 23.2% 0.5m\n", + " 700 0.1729 23.2% 0.5m\n", + " 710 0.1531 23.2% 0.5m\n", + " 720 0.1278 23.2% 0.4m\n", + " 730 0.1589 23.2% 0.4m\n", + " 740 0.1716 23.2% 0.4m\n", + " 750 0.1807 23.2% 0.4m\n", + " 760 0.1420 23.2% 0.3m\n", + " 770 0.1814 23.2% 0.3m\n", + " 780 0.1421 23.2% 0.3m\n", + " 790 0.1812 23.2% 0.2m\n", + " 800 0.1553 23.2% 0.2m\n", + " 810 0.1572 23.2% 0.2m\n", + " 820 0.1356 23.2% 0.1m\n", + " 830 0.1953 23.2% 0.1m\n", + " 840 0.1615 23.2% 0.1m\n", + " 850 0.1806 23.2% 0.0m\n", + " 860 0.1711 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 75 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 75 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16651 โ†’ Val=0.16383\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 71m\n", + "\n", + "๐Ÿ”„ Epoch 76/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2037 23.2% 2.7m\n", + " 10 0.1669 23.2% 2.7m\n", + " 20 0.1869 23.2% 2.6m\n", + " 30 0.1143 23.2% 2.6m\n", + " 40 0.1668 23.2% 2.6m\n", + " 50 0.1278 23.2% 2.5m\n", + " 60 0.2015 23.2% 2.5m\n", + " 70 0.1422 23.2% 2.5m\n", + " 80 0.1795 23.2% 2.4m\n", + " 90 0.1566 23.2% 2.4m\n", + " 100 0.1378 23.2% 2.4m\n", + " 110 0.1734 23.2% 2.4m\n", + " 120 0.1332 23.2% 2.3m\n", + " 130 0.1560 23.2% 2.3m\n", + " 140 0.1951 23.2% 2.3m\n", + " 150 0.1609 23.2% 2.2m\n", + " 160 0.1549 23.2% 2.2m\n", + " 170 0.1579 23.2% 2.2m\n", + " 180 0.1751 23.2% 2.1m\n", + " 190 0.1469 23.2% 2.1m\n", + " 200 0.1397 23.2% 2.1m\n", + " 210 0.1311 23.2% 2.0m\n", + " 220 0.1873 23.2% 2.0m\n", + " 230 0.1644 23.2% 2.0m\n", + " 240 0.1896 23.2% 1.9m\n", + " 250 0.1793 23.2% 1.9m\n", + " 260 0.1711 23.2% 1.9m\n", + " 270 0.1580 23.2% 1.9m\n", + " 280 0.1691 23.2% 1.8m\n", + " 290 0.1373 23.2% 1.8m\n", + " 300 0.1553 23.2% 1.8m\n", + " 310 0.1233 23.2% 1.7m\n", + " 320 0.1334 23.2% 1.7m\n", + " 330 0.1715 23.2% 1.7m\n", + " 340 0.1669 23.2% 1.6m\n", + " 350 0.1756 23.2% 1.6m\n", + " 360 0.1686 23.2% 1.6m\n", + " 370 0.1876 23.2% 1.5m\n", + " 380 0.1714 23.2% 1.5m\n", + " 390 0.1665 23.2% 1.5m\n", + " 400 0.1606 23.2% 1.4m\n", + " 410 0.1346 23.2% 1.4m\n", + " 420 0.1988 23.2% 1.4m\n", + " 430 0.1508 23.2% 1.4m\n", + " 440 0.1486 23.2% 1.3m\n", + " 450 0.1521 23.2% 1.3m\n", + " 460 0.1230 23.2% 1.3m\n", + " 470 0.1558 23.2% 1.2m\n", + " 480 0.1801 23.2% 1.2m\n", + " 490 0.1715 23.2% 1.2m\n", + " 500 0.1881 23.2% 1.1m\n", + " 510 0.1775 23.2% 1.1m\n", + " 520 0.1811 23.2% 1.1m\n", + " 530 0.1591 23.2% 1.0m\n", + " 540 0.1756 23.2% 1.0m\n", + " 550 0.1557 23.2% 1.0m\n", + " 560 0.1532 23.2% 0.9m\n", + " 570 0.1360 23.2% 0.9m\n", + " 580 0.1578 23.2% 0.9m\n", + " 590 0.1953 23.2% 0.9m\n", + " 600 0.1547 23.2% 0.8m\n", + " 610 0.1636 23.2% 0.8m\n", + " 620 0.1919 23.2% 0.8m\n", + " 630 0.1796 23.2% 0.7m\n", + " 640 0.1479 23.2% 0.7m\n", + " 650 0.1487 23.2% 0.7m\n", + " 660 0.1799 23.2% 0.6m\n", + " 670 0.1561 23.2% 0.6m\n", + " 680 0.1759 23.2% 0.6m\n", + " 690 0.1659 23.2% 0.5m\n", + " 700 0.1814 23.2% 0.5m\n", + " 710 0.1783 23.2% 0.5m\n", + " 720 0.1875 23.2% 0.4m\n", + " 730 0.1580 23.2% 0.4m\n", + " 740 0.1896 23.2% 0.4m\n", + " 750 0.1487 23.2% 0.4m\n", + " 760 0.1709 23.2% 0.3m\n", + " 770 0.1853 23.2% 0.3m\n", + " 780 0.1658 23.2% 0.3m\n", + " 790 0.1385 23.2% 0.2m\n", + " 800 0.1845 23.2% 0.2m\n", + " 810 0.1262 23.2% 0.2m\n", + " 820 0.1838 23.2% 0.1m\n", + " 830 0.1604 23.2% 0.1m\n", + " 840 0.1177 23.2% 0.1m\n", + " 850 0.2075 23.2% 0.0m\n", + " 860 0.1993 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 76 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 76 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16653 โ†’ Val=0.16405\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.2%\n", + " โš ๏ธ No improvement for 15/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 68m\n", + "\n", + "๐Ÿ”„ Epoch 77/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1798 23.2% 2.7m\n", + " 10 0.1751 23.2% 2.6m\n", + " 20 0.1804 23.2% 2.6m\n", + " 30 0.2148 23.2% 2.6m\n", + " 40 0.1326 23.2% 2.6m\n", + " 50 0.1738 23.2% 2.5m\n", + " 60 0.1900 23.2% 2.5m\n", + " 70 0.1890 23.2% 2.5m\n", + " 80 0.1439 23.2% 2.4m\n", + " 90 0.1755 23.2% 2.4m\n", + " 100 0.1515 23.2% 2.4m\n", + " 110 0.1823 23.2% 2.4m\n", + " 120 0.1975 23.2% 2.3m\n", + " 130 0.1786 23.2% 2.3m\n", + " 140 0.1462 23.2% 2.3m\n", + " 150 0.1516 23.2% 2.2m\n", + " 160 0.1807 23.2% 2.2m\n", + " 170 0.1990 23.2% 2.2m\n", + " 180 0.1732 23.2% 2.1m\n", + " 190 0.1819 23.2% 2.1m\n", + " 200 0.1429 23.2% 2.1m\n", + " 210 0.1629 23.2% 2.0m\n", + " 220 0.1357 23.2% 2.0m\n", + " 230 0.1663 23.2% 2.0m\n", + " 240 0.1919 23.2% 1.9m\n", + " 250 0.1703 23.2% 1.9m\n", + " 260 0.1364 23.2% 1.9m\n", + " 270 0.1414 23.2% 1.9m\n", + " 280 0.1764 23.2% 1.8m\n", + " 290 0.1405 23.2% 1.8m\n", + " 300 0.1938 23.2% 1.8m\n", + " 310 0.1808 23.2% 1.7m\n", + " 320 0.1786 23.2% 1.7m\n", + " 330 0.1984 23.2% 1.7m\n", + " 340 0.1628 23.2% 1.6m\n", + " 350 0.1894 23.2% 1.6m\n", + " 360 0.1567 23.2% 1.6m\n", + " 370 0.1705 23.2% 1.5m\n", + " 380 0.1451 23.2% 1.5m\n", + " 390 0.1432 23.2% 1.5m\n", + " 400 0.1674 23.2% 1.4m\n", + " 410 0.1528 23.2% 1.4m\n", + " 420 0.1455 23.2% 1.4m\n", + " 430 0.1896 23.2% 1.4m\n", + " 440 0.1353 23.2% 1.3m\n", + " 450 0.1626 23.2% 1.3m\n", + " 460 0.1281 23.2% 1.3m\n", + " 470 0.1764 23.2% 1.2m\n", + " 480 0.1484 23.2% 1.2m\n", + " 490 0.1847 23.2% 1.2m\n", + " 500 0.1536 23.2% 1.1m\n", + " 510 0.1771 23.2% 1.1m\n", + " 520 0.1886 23.2% 1.1m\n", + " 530 0.2017 23.2% 1.0m\n", + " 540 0.1902 23.2% 1.0m\n", + " 550 0.1580 23.2% 1.0m\n", + " 560 0.1302 23.2% 0.9m\n", + " 570 0.1989 23.2% 0.9m\n", + " 580 0.1744 23.2% 0.9m\n", + " 590 0.1519 23.2% 0.9m\n", + " 600 0.1923 23.2% 0.8m\n", + " 610 0.1786 23.2% 0.8m\n", + " 620 0.1568 23.2% 0.8m\n", + " 630 0.1961 23.2% 0.7m\n", + " 640 0.1428 23.2% 0.7m\n", + " 650 0.1798 23.2% 0.7m\n", + " 660 0.1342 23.2% 0.6m\n", + " 670 0.1386 23.2% 0.6m\n", + " 680 0.1494 23.2% 0.6m\n", + " 690 0.2066 23.2% 0.5m\n", + " 700 0.2313 23.2% 0.5m\n", + " 710 0.1557 23.2% 0.5m\n", + " 720 0.1199 23.2% 0.4m\n", + " 730 0.1434 23.2% 0.4m\n", + " 740 0.1493 23.2% 0.4m\n", + " 750 0.1681 23.2% 0.4m\n", + " 760 0.2189 23.2% 0.3m\n", + " 770 0.1352 23.2% 0.3m\n", + " 780 0.1227 23.2% 0.3m\n", + " 790 0.1951 23.2% 0.2m\n", + " 800 0.1770 23.2% 0.2m\n", + " 810 0.1724 23.2% 0.2m\n", + " 820 0.1819 23.2% 0.1m\n", + " 830 0.1714 23.2% 0.1m\n", + " 840 0.1627 23.2% 0.1m\n", + " 850 0.1381 23.2% 0.0m\n", + " 860 0.1486 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 77 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 77 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16662 โ†’ Val=0.16388\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 65m\n", + "\n", + "๐Ÿ”„ Epoch 78/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1824 23.2% 2.7m\n", + " 10 0.1967 23.2% 2.7m\n", + " 20 0.1647 23.2% 2.6m\n", + " 30 0.2025 23.2% 2.6m\n", + " 40 0.1466 23.2% 2.6m\n", + " 50 0.1983 23.2% 2.5m\n", + " 60 0.1946 23.2% 2.5m\n", + " 70 0.1583 23.2% 2.5m\n", + " 80 0.1624 23.2% 2.4m\n", + " 90 0.1833 23.2% 2.4m\n", + " 100 0.1712 23.2% 2.4m\n", + " 110 0.1471 23.2% 2.3m\n", + " 120 0.1741 23.2% 2.3m\n", + " 130 0.1675 23.2% 2.3m\n", + " 140 0.1823 23.2% 2.3m\n", + " 150 0.1681 23.2% 2.2m\n", + " 160 0.1104 23.2% 2.2m\n", + " 170 0.1678 23.2% 2.2m\n", + " 180 0.1787 23.2% 2.1m\n", + " 190 0.1473 23.2% 2.1m\n", + " 200 0.1752 23.2% 2.1m\n", + " 210 0.1301 23.2% 2.0m\n", + " 220 0.1538 23.2% 2.0m\n", + " 230 0.1710 23.2% 2.0m\n", + " 240 0.1476 23.2% 1.9m\n", + " 250 0.1587 23.2% 1.9m\n", + " 260 0.1672 23.2% 1.9m\n", + " 270 0.1664 23.2% 1.9m\n", + " 280 0.1546 23.2% 1.8m\n", + " 290 0.1344 23.2% 1.8m\n", + " 300 0.1911 23.2% 1.8m\n", + " 310 0.1652 23.2% 1.7m\n", + " 320 0.2135 23.2% 1.7m\n", + " 330 0.1640 23.2% 1.7m\n", + " 340 0.1398 23.2% 1.6m\n", + " 350 0.1833 23.2% 1.6m\n", + " 360 0.1926 23.2% 1.6m\n", + " 370 0.1547 23.2% 1.5m\n", + " 380 0.2029 23.2% 1.5m\n", + " 390 0.1787 23.2% 1.5m\n", + " 400 0.1685 23.2% 1.4m\n", + " 410 0.1120 23.2% 1.4m\n", + " 420 0.1796 23.2% 1.4m\n", + " 430 0.1672 23.2% 1.4m\n", + " 440 0.1863 23.2% 1.3m\n", + " 450 0.1562 23.2% 1.3m\n", + " 460 0.1580 23.2% 1.3m\n", + " 470 0.1322 23.2% 1.2m\n", + " 480 0.1613 23.2% 1.2m\n", + " 490 0.1461 23.2% 1.2m\n", + " 500 0.1406 23.2% 1.1m\n", + " 510 0.1466 23.2% 1.1m\n", + " 520 0.1741 23.2% 1.1m\n", + " 530 0.1856 23.2% 1.0m\n", + " 540 0.1829 23.2% 1.0m\n", + " 550 0.1706 23.2% 1.0m\n", + " 560 0.1835 23.2% 0.9m\n", + " 570 0.1644 23.2% 0.9m\n", + " 580 0.1788 23.2% 0.9m\n", + " 590 0.1676 23.2% 0.9m\n", + " 600 0.1545 23.2% 0.8m\n", + " 610 0.1873 23.2% 0.8m\n", + " 620 0.2365 23.2% 0.8m\n", + " 630 0.1748 23.2% 0.7m\n", + " 640 0.1594 23.2% 0.7m\n", + " 650 0.1771 23.2% 0.7m\n", + " 660 0.1478 23.2% 0.6m\n", + " 670 0.1662 23.2% 0.6m\n", + " 680 0.1862 23.2% 0.6m\n", + " 690 0.1618 23.2% 0.5m\n", + " 700 0.1991 23.2% 0.5m\n", + " 710 0.1905 23.2% 0.5m\n", + " 720 0.1609 23.2% 0.4m\n", + " 730 0.1463 23.2% 0.4m\n", + " 740 0.1653 23.2% 0.4m\n", + " 750 0.1506 23.2% 0.4m\n", + " 760 0.1450 23.2% 0.3m\n", + " 770 0.1453 23.2% 0.3m\n", + " 780 0.1877 23.2% 0.3m\n", + " 790 0.1444 23.2% 0.2m\n", + " 800 0.1438 23.2% 0.2m\n", + " 810 0.1962 23.2% 0.2m\n", + " 820 0.1912 23.2% 0.1m\n", + " 830 0.1197 23.2% 0.1m\n", + " 840 0.1370 23.2% 0.1m\n", + " 850 0.1571 23.2% 0.0m\n", + " 860 0.1633 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 78 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 78 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16641 โ†’ Val=0.16401\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 63m\n", + "\n", + "๐Ÿ”„ Epoch 79/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1544 23.2% 2.7m\n", + " 10 0.1667 23.2% 2.8m\n", + " 20 0.1957 23.2% 2.7m\n", + " 30 0.2024 23.2% 2.7m\n", + " 40 0.1744 23.2% 2.6m\n", + " 50 0.1680 23.2% 2.6m\n", + " 60 0.1170 23.2% 2.6m\n", + " 70 0.1793 23.2% 2.5m\n", + " 80 0.1252 23.2% 2.5m\n", + " 90 0.1689 23.2% 2.5m\n", + " 100 0.1656 23.2% 2.4m\n", + " 110 0.1716 23.2% 2.4m\n", + " 120 0.1433 23.2% 2.4m\n", + " 130 0.1499 23.2% 2.3m\n", + " 140 0.1560 23.2% 2.3m\n", + " 150 0.1407 23.2% 2.3m\n", + " 160 0.1162 23.2% 2.2m\n", + " 170 0.1979 23.2% 2.2m\n", + " 180 0.2258 23.2% 2.2m\n", + " 190 0.1888 23.2% 2.1m\n", + " 200 0.1729 23.2% 2.1m\n", + " 210 0.1514 23.2% 2.1m\n", + " 220 0.1950 23.2% 2.0m\n", + " 230 0.1774 23.2% 2.0m\n", + " 240 0.2106 23.2% 2.0m\n", + " 250 0.1561 23.2% 1.9m\n", + " 260 0.1069 23.2% 1.9m\n", + " 270 0.1616 23.2% 1.9m\n", + " 280 0.1912 23.2% 1.9m\n", + " 290 0.2264 23.2% 1.8m\n", + " 300 0.1882 23.2% 1.8m\n", + " 310 0.1570 23.2% 1.8m\n", + " 320 0.1541 23.2% 1.7m\n", + " 330 0.1350 23.2% 1.7m\n", + " 340 0.1661 23.2% 1.7m\n", + " 350 0.2033 23.2% 1.6m\n", + " 360 0.1638 23.2% 1.6m\n", + " 370 0.1729 23.2% 1.6m\n", + " 380 0.1376 23.2% 1.5m\n", + " 390 0.1508 23.2% 1.5m\n", + " 400 0.1655 23.2% 1.5m\n", + " 410 0.1700 23.2% 1.4m\n", + " 420 0.1493 23.2% 1.4m\n", + " 430 0.1445 23.2% 1.4m\n", + " 440 0.1949 23.2% 1.3m\n", + " 450 0.1896 23.2% 1.3m\n", + " 460 0.1279 23.2% 1.3m\n", + " 470 0.1323 23.2% 1.2m\n", + " 480 0.1505 23.2% 1.2m\n", + " 490 0.1708 23.2% 1.2m\n", + " 500 0.1616 23.2% 1.2m\n", + " 510 0.1328 23.2% 1.1m\n", + " 520 0.1582 23.2% 1.1m\n", + " 530 0.1460 23.2% 1.1m\n", + " 540 0.1882 23.2% 1.0m\n", + " 550 0.1387 23.2% 1.0m\n", + " 560 0.1766 23.2% 1.0m\n", + " 570 0.1780 23.2% 0.9m\n", + " 580 0.2185 23.2% 0.9m\n", + " 590 0.1718 23.2% 0.9m\n", + " 600 0.1903 23.2% 0.8m\n", + " 610 0.1646 23.2% 0.8m\n", + " 620 0.1385 23.2% 0.8m\n", + " 630 0.1913 23.2% 0.7m\n", + " 640 0.2172 23.2% 0.7m\n", + " 650 0.1629 23.2% 0.7m\n", + " 660 0.1854 23.2% 0.6m\n", + " 670 0.1500 23.2% 0.6m\n", + " 680 0.1713 23.2% 0.6m\n", + " 690 0.1726 23.2% 0.5m\n", + " 700 0.1583 23.2% 0.5m\n", + " 710 0.1601 23.2% 0.5m\n", + " 720 0.1917 23.2% 0.5m\n", + " 730 0.1866 23.2% 0.4m\n", + " 740 0.1631 23.2% 0.4m\n", + " 750 0.1991 23.2% 0.4m\n", + " 760 0.1595 23.2% 0.3m\n", + " 770 0.2224 23.2% 0.3m\n", + " 780 0.1774 23.2% 0.3m\n", + " 790 0.1644 23.2% 0.2m\n", + " 800 0.1798 23.2% 0.2m\n", + " 810 0.1756 23.2% 0.2m\n", + " 820 0.1766 23.2% 0.1m\n", + " 830 0.1557 23.2% 0.1m\n", + " 840 0.1728 23.2% 0.1m\n", + " 850 0.1486 23.2% 0.0m\n", + " 860 0.1786 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 79 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 79 SUMMARY\n", + " โฑ๏ธ Time: 173s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16661 โ†’ Val=0.16400\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 60m\n", + "\n", + "๐Ÿ”„ Epoch 80/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1145 23.2% 2.7m\n", + " 10 0.1377 23.2% 2.7m\n", + " 20 0.1785 23.2% 2.6m\n", + " 30 0.1435 23.2% 2.6m\n", + " 40 0.2029 23.2% 2.6m\n", + " 50 0.1298 23.2% 2.6m\n", + " 60 0.1778 23.2% 2.5m\n", + " 70 0.1517 23.2% 2.5m\n", + " 80 0.1739 23.2% 2.5m\n", + " 90 0.1570 23.2% 2.4m\n", + " 100 0.1989 23.2% 2.4m\n", + " 110 0.2048 23.2% 2.4m\n", + " 120 0.1880 23.2% 2.3m\n", + " 130 0.1962 23.2% 2.3m\n", + " 140 0.1374 23.2% 2.3m\n", + " 150 0.1320 23.2% 2.2m\n", + " 160 0.1995 23.2% 2.2m\n", + " 170 0.1314 23.2% 2.2m\n", + " 180 0.2035 23.2% 2.1m\n", + " 190 0.1898 23.2% 2.1m\n", + " 200 0.1866 23.2% 2.1m\n", + " 210 0.1441 23.2% 2.1m\n", + " 220 0.1825 23.2% 2.0m\n", + " 230 0.1545 23.2% 2.0m\n", + " 240 0.1576 23.2% 2.0m\n", + " 250 0.1930 23.2% 1.9m\n", + " 260 0.2070 23.2% 1.9m\n", + " 270 0.1374 23.2% 1.9m\n", + " 280 0.1814 23.2% 1.8m\n", + " 290 0.1471 23.2% 1.8m\n", + " 300 0.1686 23.2% 1.8m\n", + " 310 0.1863 23.2% 1.7m\n", + " 320 0.1584 23.2% 1.7m\n", + " 330 0.1196 23.2% 1.7m\n", + " 340 0.1634 23.2% 1.6m\n", + " 350 0.1668 23.2% 1.6m\n", + " 360 0.1660 23.2% 1.6m\n", + " 370 0.1800 23.2% 1.5m\n", + " 380 0.2121 23.2% 1.5m\n", + " 390 0.1809 23.2% 1.5m\n", + " 400 0.1916 23.2% 1.5m\n", + " 410 0.1419 23.2% 1.4m\n", + " 420 0.1849 23.2% 1.4m\n", + " 430 0.1699 23.2% 1.4m\n", + " 440 0.2032 23.2% 1.3m\n", + " 450 0.1612 23.2% 1.3m\n", + " 460 0.1666 23.2% 1.3m\n", + " 470 0.1657 23.2% 1.2m\n", + " 480 0.1576 23.2% 1.2m\n", + " 490 0.1344 23.2% 1.2m\n", + " 500 0.1182 23.2% 1.1m\n", + " 510 0.1594 23.2% 1.1m\n", + " 520 0.1594 23.2% 1.1m\n", + " 530 0.1620 23.2% 1.0m\n", + " 540 0.1896 23.2% 1.0m\n", + " 550 0.1829 23.2% 1.0m\n", + " 560 0.1763 23.2% 0.9m\n", + " 570 0.2276 23.2% 0.9m\n", + " 580 0.1653 23.2% 0.9m\n", + " 590 0.1524 23.2% 0.9m\n", + " 600 0.1824 23.2% 0.8m\n", + " 610 0.1903 23.2% 0.8m\n", + " 620 0.1606 23.2% 0.8m\n", + " 630 0.1338 23.2% 0.7m\n", + " 640 0.1431 23.2% 0.7m\n", + " 650 0.1485 23.2% 0.7m\n", + " 660 0.1624 23.2% 0.6m\n", + " 670 0.2051 23.2% 0.6m\n", + " 680 0.1739 23.2% 0.6m\n", + " 690 0.1799 23.2% 0.5m\n", + " 700 0.1633 23.2% 0.5m\n", + " 710 0.1396 23.2% 0.5m\n", + " 720 0.1535 23.2% 0.4m\n", + " 730 0.1647 23.2% 0.4m\n", + " 740 0.1384 23.2% 0.4m\n", + " 750 0.1726 23.2% 0.4m\n", + " 760 0.1645 23.2% 0.3m\n", + " 770 0.1455 23.2% 0.3m\n", + " 780 0.1752 23.2% 0.3m\n", + " 790 0.1808 23.2% 0.2m\n", + " 800 0.1699 23.2% 0.2m\n", + " 810 0.1662 23.2% 0.2m\n", + " 820 0.1892 23.2% 0.1m\n", + " 830 0.1632 23.2% 0.1m\n", + " 840 0.1821 23.2% 0.1m\n", + " 850 0.2102 23.2% 0.0m\n", + " 860 0.1622 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 80 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 80 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16646 โ†’ Val=0.16386\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 57m\n", + "\n", + "๐Ÿ”„ Epoch 81/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1867 23.2% 2.6m\n", + " 10 0.1815 23.2% 2.7m\n", + " 20 0.2033 23.2% 2.7m\n", + " 30 0.1737 23.2% 2.6m\n", + " 40 0.1712 23.2% 2.6m\n", + " 50 0.1817 23.2% 2.6m\n", + " 60 0.1820 23.2% 2.5m\n", + " 70 0.1389 23.2% 2.5m\n", + " 80 0.1422 23.2% 2.5m\n", + " 90 0.1733 23.2% 2.4m\n", + " 100 0.1714 23.2% 2.4m\n", + " 110 0.2039 23.2% 2.4m\n", + " 120 0.1535 23.2% 2.4m\n", + " 130 0.1715 23.2% 2.3m\n", + " 140 0.1465 23.2% 2.3m\n", + " 150 0.1833 23.2% 2.3m\n", + " 160 0.1446 23.2% 2.2m\n", + " 170 0.1506 23.2% 2.2m\n", + " 180 0.1858 23.2% 2.2m\n", + " 190 0.1941 23.2% 2.1m\n", + " 200 0.1753 23.2% 2.1m\n", + " 210 0.1823 23.2% 2.1m\n", + " 220 0.1700 23.2% 2.0m\n", + " 230 0.1744 23.2% 2.0m\n", + " 240 0.1403 23.2% 2.0m\n", + " 250 0.1276 23.2% 2.0m\n", + " 260 0.1814 23.2% 1.9m\n", + " 270 0.1889 23.2% 1.9m\n", + " 280 0.1803 23.2% 1.9m\n", + " 290 0.1066 23.2% 1.8m\n", + " 300 0.1767 23.2% 1.8m\n", + " 310 0.1639 23.2% 1.8m\n", + " 320 0.1099 23.2% 1.7m\n", + " 330 0.1657 23.2% 1.7m\n", + " 340 0.1893 23.2% 1.7m\n", + " 350 0.1560 23.2% 1.6m\n", + " 360 0.1700 23.2% 1.6m\n", + " 370 0.1627 23.2% 1.6m\n", + " 380 0.1800 23.2% 1.5m\n", + " 390 0.1887 23.2% 1.5m\n", + " 400 0.1272 23.2% 1.5m\n", + " 410 0.1626 23.2% 1.4m\n", + " 420 0.1548 23.2% 1.4m\n", + " 430 0.1623 23.2% 1.4m\n", + " 440 0.1599 23.2% 1.3m\n", + " 450 0.1435 23.2% 1.3m\n", + " 460 0.1784 23.2% 1.3m\n", + " 470 0.1875 23.2% 1.3m\n", + " 480 0.1565 23.2% 1.2m\n", + " 490 0.1513 23.2% 1.2m\n", + " 500 0.1556 23.2% 1.2m\n", + " 510 0.1269 23.2% 1.1m\n", + " 520 0.1744 23.2% 1.1m\n", + " 530 0.1582 23.2% 1.1m\n", + " 540 0.1674 23.2% 1.0m\n", + " 550 0.1435 23.2% 1.0m\n", + " 560 0.1681 23.2% 1.0m\n", + " 570 0.1710 23.2% 0.9m\n", + " 580 0.1358 23.2% 0.9m\n", + " 590 0.1564 23.2% 0.9m\n", + " 600 0.1558 23.2% 0.8m\n", + " 610 0.1586 23.2% 0.8m\n", + " 620 0.1895 23.2% 0.8m\n", + " 630 0.2110 23.2% 0.7m\n", + " 640 0.1856 23.2% 0.7m\n", + " 650 0.1719 23.2% 0.7m\n", + " 660 0.2094 23.2% 0.6m\n", + " 670 0.2307 23.2% 0.6m\n", + " 680 0.1966 23.2% 0.6m\n", + " 690 0.1620 23.2% 0.6m\n", + " 700 0.1458 23.2% 0.5m\n", + " 710 0.1335 23.2% 0.5m\n", + " 720 0.1537 23.2% 0.5m\n", + " 730 0.1650 23.2% 0.4m\n", + " 740 0.1728 23.2% 0.4m\n", + " 750 0.1659 23.2% 0.4m\n", + " 760 0.1765 23.2% 0.3m\n", + " 770 0.1775 23.2% 0.3m\n", + " 780 0.1868 23.2% 0.3m\n", + " 790 0.1982 23.2% 0.2m\n", + " 800 0.1486 23.2% 0.2m\n", + " 810 0.1655 23.2% 0.2m\n", + " 820 0.0965 23.2% 0.1m\n", + " 830 0.1582 23.2% 0.1m\n", + " 840 0.1433 23.2% 0.1m\n", + " 850 0.1293 23.2% 0.0m\n", + " 860 0.1564 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 81 COMPLETE in 165s\n", + "\n", + "โœ… EPOCH 81 SUMMARY\n", + " โฑ๏ธ Time: 173s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16645 โ†’ Val=0.16451\n", + " ๐Ÿ“Š Acc: EC=81.4% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โš ๏ธ No improvement for 20/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 55m\n", + "\n", + "๐Ÿ”„ Epoch 82/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1559 23.2% 2.6m\n", + " 10 0.2327 23.2% 2.7m\n", + " 20 0.1985 23.2% 2.7m\n", + " 30 0.1846 23.2% 2.7m\n", + " 40 0.1847 23.2% 2.6m\n", + " 50 0.1564 23.2% 2.6m\n", + " 60 0.1450 23.2% 2.5m\n", + " 70 0.1439 23.2% 2.5m\n", + " 80 0.1378 23.2% 2.5m\n", + " 90 0.1473 23.2% 2.5m\n", + " 100 0.1918 23.2% 2.4m\n", + " 110 0.2194 23.2% 2.4m\n", + " 120 0.1740 23.2% 2.4m\n", + " 130 0.1621 23.2% 2.3m\n", + " 140 0.1632 23.2% 2.3m\n", + " 150 0.1396 23.2% 2.3m\n", + " 160 0.1954 23.2% 2.2m\n", + " 170 0.1847 23.2% 2.2m\n", + " 180 0.1653 23.2% 2.2m\n", + " 190 0.1929 23.2% 2.1m\n", + " 200 0.1902 23.2% 2.1m\n", + " 210 0.1407 23.2% 2.1m\n", + " 220 0.1575 23.2% 2.0m\n", + " 230 0.1380 23.2% 2.0m\n", + " 240 0.1767 23.2% 2.0m\n", + " 250 0.1757 23.2% 1.9m\n", + " 260 0.1639 23.2% 1.9m\n", + " 270 0.1640 23.2% 1.9m\n", + " 280 0.1679 23.2% 1.9m\n", + " 290 0.1492 23.2% 1.8m\n", + " 300 0.1768 23.2% 1.8m\n", + " 310 0.1792 23.2% 1.8m\n", + " 320 0.2023 23.2% 1.7m\n", + " 330 0.1616 23.2% 1.7m\n", + " 340 0.1662 23.2% 1.7m\n", + " 350 0.1519 23.2% 1.6m\n", + " 360 0.2021 23.2% 1.6m\n", + " 370 0.1634 23.2% 1.6m\n", + " 380 0.1579 23.2% 1.5m\n", + " 390 0.1886 23.2% 1.5m\n", + " 400 0.1744 23.2% 1.5m\n", + " 410 0.1791 23.2% 1.4m\n", + " 420 0.1868 23.2% 1.4m\n", + " 430 0.1551 23.2% 1.4m\n", + " 440 0.1292 23.2% 1.3m\n", + " 450 0.1671 23.2% 1.3m\n", + " 460 0.1782 23.2% 1.3m\n", + " 470 0.1976 23.2% 1.2m\n", + " 480 0.1672 23.2% 1.2m\n", + " 490 0.1519 23.2% 1.2m\n", + " 500 0.1600 23.2% 1.2m\n", + " 510 0.1430 23.2% 1.1m\n", + " 520 0.1449 23.2% 1.1m\n", + " 530 0.1561 23.2% 1.1m\n", + " 540 0.1631 23.2% 1.0m\n", + " 550 0.1747 23.2% 1.0m\n", + " 560 0.1742 23.2% 1.0m\n", + " 570 0.1817 23.2% 0.9m\n", + " 580 0.1813 23.2% 0.9m\n", + " 590 0.1615 23.2% 0.9m\n", + " 600 0.1755 23.2% 0.8m\n", + " 610 0.1908 23.2% 0.8m\n", + " 620 0.1899 23.2% 0.8m\n", + " 630 0.1790 23.2% 0.7m\n", + " 640 0.1204 23.2% 0.7m\n", + " 650 0.1654 23.2% 0.7m\n", + " 660 0.1455 23.2% 0.6m\n", + " 670 0.1572 23.2% 0.6m\n", + " 680 0.1800 23.2% 0.6m\n", + " 690 0.1423 23.2% 0.5m\n", + " 700 0.1510 23.2% 0.5m\n", + " 710 0.1483 23.2% 0.5m\n", + " 720 0.1831 23.2% 0.5m\n", + " 730 0.1802 23.2% 0.4m\n", + " 740 0.1516 23.2% 0.4m\n", + " 750 0.1879 23.2% 0.4m\n", + " 760 0.1797 23.2% 0.3m\n", + " 770 0.1726 23.2% 0.3m\n", + " 780 0.2022 23.2% 0.3m\n", + " 790 0.1827 23.2% 0.2m\n", + " 800 0.1292 23.2% 0.2m\n", + " 810 0.1458 23.2% 0.2m\n", + " 820 0.1461 23.2% 0.1m\n", + " 830 0.1278 23.2% 0.1m\n", + " 840 0.1801 23.2% 0.1m\n", + " 850 0.1471 23.2% 0.0m\n", + " 860 0.1978 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 82 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 82 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16654 โ†’ Val=0.16407\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.8% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 52m\n", + "\n", + "๐Ÿ”„ Epoch 83/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1728 23.2% 2.8m\n", + " 10 0.1666 23.2% 2.7m\n", + " 20 0.1575 23.2% 2.7m\n", + " 30 0.1267 23.2% 2.6m\n", + " 40 0.1826 23.2% 2.6m\n", + " 50 0.1527 23.2% 2.6m\n", + " 60 0.1916 23.2% 2.5m\n", + " 70 0.1591 23.2% 2.5m\n", + " 80 0.1425 23.2% 2.5m\n", + " 90 0.1285 23.2% 2.4m\n", + " 100 0.1519 23.2% 2.4m\n", + " 110 0.1392 23.2% 2.4m\n", + " 120 0.2215 23.2% 2.3m\n", + " 130 0.2009 23.2% 2.3m\n", + " 140 0.1454 23.2% 2.3m\n", + " 150 0.1226 23.2% 2.2m\n", + " 160 0.1675 23.2% 2.2m\n", + " 170 0.1669 23.2% 2.2m\n", + " 180 0.1324 23.2% 2.1m\n", + " 190 0.1925 23.2% 2.1m\n", + " 200 0.1696 23.2% 2.1m\n", + " 210 0.2047 23.2% 2.0m\n", + " 220 0.1946 23.2% 2.0m\n", + " 230 0.1476 23.2% 2.0m\n", + " 240 0.1424 23.2% 2.0m\n", + " 250 0.1837 23.2% 1.9m\n", + " 260 0.1817 23.2% 1.9m\n", + " 270 0.1717 23.2% 1.9m\n", + " 280 0.1644 23.2% 1.8m\n", + " 290 0.1302 23.2% 1.8m\n", + " 300 0.1844 23.2% 1.8m\n", + " 310 0.1432 23.2% 1.7m\n", + " 320 0.1508 23.2% 1.7m\n", + " 330 0.1744 23.2% 1.7m\n", + " 340 0.2077 23.2% 1.6m\n", + " 350 0.1764 23.2% 1.6m\n", + " 360 0.1954 23.2% 1.6m\n", + " 370 0.1818 23.2% 1.5m\n", + " 380 0.1582 23.2% 1.5m\n", + " 390 0.1597 23.2% 1.5m\n", + " 400 0.2034 23.2% 1.4m\n", + " 410 0.1640 23.2% 1.4m\n", + " 420 0.1683 23.2% 1.4m\n", + " 430 0.2173 23.2% 1.4m\n", + " 440 0.1558 23.2% 1.3m\n", + " 450 0.1501 23.2% 1.3m\n", + " 460 0.1798 23.2% 1.3m\n", + " 470 0.1463 23.2% 1.2m\n", + " 480 0.1692 23.2% 1.2m\n", + " 490 0.1935 23.2% 1.2m\n", + " 500 0.1455 23.2% 1.1m\n", + " 510 0.1735 23.2% 1.1m\n", + " 520 0.1460 23.2% 1.1m\n", + " 530 0.1906 23.2% 1.0m\n", + " 540 0.1497 23.2% 1.0m\n", + " 550 0.2111 23.2% 1.0m\n", + " 560 0.1573 23.2% 0.9m\n", + " 570 0.1792 23.2% 0.9m\n", + " 580 0.1763 23.2% 0.9m\n", + " 590 0.1382 23.2% 0.9m\n", + " 600 0.1543 23.2% 0.8m\n", + " 610 0.2190 23.2% 0.8m\n", + " 620 0.1591 23.2% 0.8m\n", + " 630 0.1921 23.2% 0.7m\n", + " 640 0.1530 23.2% 0.7m\n", + " 650 0.1670 23.2% 0.7m\n", + " 660 0.1109 23.2% 0.6m\n", + " 670 0.1450 23.2% 0.6m\n", + " 680 0.1521 23.2% 0.6m\n", + " 690 0.1351 23.2% 0.5m\n", + " 700 0.1735 23.2% 0.5m\n", + " 710 0.1586 23.2% 0.5m\n", + " 720 0.1748 23.2% 0.4m\n", + " 730 0.1509 23.2% 0.4m\n", + " 740 0.1318 23.2% 0.4m\n", + " 750 0.1252 23.2% 0.4m\n", + " 760 0.1864 23.2% 0.3m\n", + " 770 0.1830 23.2% 0.3m\n", + " 780 0.1465 23.2% 0.3m\n", + " 790 0.1452 23.2% 0.2m\n", + " 800 0.1217 23.2% 0.2m\n", + " 810 0.1684 23.2% 0.2m\n", + " 820 0.1817 23.2% 0.1m\n", + " 830 0.1781 23.2% 0.1m\n", + " 840 0.1946 23.2% 0.1m\n", + " 850 0.1587 23.2% 0.0m\n", + " 860 0.1595 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 83 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 83 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16633 โ†’ Val=0.16375\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16375)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 48m\n", + "\n", + "๐Ÿ”„ Epoch 84/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.2196 23.2% 2.7m\n", + " 10 0.1989 23.2% 2.7m\n", + " 20 0.1943 23.2% 2.6m\n", + " 30 0.1574 23.2% 2.6m\n", + " 40 0.1464 23.2% 2.6m\n", + " 50 0.1506 23.2% 2.5m\n", + " 60 0.1831 23.2% 2.5m\n", + " 70 0.1698 23.2% 2.5m\n", + " 80 0.1998 23.2% 2.4m\n", + " 90 0.1519 23.2% 2.4m\n", + " 100 0.1856 23.2% 2.4m\n", + " 110 0.1482 23.2% 2.4m\n", + " 120 0.2076 23.2% 2.3m\n", + " 130 0.1909 23.2% 2.3m\n", + " 140 0.1783 23.2% 2.3m\n", + " 150 0.1873 23.2% 2.2m\n", + " 160 0.1705 23.2% 2.2m\n", + " 170 0.1639 23.2% 2.2m\n", + " 180 0.1681 23.2% 2.1m\n", + " 190 0.1742 23.2% 2.1m\n", + " 200 0.1506 23.2% 2.1m\n", + " 210 0.1984 23.2% 2.1m\n", + " 220 0.2072 23.2% 2.0m\n", + " 230 0.1310 23.2% 2.0m\n", + " 240 0.1896 23.2% 2.0m\n", + " 250 0.1673 23.2% 1.9m\n", + " 260 0.1608 23.2% 1.9m\n", + " 270 0.1625 23.2% 1.9m\n", + " 280 0.1454 23.2% 1.8m\n", + " 290 0.1968 23.2% 1.8m\n", + " 300 0.1819 23.2% 1.8m\n", + " 310 0.1403 23.2% 1.7m\n", + " 320 0.2052 23.2% 1.7m\n", + " 330 0.1856 23.2% 1.7m\n", + " 340 0.1537 23.2% 1.6m\n", + " 350 0.1661 23.2% 1.6m\n", + " 360 0.1979 23.2% 1.6m\n", + " 370 0.1332 23.2% 1.6m\n", + " 380 0.1606 23.2% 1.5m\n", + " 390 0.1908 23.2% 1.5m\n", + " 400 0.1860 23.2% 1.5m\n", + " 410 0.1838 23.2% 1.4m\n", + " 420 0.1941 23.2% 1.4m\n", + " 430 0.1783 23.2% 1.4m\n", + " 440 0.1137 23.2% 1.3m\n", + " 450 0.1990 23.2% 1.3m\n", + " 460 0.1789 23.2% 1.3m\n", + " 470 0.1569 23.2% 1.2m\n", + " 480 0.1505 23.2% 1.2m\n", + " 490 0.1328 23.2% 1.2m\n", + " 500 0.1706 23.2% 1.1m\n", + " 510 0.1528 23.2% 1.1m\n", + " 520 0.1422 23.2% 1.1m\n", + " 530 0.1577 23.2% 1.0m\n", + " 540 0.1716 23.2% 1.0m\n", + " 550 0.1173 23.2% 1.0m\n", + " 560 0.1646 23.2% 1.0m\n", + " 570 0.1858 23.2% 0.9m\n", + " 580 0.1280 23.2% 0.9m\n", + " 590 0.1856 23.2% 0.9m\n", + " 600 0.1392 23.2% 0.8m\n", + " 610 0.1302 23.2% 0.8m\n", + " 620 0.1970 23.2% 0.8m\n", + " 630 0.1684 23.2% 0.7m\n", + " 640 0.1614 23.2% 0.7m\n", + " 650 0.1565 23.2% 0.7m\n", + " 660 0.1673 23.2% 0.6m\n", + " 670 0.1339 23.2% 0.6m\n", + " 680 0.1434 23.2% 0.6m\n", + " 690 0.1582 23.2% 0.5m\n", + " 700 0.1706 23.2% 0.5m\n", + " 710 0.1893 23.2% 0.5m\n", + " 720 0.1130 23.2% 0.5m\n", + " 730 0.2408 23.2% 0.4m\n", + " 740 0.1480 23.2% 0.4m\n", + " 750 0.1584 23.2% 0.4m\n", + " 760 0.2042 23.2% 0.3m\n", + " 770 0.1908 23.2% 0.3m\n", + " 780 0.1659 23.2% 0.3m\n", + " 790 0.1888 23.2% 0.2m\n", + " 800 0.1531 23.2% 0.2m\n", + " 810 0.1502 23.2% 0.2m\n", + " 820 0.1340 23.2% 0.1m\n", + " 830 0.1616 23.2% 0.1m\n", + " 840 0.1577 23.2% 0.1m\n", + " 850 0.1519 23.2% 0.0m\n", + " 860 0.1394 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 84 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 84 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16649 โ†’ Val=0.16390\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 46m\n", + "\n", + "๐Ÿ”„ Epoch 85/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1673 23.2% 2.7m\n", + " 10 0.1696 23.2% 2.7m\n", + " 20 0.1413 23.2% 2.6m\n", + " 30 0.1383 23.2% 2.6m\n", + " 40 0.1906 23.2% 2.6m\n", + " 50 0.1856 23.2% 2.5m\n", + " 60 0.2284 23.2% 2.5m\n", + " 70 0.1386 23.2% 2.5m\n", + " 80 0.1850 23.2% 2.5m\n", + " 90 0.1895 23.2% 2.4m\n", + " 100 0.1278 23.2% 2.4m\n", + " 110 0.1562 23.2% 2.4m\n", + " 120 0.1556 23.2% 2.3m\n", + " 130 0.1327 23.2% 2.3m\n", + " 140 0.1472 23.2% 2.3m\n", + " 150 0.1342 23.2% 2.2m\n", + " 160 0.1694 23.2% 2.2m\n", + " 170 0.1518 23.2% 2.2m\n", + " 180 0.1741 23.2% 2.1m\n", + " 190 0.1343 23.2% 2.1m\n", + " 200 0.1838 23.2% 2.1m\n", + " 210 0.1976 23.2% 2.0m\n", + " 220 0.1733 23.2% 2.0m\n", + " 230 0.1087 23.2% 2.0m\n", + " 240 0.1744 23.2% 2.0m\n", + " 250 0.1793 23.2% 1.9m\n", + " 260 0.1654 23.2% 1.9m\n", + " 270 0.1789 23.2% 1.9m\n", + " 280 0.1591 23.2% 1.8m\n", + " 290 0.1979 23.2% 1.8m\n", + " 300 0.1492 23.2% 1.8m\n", + " 310 0.1399 23.2% 1.7m\n", + " 320 0.1815 23.2% 1.7m\n", + " 330 0.1660 23.2% 1.7m\n", + " 340 0.1974 23.2% 1.6m\n", + " 350 0.1478 23.2% 1.6m\n", + " 360 0.1572 23.2% 1.6m\n", + " 370 0.1573 23.2% 1.6m\n", + " 380 0.1373 23.2% 1.5m\n", + " 390 0.1566 23.2% 1.5m\n", + " 400 0.1757 23.2% 1.5m\n", + " 410 0.1765 23.2% 1.4m\n", + " 420 0.1908 23.2% 1.4m\n", + " 430 0.1496 23.2% 1.4m\n", + " 440 0.1677 23.2% 1.3m\n", + " 450 0.1519 23.2% 1.3m\n", + " 460 0.1861 23.2% 1.3m\n", + " 470 0.1370 23.2% 1.2m\n", + " 480 0.1597 23.2% 1.2m\n", + " 490 0.1813 23.2% 1.2m\n", + " 500 0.1476 23.2% 1.1m\n", + " 510 0.1711 23.2% 1.1m\n", + " 520 0.1566 23.2% 1.1m\n", + " 530 0.1338 23.2% 1.0m\n", + " 540 0.2030 23.2% 1.0m\n", + " 550 0.1937 23.2% 1.0m\n", + " 560 0.1442 23.2% 1.0m\n", + " 570 0.1786 23.2% 0.9m\n", + " 580 0.1492 23.2% 0.9m\n", + " 590 0.1668 23.2% 0.9m\n", + " 600 0.1765 23.2% 0.8m\n", + " 610 0.1896 23.2% 0.8m\n", + " 620 0.1688 23.2% 0.8m\n", + " 630 0.1577 23.2% 0.7m\n", + " 640 0.1329 23.2% 0.7m\n", + " 650 0.1647 23.2% 0.7m\n", + " 660 0.1348 23.2% 0.6m\n", + " 670 0.1161 23.2% 0.6m\n", + " 680 0.1465 23.2% 0.6m\n", + " 690 0.1626 23.2% 0.5m\n", + " 700 0.1492 23.2% 0.5m\n", + " 710 0.1741 23.2% 0.5m\n", + " 720 0.2034 23.2% 0.5m\n", + " 730 0.1427 23.2% 0.4m\n", + " 740 0.1783 23.2% 0.4m\n", + " 750 0.1857 23.2% 0.4m\n", + " 760 0.1598 23.2% 0.3m\n", + " 770 0.1576 23.2% 0.3m\n", + " 780 0.1865 23.2% 0.3m\n", + " 790 0.2265 23.2% 0.2m\n", + " 800 0.1899 23.2% 0.2m\n", + " 810 0.1951 23.2% 0.2m\n", + " 820 0.1446 23.2% 0.1m\n", + " 830 0.1574 23.2% 0.1m\n", + " 840 0.2147 23.2% 0.1m\n", + " 850 0.1585 23.2% 0.0m\n", + " 860 0.1864 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 85 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 85 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16640 โ†’ Val=0.16390\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 43m\n", + "\n", + "๐Ÿ”„ Epoch 86/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1752 23.2% 2.7m\n", + " 10 0.1479 23.2% 2.7m\n", + " 20 0.1756 23.2% 2.7m\n", + " 30 0.1785 23.2% 2.7m\n", + " 40 0.1752 23.2% 2.6m\n", + " 50 0.1536 23.2% 2.6m\n", + " 60 0.1473 23.2% 2.6m\n", + " 70 0.1871 23.2% 2.5m\n", + " 80 0.1801 23.2% 2.5m\n", + " 90 0.1550 23.2% 2.5m\n", + " 100 0.1179 23.2% 2.4m\n", + " 110 0.1297 23.2% 2.4m\n", + " 120 0.1462 23.2% 2.4m\n", + " 130 0.1417 23.2% 2.3m\n", + " 140 0.1338 23.2% 2.3m\n", + " 150 0.1675 23.2% 2.3m\n", + " 160 0.1859 23.2% 2.2m\n", + " 170 0.2143 23.2% 2.2m\n", + " 180 0.1792 23.2% 2.2m\n", + " 190 0.1460 23.2% 2.1m\n", + " 200 0.1754 23.2% 2.1m\n", + " 210 0.1362 23.2% 2.1m\n", + " 220 0.1562 23.2% 2.0m\n", + " 230 0.1620 23.2% 2.0m\n", + " 240 0.1478 23.2% 2.0m\n", + " 250 0.1589 23.2% 1.9m\n", + " 260 0.1864 23.2% 1.9m\n", + " 270 0.1775 23.2% 1.9m\n", + " 280 0.1247 23.2% 1.8m\n", + " 290 0.1703 23.2% 1.8m\n", + " 300 0.1890 23.2% 1.8m\n", + " 310 0.1521 23.2% 1.8m\n", + " 320 0.1631 23.2% 1.7m\n", + " 330 0.1553 23.2% 1.7m\n", + " 340 0.1626 23.2% 1.7m\n", + " 350 0.1562 23.2% 1.6m\n", + " 360 0.1629 23.2% 1.6m\n", + " 370 0.1585 23.2% 1.6m\n", + " 380 0.1694 23.2% 1.5m\n", + " 390 0.1785 23.2% 1.5m\n", + " 400 0.1855 23.2% 1.5m\n", + " 410 0.1611 23.2% 1.4m\n", + " 420 0.1631 23.2% 1.4m\n", + " 430 0.1733 23.2% 1.4m\n", + " 440 0.1627 23.2% 1.3m\n", + " 450 0.1766 23.2% 1.3m\n", + " 460 0.1595 23.2% 1.3m\n", + " 470 0.1596 23.2% 1.2m\n", + " 480 0.1436 23.2% 1.2m\n", + " 490 0.1836 23.2% 1.2m\n", + " 500 0.1516 23.2% 1.1m\n", + " 510 0.1416 23.2% 1.1m\n", + " 520 0.1567 23.2% 1.1m\n", + " 530 0.1485 23.2% 1.1m\n", + " 540 0.1872 23.2% 1.0m\n", + " 550 0.1776 23.2% 1.0m\n", + " 560 0.1767 23.2% 1.0m\n", + " 570 0.1842 23.2% 0.9m\n", + " 580 0.2200 23.2% 0.9m\n", + " 590 0.1986 23.2% 0.9m\n", + " 600 0.1805 23.2% 0.8m\n", + " 610 0.1784 23.2% 0.8m\n", + " 620 0.1929 23.2% 0.8m\n", + " 630 0.1515 23.2% 0.7m\n", + " 640 0.1492 23.2% 0.7m\n", + " 650 0.1648 23.2% 0.7m\n", + " 660 0.1916 23.2% 0.6m\n", + " 670 0.1918 23.2% 0.6m\n", + " 680 0.1559 23.2% 0.6m\n", + " 690 0.1443 23.2% 0.5m\n", + " 700 0.1969 23.2% 0.5m\n", + " 710 0.1865 23.2% 0.5m\n", + " 720 0.1443 23.2% 0.5m\n", + " 730 0.1594 23.2% 0.4m\n", + " 740 0.1124 23.2% 0.4m\n", + " 750 0.1738 23.2% 0.4m\n", + " 760 0.1750 23.2% 0.3m\n", + " 770 0.1534 23.2% 0.3m\n", + " 780 0.1793 23.2% 0.3m\n", + " 790 0.1580 23.2% 0.2m\n", + " 800 0.1931 23.2% 0.2m\n", + " 810 0.1522 23.2% 0.2m\n", + " 820 0.1354 23.2% 0.1m\n", + " 830 0.1657 23.2% 0.1m\n", + " 840 0.1469 23.2% 0.1m\n", + " 850 0.1820 23.2% 0.0m\n", + " 860 0.2048 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 86 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 86 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16645 โ†’ Val=0.16387\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 40m\n", + "\n", + "๐Ÿ”„ Epoch 87/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1703 23.2% 2.7m\n", + " 10 0.1521 23.2% 2.7m\n", + " 20 0.1768 23.2% 2.7m\n", + " 30 0.1739 23.2% 2.7m\n", + " 40 0.1348 23.2% 2.6m\n", + " 50 0.1530 23.2% 2.6m\n", + " 60 0.1605 23.2% 2.6m\n", + " 70 0.2008 23.2% 2.5m\n", + " 80 0.1540 23.2% 2.5m\n", + " 90 0.1730 23.2% 2.5m\n", + " 100 0.1680 23.2% 2.4m\n", + " 110 0.1436 23.2% 2.4m\n", + " 120 0.1582 23.2% 2.4m\n", + " 130 0.1830 23.2% 2.3m\n", + " 140 0.1606 23.2% 2.3m\n", + " 150 0.1338 23.2% 2.3m\n", + " 160 0.1850 23.2% 2.2m\n", + " 170 0.1950 23.2% 2.2m\n", + " 180 0.1441 23.2% 2.2m\n", + " 190 0.1392 23.2% 2.1m\n", + " 200 0.2093 23.2% 2.1m\n", + " 210 0.1873 23.2% 2.1m\n", + " 220 0.1637 23.2% 2.0m\n", + " 230 0.1609 23.2% 2.0m\n", + " 240 0.1648 23.2% 2.0m\n", + " 250 0.1418 23.2% 1.9m\n", + " 260 0.1560 23.2% 1.9m\n", + " 270 0.1414 23.2% 1.9m\n", + " 280 0.1904 23.2% 1.8m\n", + " 290 0.1609 23.2% 1.8m\n", + " 300 0.1679 23.2% 1.8m\n", + " 310 0.1291 23.2% 1.8m\n", + " 320 0.1432 23.2% 1.7m\n", + " 330 0.1543 23.2% 1.7m\n", + " 340 0.1802 23.2% 1.7m\n", + " 350 0.1557 23.2% 1.6m\n", + " 360 0.1540 23.2% 1.6m\n", + " 370 0.1521 23.2% 1.6m\n", + " 380 0.1157 23.2% 1.5m\n", + " 390 0.1699 23.2% 1.5m\n", + " 400 0.1638 23.2% 1.5m\n", + " 410 0.1587 23.2% 1.4m\n", + " 420 0.1708 23.2% 1.4m\n", + " 430 0.1910 23.2% 1.4m\n", + " 440 0.1449 23.2% 1.3m\n", + " 450 0.2045 23.2% 1.3m\n", + " 460 0.1468 23.2% 1.3m\n", + " 470 0.1925 23.2% 1.2m\n", + " 480 0.1371 23.2% 1.2m\n", + " 490 0.1730 23.2% 1.2m\n", + " 500 0.1301 23.2% 1.2m\n", + " 510 0.1354 23.2% 1.1m\n", + " 520 0.1977 23.2% 1.1m\n", + " 530 0.1615 23.2% 1.1m\n", + " 540 0.1602 23.2% 1.0m\n", + " 550 0.2030 23.2% 1.0m\n", + " 560 0.1764 23.2% 1.0m\n", + " 570 0.1476 23.2% 0.9m\n", + " 580 0.1984 23.2% 0.9m\n", + " 590 0.1868 23.2% 0.9m\n", + " 600 0.1300 23.2% 0.8m\n", + " 610 0.1410 23.2% 0.8m\n", + " 620 0.1621 23.2% 0.8m\n", + " 630 0.1418 23.2% 0.7m\n", + " 640 0.1690 23.2% 0.7m\n", + " 650 0.2315 23.2% 0.7m\n", + " 660 0.1836 23.2% 0.6m\n", + " 670 0.1427 23.2% 0.6m\n", + " 680 0.1996 23.2% 0.6m\n", + " 690 0.1454 23.2% 0.5m\n", + " 700 0.1329 23.2% 0.5m\n", + " 710 0.1811 23.2% 0.5m\n", + " 720 0.1289 23.2% 0.5m\n", + " 730 0.1799 23.2% 0.4m\n", + " 740 0.1399 23.2% 0.4m\n", + " 750 0.1834 23.2% 0.4m\n", + " 760 0.1738 23.2% 0.3m\n", + " 770 0.1585 23.2% 0.3m\n", + " 780 0.1229 23.2% 0.3m\n", + " 790 0.1784 23.2% 0.2m\n", + " 800 0.1602 23.2% 0.2m\n", + " 810 0.1566 23.2% 0.2m\n", + " 820 0.1672 23.2% 0.1m\n", + " 830 0.1447 23.2% 0.1m\n", + " 840 0.1848 23.2% 0.1m\n", + " 850 0.1496 23.2% 0.0m\n", + " 860 0.1428 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 87 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 87 SUMMARY\n", + " โฑ๏ธ Time: 172s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16653 โ†’ Val=0.16399\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 37m\n", + "\n", + "๐Ÿ”„ Epoch 88/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1544 23.2% 2.6m\n", + " 10 0.1453 23.2% 2.7m\n", + " 20 0.1213 23.2% 2.6m\n", + " 30 0.1431 23.2% 2.6m\n", + " 40 0.1955 23.2% 2.6m\n", + " 50 0.1517 23.2% 2.5m\n", + " 60 0.1542 23.2% 2.5m\n", + " 70 0.1471 23.2% 2.5m\n", + " 80 0.1829 23.2% 2.5m\n", + " 90 0.1514 23.2% 2.4m\n", + " 100 0.1366 23.2% 2.4m\n", + " 110 0.1729 23.2% 2.4m\n", + " 120 0.1312 23.2% 2.3m\n", + " 130 0.1585 23.2% 2.3m\n", + " 140 0.2054 23.2% 2.3m\n", + " 150 0.1457 23.2% 2.2m\n", + " 160 0.1342 23.2% 2.2m\n", + " 170 0.1759 23.2% 2.2m\n", + " 180 0.1454 23.2% 2.2m\n", + " 190 0.2016 23.2% 2.1m\n", + " 200 0.2042 23.2% 2.1m\n", + " 210 0.1907 23.2% 2.1m\n", + " 220 0.1784 23.2% 2.0m\n", + " 230 0.1312 23.2% 2.0m\n", + " 240 0.1474 23.2% 2.0m\n", + " 250 0.1959 23.2% 1.9m\n", + " 260 0.1371 23.2% 1.9m\n", + " 270 0.1488 23.2% 1.9m\n", + " 280 0.1865 23.2% 1.8m\n", + " 290 0.1246 23.2% 1.8m\n", + " 300 0.1801 23.2% 1.8m\n", + " 310 0.1828 23.2% 1.7m\n", + " 320 0.1632 23.2% 1.7m\n", + " 330 0.1749 23.2% 1.7m\n", + " 340 0.1676 23.2% 1.7m\n", + " 350 0.1558 23.2% 1.6m\n", + " 360 0.1607 23.2% 1.6m\n", + " 370 0.1902 23.2% 1.6m\n", + " 380 0.1307 23.2% 1.5m\n", + " 390 0.1373 23.2% 1.5m\n", + " 400 0.2051 23.2% 1.5m\n", + " 410 0.1494 23.2% 1.4m\n", + " 420 0.1339 23.2% 1.4m\n", + " 430 0.1838 23.2% 1.4m\n", + " 440 0.1611 23.2% 1.3m\n", + " 450 0.1524 23.2% 1.3m\n", + " 460 0.1721 23.2% 1.3m\n", + " 470 0.1462 23.2% 1.2m\n", + " 480 0.1800 23.2% 1.2m\n", + " 490 0.1628 23.2% 1.2m\n", + " 500 0.1566 23.2% 1.1m\n", + " 510 0.1719 23.2% 1.1m\n", + " 520 0.1837 23.2% 1.1m\n", + " 530 0.2195 23.2% 1.1m\n", + " 540 0.1796 23.2% 1.0m\n", + " 550 0.1903 23.2% 1.0m\n", + " 560 0.1493 23.2% 1.0m\n", + " 570 0.1900 23.2% 0.9m\n", + " 580 0.1985 23.2% 0.9m\n", + " 590 0.1465 23.2% 0.9m\n", + " 600 0.1748 23.2% 0.8m\n", + " 610 0.1522 23.2% 0.8m\n", + " 620 0.1711 23.2% 0.8m\n", + " 630 0.1501 23.2% 0.7m\n", + " 640 0.1654 23.2% 0.7m\n", + " 650 0.1681 23.2% 0.7m\n", + " 660 0.1623 23.2% 0.6m\n", + " 670 0.2033 23.2% 0.6m\n", + " 680 0.1691 23.2% 0.6m\n", + " 690 0.1440 23.2% 0.5m\n", + " 700 0.1656 23.2% 0.5m\n", + " 710 0.1529 23.2% 0.5m\n", + " 720 0.1778 23.2% 0.5m\n", + " 730 0.1479 23.2% 0.4m\n", + " 740 0.1430 23.2% 0.4m\n", + " 750 0.1822 23.2% 0.4m\n", + " 760 0.1601 23.2% 0.3m\n", + " 770 0.1762 23.2% 0.3m\n", + " 780 0.1607 23.2% 0.3m\n", + " 790 0.1504 23.2% 0.2m\n", + " 800 0.1763 23.2% 0.2m\n", + " 810 0.1196 23.2% 0.2m\n", + " 820 0.1805 23.2% 0.1m\n", + " 830 0.1379 23.2% 0.1m\n", + " 840 0.1972 23.2% 0.1m\n", + " 850 0.1954 23.2% 0.0m\n", + " 860 0.1727 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 88 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 88 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16646 โ†’ Val=0.16370\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โญ NEW BEST MODEL SAVED (val_loss=0.16370)\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 34m\n", + "\n", + "๐Ÿ”„ Epoch 89/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1834 23.2% 2.8m\n", + " 10 0.1171 23.2% 2.8m\n", + " 20 0.1369 23.2% 2.7m\n", + " 30 0.1613 23.2% 2.7m\n", + " 40 0.1763 23.2% 2.6m\n", + " 50 0.1230 23.2% 2.6m\n", + " 60 0.1414 23.2% 2.5m\n", + " 70 0.1559 23.2% 2.5m\n", + " 80 0.1581 23.2% 2.5m\n", + " 90 0.1476 23.2% 2.5m\n", + " 100 0.1711 23.2% 2.4m\n", + " 110 0.1903 23.2% 2.4m\n", + " 120 0.1857 23.2% 2.4m\n", + " 130 0.1541 23.2% 2.3m\n", + " 140 0.2083 23.2% 2.3m\n", + " 150 0.1628 23.2% 2.2m\n", + " 160 0.1817 23.2% 2.2m\n", + " 170 0.1629 23.2% 2.2m\n", + " 180 0.1748 23.2% 2.2m\n", + " 190 0.1442 23.2% 2.1m\n", + " 200 0.1880 23.2% 2.1m\n", + " 210 0.1461 23.2% 2.1m\n", + " 220 0.1732 23.2% 2.0m\n", + " 230 0.1505 23.2% 2.0m\n", + " 240 0.1728 23.2% 2.0m\n", + " 250 0.1475 23.2% 1.9m\n", + " 260 0.1724 23.2% 1.9m\n", + " 270 0.1613 23.2% 1.9m\n", + " 280 0.2107 23.2% 1.8m\n", + " 290 0.1632 23.2% 1.8m\n", + " 300 0.1709 23.2% 1.8m\n", + " 310 0.1513 23.2% 1.7m\n", + " 320 0.1901 23.2% 1.7m\n", + " 330 0.1709 23.2% 1.7m\n", + " 340 0.1764 23.2% 1.6m\n", + " 350 0.2090 23.2% 1.6m\n", + " 360 0.2090 23.2% 1.6m\n", + " 370 0.1598 23.2% 1.5m\n", + " 380 0.1658 23.2% 1.5m\n", + " 390 0.1496 23.2% 1.5m\n", + " 400 0.2036 23.2% 1.5m\n", + " 410 0.1897 23.2% 1.4m\n", + " 420 0.1558 23.2% 1.4m\n", + " 430 0.1696 23.2% 1.4m\n", + " 440 0.1900 23.2% 1.3m\n", + " 450 0.1459 23.2% 1.3m\n", + " 460 0.1600 23.2% 1.3m\n", + " 470 0.1852 23.2% 1.2m\n", + " 480 0.1681 23.2% 1.2m\n", + " 490 0.1869 23.2% 1.2m\n", + " 500 0.1465 23.2% 1.1m\n", + " 510 0.1694 23.2% 1.1m\n", + " 520 0.2163 23.2% 1.1m\n", + " 530 0.1687 23.2% 1.0m\n", + " 540 0.1404 23.2% 1.0m\n", + " 550 0.1536 23.2% 1.0m\n", + " 560 0.1329 23.2% 1.0m\n", + " 570 0.1426 23.2% 0.9m\n", + " 580 0.1868 23.2% 0.9m\n", + " 590 0.1891 23.2% 0.9m\n", + " 600 0.1739 23.2% 0.8m\n", + " 610 0.1921 23.2% 0.8m\n", + " 620 0.2020 23.2% 0.8m\n", + " 630 0.1770 23.2% 0.7m\n", + " 640 0.1126 23.2% 0.7m\n", + " 650 0.1693 23.2% 0.7m\n", + " 660 0.1554 23.2% 0.6m\n", + " 670 0.2023 23.2% 0.6m\n", + " 680 0.1390 23.2% 0.6m\n", + " 690 0.1759 23.2% 0.5m\n", + " 700 0.1561 23.2% 0.5m\n", + " 710 0.1651 23.2% 0.5m\n", + " 720 0.1657 23.2% 0.4m\n", + " 730 0.1946 23.2% 0.4m\n", + " 740 0.1295 23.2% 0.4m\n", + " 750 0.1974 23.2% 0.4m\n", + " 760 0.1460 23.2% 0.3m\n", + " 770 0.1341 23.2% 0.3m\n", + " 780 0.1804 23.2% 0.3m\n", + " 790 0.1626 23.2% 0.2m\n", + " 800 0.1673 23.2% 0.2m\n", + " 810 0.1836 23.2% 0.2m\n", + " 820 0.1163 23.2% 0.1m\n", + " 830 0.1952 23.2% 0.1m\n", + " 840 0.1726 23.2% 0.1m\n", + " 850 0.1467 23.2% 0.0m\n", + " 860 0.2200 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 89 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 89 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16645 โ†’ Val=0.16424\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 31m\n", + "\n", + "๐Ÿ”„ Epoch 90/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1484 23.2% 2.6m\n", + " 10 0.1744 23.2% 2.7m\n", + " 20 0.1624 23.2% 2.7m\n", + " 30 0.1702 23.2% 2.6m\n", + " 40 0.1564 23.2% 2.6m\n", + " 50 0.1836 23.2% 2.6m\n", + " 60 0.1910 23.2% 2.5m\n", + " 70 0.1694 23.2% 2.5m\n", + " 80 0.1919 23.2% 2.5m\n", + " 90 0.1636 23.2% 2.4m\n", + " 100 0.1490 23.2% 2.4m\n", + " 110 0.1568 23.2% 2.4m\n", + " 120 0.1609 23.2% 2.3m\n", + " 130 0.1911 23.2% 2.3m\n", + " 140 0.1757 23.2% 2.3m\n", + " 150 0.2042 23.2% 2.2m\n", + " 160 0.1358 23.2% 2.2m\n", + " 170 0.1810 23.2% 2.2m\n", + " 180 0.1839 23.2% 2.2m\n", + " 190 0.1180 23.2% 2.1m\n", + " 200 0.1796 23.2% 2.1m\n", + " 210 0.1443 23.2% 2.1m\n", + " 220 0.1526 23.2% 2.0m\n", + " 230 0.1807 23.2% 2.0m\n", + " 240 0.1336 23.2% 2.0m\n", + " 250 0.1760 23.2% 1.9m\n", + " 260 0.1494 23.2% 1.9m\n", + " 270 0.1736 23.2% 1.9m\n", + " 280 0.1422 23.2% 1.8m\n", + " 290 0.1644 23.2% 1.8m\n", + " 300 0.1737 23.2% 1.8m\n", + " 310 0.1783 23.2% 1.7m\n", + " 320 0.1320 23.2% 1.7m\n", + " 330 0.1277 23.2% 1.7m\n", + " 340 0.1569 23.2% 1.6m\n", + " 350 0.2258 23.2% 1.6m\n", + " 360 0.1257 23.2% 1.6m\n", + " 370 0.1760 23.2% 1.6m\n", + " 380 0.1673 23.2% 1.5m\n", + " 390 0.1771 23.2% 1.5m\n", + " 400 0.1647 23.2% 1.5m\n", + " 410 0.1699 23.2% 1.4m\n", + " 420 0.1839 23.2% 1.4m\n", + " 430 0.1643 23.2% 1.4m\n", + " 440 0.1931 23.2% 1.3m\n", + " 450 0.1667 23.2% 1.3m\n", + " 460 0.2075 23.2% 1.3m\n", + " 470 0.1510 23.2% 1.2m\n", + " 480 0.1579 23.2% 1.2m\n", + " 490 0.1495 23.2% 1.2m\n", + " 500 0.1452 23.2% 1.1m\n", + " 510 0.1546 23.2% 1.1m\n", + " 520 0.2026 23.2% 1.1m\n", + " 530 0.1812 23.2% 1.0m\n", + " 540 0.1586 23.2% 1.0m\n", + " 550 0.1488 23.2% 1.0m\n", + " 560 0.1794 23.2% 1.0m\n", + " 570 0.1878 23.2% 0.9m\n", + " 580 0.1940 23.2% 0.9m\n", + " 590 0.1374 23.2% 0.9m\n", + " 600 0.1805 23.2% 0.8m\n", + " 610 0.1545 23.2% 0.8m\n", + " 620 0.1457 23.2% 0.8m\n", + " 630 0.1685 23.2% 0.7m\n", + " 640 0.1230 23.2% 0.7m\n", + " 650 0.1693 23.2% 0.7m\n", + " 660 0.1542 23.2% 0.6m\n", + " 670 0.1477 23.2% 0.6m\n", + " 680 0.1508 23.2% 0.6m\n", + " 690 0.1601 23.2% 0.5m\n", + " 700 0.1837 23.2% 0.5m\n", + " 710 0.1743 23.2% 0.5m\n", + " 720 0.1974 23.2% 0.5m\n", + " 730 0.1489 23.2% 0.4m\n", + " 740 0.1581 23.2% 0.4m\n", + " 750 0.1955 23.2% 0.4m\n", + " 760 0.1489 23.2% 0.3m\n", + " 770 0.1660 23.2% 0.3m\n", + " 780 0.1765 23.2% 0.3m\n", + " 790 0.1523 23.2% 0.2m\n", + " 800 0.1554 23.2% 0.2m\n", + " 810 0.1705 23.2% 0.2m\n", + " 820 0.1689 23.2% 0.1m\n", + " 830 0.1546 23.2% 0.1m\n", + " 840 0.1302 23.2% 0.1m\n", + " 850 0.1714 23.2% 0.0m\n", + " 860 0.1857 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 90 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 90 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16634 โ†’ Val=0.16382\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 29m\n", + "\n", + "๐Ÿ”„ Epoch 91/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1624 23.2% 2.7m\n", + " 10 0.1689 23.2% 2.7m\n", + " 20 0.1982 23.2% 2.7m\n", + " 30 0.1617 23.2% 2.6m\n", + " 40 0.1275 23.2% 2.6m\n", + " 50 0.1477 23.2% 2.5m\n", + " 60 0.1822 23.2% 2.5m\n", + " 70 0.1446 23.2% 2.5m\n", + " 80 0.1724 23.2% 2.4m\n", + " 90 0.2023 23.2% 2.4m\n", + " 100 0.1926 23.2% 2.4m\n", + " 110 0.1617 23.2% 2.4m\n", + " 120 0.1724 23.2% 2.3m\n", + " 130 0.1770 23.2% 2.3m\n", + " 140 0.1098 23.2% 2.3m\n", + " 150 0.1899 23.2% 2.2m\n", + " 160 0.1825 23.2% 2.2m\n", + " 170 0.1641 23.2% 2.2m\n", + " 180 0.1904 23.2% 2.1m\n", + " 190 0.2047 23.2% 2.1m\n", + " 200 0.1817 23.2% 2.1m\n", + " 210 0.1928 23.2% 2.1m\n", + " 220 0.1389 23.2% 2.0m\n", + " 230 0.1743 23.2% 2.0m\n", + " 240 0.1652 23.2% 2.0m\n", + " 250 0.1290 23.2% 1.9m\n", + " 260 0.1839 23.2% 1.9m\n", + " 270 0.1933 23.2% 1.9m\n", + " 280 0.1576 23.2% 1.8m\n", + " 290 0.1607 23.2% 1.8m\n", + " 300 0.1741 23.2% 1.8m\n", + " 310 0.1621 23.2% 1.7m\n", + " 320 0.1740 23.2% 1.7m\n", + " 330 0.1549 23.2% 1.7m\n", + " 340 0.2375 23.2% 1.6m\n", + " 350 0.1914 23.2% 1.6m\n", + " 360 0.1736 23.2% 1.6m\n", + " 370 0.1484 23.2% 1.5m\n", + " 380 0.1195 23.2% 1.5m\n", + " 390 0.1607 23.2% 1.5m\n", + " 400 0.1730 23.2% 1.5m\n", + " 410 0.1677 23.2% 1.4m\n", + " 420 0.1418 23.2% 1.4m\n", + " 430 0.1444 23.2% 1.4m\n", + " 440 0.1757 23.2% 1.3m\n", + " 450 0.1882 23.2% 1.3m\n", + " 460 0.1655 23.2% 1.3m\n", + " 470 0.2327 23.2% 1.2m\n", + " 480 0.1648 23.2% 1.2m\n", + " 490 0.1732 23.2% 1.2m\n", + " 500 0.1689 23.2% 1.1m\n", + " 510 0.1598 23.2% 1.1m\n", + " 520 0.1344 23.2% 1.1m\n", + " 530 0.1642 23.2% 1.0m\n", + " 540 0.1675 23.2% 1.0m\n", + " 550 0.1852 23.2% 1.0m\n", + " 560 0.1691 23.2% 1.0m\n", + " 570 0.1788 23.2% 0.9m\n", + " 580 0.1628 23.2% 0.9m\n", + " 590 0.1295 23.2% 0.9m\n", + " 600 0.1509 23.2% 0.8m\n", + " 610 0.1700 23.2% 0.8m\n", + " 620 0.1482 23.2% 0.8m\n", + " 630 0.1549 23.2% 0.7m\n", + " 640 0.1775 23.2% 0.7m\n", + " 650 0.1676 23.2% 0.7m\n", + " 660 0.1548 23.2% 0.6m\n", + " 670 0.1674 23.2% 0.6m\n", + " 680 0.1476 23.2% 0.6m\n", + " 690 0.1693 23.2% 0.5m\n", + " 700 0.1521 23.2% 0.5m\n", + " 710 0.1259 23.2% 0.5m\n", + " 720 0.1593 23.2% 0.4m\n", + " 730 0.1841 23.2% 0.4m\n", + " 740 0.1956 23.2% 0.4m\n", + " 750 0.1522 23.2% 0.4m\n", + " 760 0.1685 23.2% 0.3m\n", + " 770 0.1622 23.2% 0.3m\n", + " 780 0.2277 23.2% 0.3m\n", + " 790 0.1002 23.2% 0.2m\n", + " 800 0.1703 23.2% 0.2m\n", + " 810 0.1654 23.2% 0.2m\n", + " 820 0.1453 23.2% 0.1m\n", + " 830 0.1740 23.2% 0.1m\n", + " 840 0.1627 23.2% 0.1m\n", + " 850 0.1272 23.2% 0.0m\n", + " 860 0.1358 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 91 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 91 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16643 โ†’ Val=0.16393\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 26m\n", + "\n", + "๐Ÿ”„ Epoch 92/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1582 23.2% 2.6m\n", + " 10 0.1798 23.2% 2.6m\n", + " 20 0.1859 23.2% 2.6m\n", + " 30 0.1803 23.2% 2.6m\n", + " 40 0.1780 23.2% 2.6m\n", + " 50 0.1305 23.2% 2.6m\n", + " 60 0.2080 23.2% 2.5m\n", + " 70 0.1541 23.2% 2.5m\n", + " 80 0.1937 23.2% 2.5m\n", + " 90 0.1630 23.2% 2.4m\n", + " 100 0.1729 23.2% 2.4m\n", + " 110 0.1681 23.2% 2.4m\n", + " 120 0.1542 23.2% 2.3m\n", + " 130 0.1763 23.2% 2.3m\n", + " 140 0.1306 23.2% 2.3m\n", + " 150 0.1607 23.2% 2.2m\n", + " 160 0.2150 23.2% 2.2m\n", + " 170 0.1650 23.2% 2.2m\n", + " 180 0.1599 23.2% 2.1m\n", + " 190 0.1770 23.2% 2.1m\n", + " 200 0.0967 23.2% 2.1m\n", + " 210 0.1726 23.2% 2.1m\n", + " 220 0.1709 23.2% 2.0m\n", + " 230 0.1405 23.2% 2.0m\n", + " 240 0.2076 23.2% 2.0m\n", + " 250 0.2011 23.2% 1.9m\n", + " 260 0.1507 23.2% 1.9m\n", + " 270 0.1788 23.2% 1.9m\n", + " 280 0.1593 23.2% 1.8m\n", + " 290 0.1656 23.2% 1.8m\n", + " 300 0.1539 23.2% 1.8m\n", + " 310 0.1708 23.2% 1.7m\n", + " 320 0.1245 23.2% 1.7m\n", + " 330 0.1584 23.2% 1.7m\n", + " 340 0.1386 23.2% 1.6m\n", + " 350 0.1713 23.2% 1.6m\n", + " 360 0.1513 23.2% 1.6m\n", + " 370 0.1767 23.2% 1.5m\n", + " 380 0.1875 23.2% 1.5m\n", + " 390 0.1949 23.2% 1.5m\n", + " 400 0.1667 23.2% 1.5m\n", + " 410 0.1572 23.2% 1.4m\n", + " 420 0.1750 23.2% 1.4m\n", + " 430 0.1964 23.2% 1.4m\n", + " 440 0.1799 23.2% 1.3m\n", + " 450 0.2112 23.2% 1.3m\n", + " 460 0.1778 23.2% 1.3m\n", + " 470 0.2012 23.2% 1.2m\n", + " 480 0.1732 23.2% 1.2m\n", + " 490 0.1646 23.2% 1.2m\n", + " 500 0.1993 23.2% 1.1m\n", + " 510 0.1492 23.2% 1.1m\n", + " 520 0.1926 23.2% 1.1m\n", + " 530 0.1698 23.2% 1.0m\n", + " 540 0.1470 23.2% 1.0m\n", + " 550 0.1642 23.2% 1.0m\n", + " 560 0.1747 23.2% 0.9m\n", + " 570 0.1645 23.2% 0.9m\n", + " 580 0.1821 23.2% 0.9m\n", + " 590 0.1349 23.2% 0.9m\n", + " 600 0.1545 23.2% 0.8m\n", + " 610 0.1338 23.2% 0.8m\n", + " 620 0.1720 23.2% 0.8m\n", + " 630 0.1865 23.2% 0.7m\n", + " 640 0.1387 23.2% 0.7m\n", + " 650 0.1559 23.2% 0.7m\n", + " 660 0.1711 23.2% 0.6m\n", + " 670 0.1177 23.2% 0.6m\n", + " 680 0.1842 23.2% 0.6m\n", + " 690 0.1472 23.2% 0.5m\n", + " 700 0.1900 23.2% 0.5m\n", + " 710 0.1893 23.2% 0.5m\n", + " 720 0.1365 23.2% 0.4m\n", + " 730 0.1603 23.2% 0.4m\n", + " 740 0.1692 23.2% 0.4m\n", + " 750 0.2126 23.2% 0.4m\n", + " 760 0.1801 23.2% 0.3m\n", + " 770 0.1733 23.2% 0.3m\n", + " 780 0.1534 23.2% 0.3m\n", + " 790 0.1247 23.2% 0.2m\n", + " 800 0.1417 23.2% 0.2m\n", + " 810 0.1823 23.2% 0.2m\n", + " 820 0.1403 23.2% 0.1m\n", + " 830 0.1583 23.2% 0.1m\n", + " 840 0.1629 23.2% 0.1m\n", + " 850 0.1409 23.2% 0.0m\n", + " 860 0.1573 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 92 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 92 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16639 โ†’ Val=0.16397\n", + " ๐Ÿ“Š Acc: EC=81.5% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 23m\n", + "\n", + "๐Ÿ”„ Epoch 93/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1432 23.2% 2.7m\n", + " 10 0.2085 23.2% 2.7m\n", + " 20 0.1870 23.2% 2.6m\n", + " 30 0.2004 23.2% 2.6m\n", + " 40 0.1500 23.2% 2.6m\n", + " 50 0.1561 23.2% 2.6m\n", + " 60 0.1011 23.2% 2.5m\n", + " 70 0.1519 23.2% 2.5m\n", + " 80 0.1618 23.2% 2.5m\n", + " 90 0.1910 23.2% 2.4m\n", + " 100 0.2227 23.2% 2.4m\n", + " 110 0.1424 23.2% 2.4m\n", + " 120 0.1418 23.2% 2.3m\n", + " 130 0.1342 23.2% 2.3m\n", + " 140 0.1495 23.2% 2.3m\n", + " 150 0.1726 23.2% 2.2m\n", + " 160 0.1554 23.2% 2.2m\n", + " 170 0.1747 23.2% 2.2m\n", + " 180 0.1288 23.2% 2.2m\n", + " 190 0.1854 23.2% 2.1m\n", + " 200 0.1852 23.2% 2.1m\n", + " 210 0.1457 23.2% 2.1m\n", + " 220 0.1471 23.2% 2.0m\n", + " 230 0.1635 23.2% 2.0m\n", + " 240 0.1405 23.2% 2.0m\n", + " 250 0.1778 23.2% 1.9m\n", + " 260 0.1652 23.2% 1.9m\n", + " 270 0.1728 23.2% 1.9m\n", + " 280 0.1459 23.2% 1.8m\n", + " 290 0.1801 23.2% 1.8m\n", + " 300 0.1985 23.2% 1.8m\n", + " 310 0.1641 23.2% 1.7m\n", + " 320 0.1504 23.2% 1.7m\n", + " 330 0.1730 23.2% 1.7m\n", + " 340 0.1674 23.2% 1.6m\n", + " 350 0.1475 23.2% 1.6m\n", + " 360 0.1241 23.2% 1.6m\n", + " 370 0.1375 23.2% 1.5m\n", + " 380 0.1922 23.2% 1.5m\n", + " 390 0.1709 23.2% 1.5m\n", + " 400 0.1846 23.2% 1.5m\n", + " 410 0.1252 23.2% 1.4m\n", + " 420 0.1814 23.2% 1.4m\n", + " 430 0.1802 23.2% 1.4m\n", + " 440 0.1786 23.2% 1.3m\n", + " 450 0.1484 23.2% 1.3m\n", + " 460 0.1842 23.2% 1.3m\n", + " 470 0.2081 23.2% 1.2m\n", + " 480 0.1575 23.2% 1.2m\n", + " 490 0.1740 23.2% 1.2m\n", + " 500 0.1536 23.2% 1.1m\n", + " 510 0.1953 23.2% 1.1m\n", + " 520 0.1216 23.2% 1.1m\n", + " 530 0.1934 23.2% 1.0m\n", + " 540 0.1883 23.2% 1.0m\n", + " 550 0.1905 23.2% 1.0m\n", + " 560 0.1532 23.2% 1.0m\n", + " 570 0.1538 23.2% 0.9m\n", + " 580 0.1146 23.2% 0.9m\n", + " 590 0.1390 23.2% 0.9m\n", + " 600 0.1382 23.2% 0.8m\n", + " 610 0.1731 23.2% 0.8m\n", + " 620 0.1830 23.2% 0.8m\n", + " 630 0.1892 23.2% 0.7m\n", + " 640 0.1862 23.2% 0.7m\n", + " 650 0.2000 23.2% 0.7m\n", + " 660 0.2005 23.2% 0.6m\n", + " 670 0.1828 23.2% 0.6m\n", + " 680 0.1939 23.2% 0.6m\n", + " 690 0.1553 23.2% 0.5m\n", + " 700 0.1816 23.2% 0.5m\n", + " 710 0.1883 23.2% 0.5m\n", + " 720 0.1787 23.2% 0.4m\n", + " 730 0.1803 23.2% 0.4m\n", + " 740 0.1841 23.2% 0.4m\n", + " 750 0.1143 23.2% 0.4m\n", + " 760 0.1854 23.2% 0.3m\n", + " 770 0.1590 23.2% 0.3m\n", + " 780 0.1947 23.2% 0.3m\n", + " 790 0.1366 23.2% 0.2m\n", + " 800 0.1617 23.2% 0.2m\n", + " 810 0.1845 23.2% 0.2m\n", + " 820 0.1405 23.2% 0.1m\n", + " 830 0.1957 23.2% 0.1m\n", + " 840 0.1693 23.2% 0.1m\n", + " 850 0.1803 23.2% 0.0m\n", + " 860 0.1759 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 93 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 93 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16643 โ†’ Val=0.16386\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โš ๏ธ No improvement for 5/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 20m\n", + "\n", + "๐Ÿ”„ Epoch 94/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1271 23.2% 2.7m\n", + " 10 0.1561 23.2% 2.7m\n", + " 20 0.1518 23.2% 2.6m\n", + " 30 0.1786 23.2% 2.6m\n", + " 40 0.1765 23.2% 2.6m\n", + " 50 0.1521 23.2% 2.5m\n", + " 60 0.2019 23.2% 2.5m\n", + " 70 0.1572 23.2% 2.5m\n", + " 80 0.1315 23.2% 2.5m\n", + " 90 0.1633 23.2% 2.4m\n", + " 100 0.1552 23.2% 2.4m\n", + " 110 0.1720 23.2% 2.4m\n", + " 120 0.1982 23.2% 2.3m\n", + " 130 0.1759 23.2% 2.3m\n", + " 140 0.1416 23.2% 2.3m\n", + " 150 0.1521 23.2% 2.2m\n", + " 160 0.1742 23.2% 2.2m\n", + " 170 0.1495 23.2% 2.2m\n", + " 180 0.1729 23.2% 2.1m\n", + " 190 0.2122 23.2% 2.1m\n", + " 200 0.1784 23.2% 2.1m\n", + " 210 0.1983 23.2% 2.0m\n", + " 220 0.1683 23.2% 2.0m\n", + " 230 0.1511 23.2% 2.0m\n", + " 240 0.1664 23.2% 2.0m\n", + " 250 0.1323 23.2% 1.9m\n", + " 260 0.1859 23.2% 1.9m\n", + " 270 0.1233 23.2% 1.9m\n", + " 280 0.1440 23.2% 1.8m\n", + " 290 0.1832 23.2% 1.8m\n", + " 300 0.1623 23.2% 1.8m\n", + " 310 0.1819 23.2% 1.7m\n", + " 320 0.1775 23.2% 1.7m\n", + " 330 0.1524 23.2% 1.7m\n", + " 340 0.1851 23.2% 1.6m\n", + " 350 0.1573 23.2% 1.6m\n", + " 360 0.2056 23.2% 1.6m\n", + " 370 0.1487 23.2% 1.6m\n", + " 380 0.1903 23.2% 1.5m\n", + " 390 0.1682 23.2% 1.5m\n", + " 400 0.1594 23.2% 1.5m\n", + " 410 0.1736 23.2% 1.4m\n", + " 420 0.1379 23.2% 1.4m\n", + " 430 0.1284 23.2% 1.4m\n", + " 440 0.1627 23.2% 1.3m\n", + " 450 0.1611 23.2% 1.3m\n", + " 460 0.1521 23.2% 1.3m\n", + " 470 0.1598 23.2% 1.2m\n", + " 480 0.1900 23.2% 1.2m\n", + " 490 0.1730 23.2% 1.2m\n", + " 500 0.1358 23.2% 1.1m\n", + " 510 0.2020 23.2% 1.1m\n", + " 520 0.2029 23.2% 1.1m\n", + " 530 0.1685 23.2% 1.0m\n", + " 540 0.1588 23.2% 1.0m\n", + " 550 0.1401 23.2% 1.0m\n", + " 560 0.1637 23.2% 1.0m\n", + " 570 0.1293 23.2% 0.9m\n", + " 580 0.2091 23.2% 0.9m\n", + " 590 0.1831 23.2% 0.9m\n", + " 600 0.1800 23.2% 0.8m\n", + " 610 0.2062 23.2% 0.8m\n", + " 620 0.1656 23.2% 0.8m\n", + " 630 0.1336 23.2% 0.7m\n", + " 640 0.1581 23.2% 0.7m\n", + " 650 0.1816 23.2% 0.7m\n", + " 660 0.1663 23.2% 0.6m\n", + " 670 0.1911 23.2% 0.6m\n", + " 680 0.1945 23.2% 0.6m\n", + " 690 0.1774 23.2% 0.5m\n", + " 700 0.1689 23.2% 0.5m\n", + " 710 0.2115 23.2% 0.5m\n", + " 720 0.1439 23.2% 0.4m\n", + " 730 0.1500 23.2% 0.4m\n", + " 740 0.1703 23.2% 0.4m\n", + " 750 0.1355 23.2% 0.4m\n", + " 760 0.1738 23.2% 0.3m\n", + " 770 0.1859 23.2% 0.3m\n", + " 780 0.1782 23.2% 0.3m\n", + " 790 0.1372 23.2% 0.2m\n", + " 800 0.1704 23.2% 0.2m\n", + " 810 0.1754 23.2% 0.2m\n", + " 820 0.1996 23.2% 0.1m\n", + " 830 0.1722 23.2% 0.1m\n", + " 840 0.1713 23.2% 0.1m\n", + " 850 0.1587 23.2% 0.0m\n", + " 860 0.1688 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 94 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 94 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16642 โ†’ Val=0.16403\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 17m\n", + "\n", + "๐Ÿ”„ Epoch 95/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1635 23.2% 2.7m\n", + " 10 0.1516 23.2% 2.6m\n", + " 20 0.1513 23.2% 2.7m\n", + " 30 0.1838 23.2% 2.6m\n", + " 40 0.1996 23.2% 2.6m\n", + " 50 0.1959 23.2% 2.6m\n", + " 60 0.2000 23.2% 2.5m\n", + " 70 0.1558 23.2% 2.5m\n", + " 80 0.1952 23.2% 2.5m\n", + " 90 0.1972 23.2% 2.4m\n", + " 100 0.1964 23.2% 2.4m\n", + " 110 0.1584 23.2% 2.4m\n", + " 120 0.1979 23.2% 2.3m\n", + " 130 0.1485 23.2% 2.3m\n", + " 140 0.1834 23.2% 2.3m\n", + " 150 0.1809 23.2% 2.2m\n", + " 160 0.1365 23.2% 2.2m\n", + " 170 0.2236 23.2% 2.2m\n", + " 180 0.1695 23.2% 2.1m\n", + " 190 0.1895 23.2% 2.1m\n", + " 200 0.1481 23.2% 2.1m\n", + " 210 0.1749 23.2% 2.1m\n", + " 220 0.1937 23.2% 2.0m\n", + " 230 0.1590 23.2% 2.0m\n", + " 240 0.1902 23.2% 2.0m\n", + " 250 0.1448 23.2% 1.9m\n", + " 260 0.1692 23.2% 1.9m\n", + " 270 0.1263 23.2% 1.9m\n", + " 280 0.1450 23.2% 1.8m\n", + " 290 0.1597 23.2% 1.8m\n", + " 300 0.1957 23.2% 1.8m\n", + " 310 0.1765 23.2% 1.7m\n", + " 320 0.2218 23.2% 1.7m\n", + " 330 0.1665 23.2% 1.7m\n", + " 340 0.1617 23.2% 1.6m\n", + " 350 0.1437 23.2% 1.6m\n", + " 360 0.1690 23.2% 1.6m\n", + " 370 0.1760 23.2% 1.6m\n", + " 380 0.1491 23.2% 1.5m\n", + " 390 0.1415 23.2% 1.5m\n", + " 400 0.1841 23.2% 1.5m\n", + " 410 0.1824 23.2% 1.4m\n", + " 420 0.1403 23.2% 1.4m\n", + " 430 0.1825 23.2% 1.4m\n", + " 440 0.1213 23.2% 1.3m\n", + " 450 0.1508 23.2% 1.3m\n", + " 460 0.1567 23.2% 1.3m\n", + " 470 0.1325 23.2% 1.2m\n", + " 480 0.1805 23.2% 1.2m\n", + " 490 0.1282 23.2% 1.2m\n", + " 500 0.1940 23.2% 1.1m\n", + " 510 0.1706 23.2% 1.1m\n", + " 520 0.1539 23.2% 1.1m\n", + " 530 0.1924 23.2% 1.0m\n", + " 540 0.1596 23.2% 1.0m\n", + " 550 0.1378 23.2% 1.0m\n", + " 560 0.1674 23.2% 1.0m\n", + " 570 0.1386 23.2% 0.9m\n", + " 580 0.1411 23.2% 0.9m\n", + " 590 0.1269 23.2% 0.9m\n", + " 600 0.2017 23.2% 0.8m\n", + " 610 0.1846 23.2% 0.8m\n", + " 620 0.1646 23.2% 0.8m\n", + " 630 0.1371 23.2% 0.7m\n", + " 640 0.1270 23.2% 0.7m\n", + " 650 0.1447 23.2% 0.7m\n", + " 660 0.1459 23.2% 0.6m\n", + " 670 0.1771 23.2% 0.6m\n", + " 680 0.1958 23.2% 0.6m\n", + " 690 0.1706 23.2% 0.5m\n", + " 700 0.1783 23.2% 0.5m\n", + " 710 0.1602 23.2% 0.5m\n", + " 720 0.1428 23.2% 0.5m\n", + " 730 0.1272 23.2% 0.4m\n", + " 740 0.1866 23.2% 0.4m\n", + " 750 0.2257 23.2% 0.4m\n", + " 760 0.1987 23.2% 0.3m\n", + " 770 0.1421 23.2% 0.3m\n", + " 780 0.1392 23.2% 0.3m\n", + " 790 0.1416 23.2% 0.2m\n", + " 800 0.1762 23.2% 0.2m\n", + " 810 0.1520 23.2% 0.2m\n", + " 820 0.2048 23.2% 0.1m\n", + " 830 0.1829 23.2% 0.1m\n", + " 840 0.1423 23.2% 0.1m\n", + " 850 0.1807 23.2% 0.0m\n", + " 860 0.2003 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 95 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 95 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16650 โ†’ Val=0.16388\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.4% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 14m\n", + "\n", + "๐Ÿ”„ Epoch 96/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1438 23.2% 2.7m\n", + " 10 0.1455 23.2% 2.8m\n", + " 20 0.1749 23.2% 2.7m\n", + " 30 0.2083 23.2% 2.7m\n", + " 40 0.1603 23.2% 2.6m\n", + " 50 0.1399 23.2% 2.6m\n", + " 60 0.1778 23.2% 2.6m\n", + " 70 0.1538 23.2% 2.5m\n", + " 80 0.2118 23.2% 2.5m\n", + " 90 0.1396 23.2% 2.5m\n", + " 100 0.1585 23.2% 2.4m\n", + " 110 0.1674 23.2% 2.4m\n", + " 120 0.1620 23.2% 2.4m\n", + " 130 0.1670 23.2% 2.3m\n", + " 140 0.1774 23.2% 2.3m\n", + " 150 0.1605 23.2% 2.2m\n", + " 160 0.2029 23.2% 2.2m\n", + " 170 0.1221 23.2% 2.2m\n", + " 180 0.1487 23.2% 2.1m\n", + " 190 0.1869 23.2% 2.1m\n", + " 200 0.1600 23.2% 2.1m\n", + " 210 0.1617 23.2% 2.1m\n", + " 220 0.1745 23.2% 2.0m\n", + " 230 0.2005 23.2% 2.0m\n", + " 240 0.1590 23.2% 2.0m\n", + " 250 0.1661 23.2% 1.9m\n", + " 260 0.1884 23.2% 1.9m\n", + " 270 0.1086 23.2% 1.9m\n", + " 280 0.1462 23.2% 1.8m\n", + " 290 0.1428 23.2% 1.8m\n", + " 300 0.1638 23.2% 1.8m\n", + " 310 0.1937 23.2% 1.7m\n", + " 320 0.1969 23.2% 1.7m\n", + " 330 0.1573 23.2% 1.7m\n", + " 340 0.1646 23.2% 1.6m\n", + " 350 0.1612 23.2% 1.6m\n", + " 360 0.2004 23.2% 1.6m\n", + " 370 0.1217 23.2% 1.5m\n", + " 380 0.1873 23.2% 1.5m\n", + " 390 0.1811 23.2% 1.5m\n", + " 400 0.1457 23.2% 1.5m\n", + " 410 0.1847 23.2% 1.4m\n", + " 420 0.1700 23.2% 1.4m\n", + " 430 0.1528 23.2% 1.4m\n", + " 440 0.1389 23.2% 1.3m\n", + " 450 0.1592 23.2% 1.3m\n", + " 460 0.1681 23.2% 1.3m\n", + " 470 0.1424 23.2% 1.2m\n", + " 480 0.2193 23.2% 1.2m\n", + " 490 0.1334 23.2% 1.2m\n", + " 500 0.1585 23.2% 1.1m\n", + " 510 0.1607 23.2% 1.1m\n", + " 520 0.1534 23.2% 1.1m\n", + " 530 0.1605 23.2% 1.0m\n", + " 540 0.2289 23.2% 1.0m\n", + " 550 0.1843 23.2% 1.0m\n", + " 560 0.1935 23.2% 0.9m\n", + " 570 0.1798 23.2% 0.9m\n", + " 580 0.1924 23.2% 0.9m\n", + " 590 0.1685 23.2% 0.9m\n", + " 600 0.1480 23.2% 0.8m\n", + " 610 0.1555 23.2% 0.8m\n", + " 620 0.1798 23.2% 0.8m\n", + " 630 0.1884 23.2% 0.7m\n", + " 640 0.1738 23.2% 0.7m\n", + " 650 0.1587 23.2% 0.7m\n", + " 660 0.1844 23.2% 0.6m\n", + " 670 0.1955 23.2% 0.6m\n", + " 680 0.1683 23.2% 0.6m\n", + " 690 0.1457 23.2% 0.5m\n", + " 700 0.1498 23.2% 0.5m\n", + " 710 0.1455 23.2% 0.5m\n", + " 720 0.1448 23.2% 0.4m\n", + " 730 0.1848 23.2% 0.4m\n", + " 740 0.1875 23.2% 0.4m\n", + " 750 0.1932 23.2% 0.4m\n", + " 760 0.1657 23.2% 0.3m\n", + " 770 0.2021 23.2% 0.3m\n", + " 780 0.1802 23.2% 0.3m\n", + " 790 0.1460 23.2% 0.2m\n", + " 800 0.1683 23.2% 0.2m\n", + " 810 0.1632 23.2% 0.2m\n", + " 820 0.1377 23.2% 0.1m\n", + " 830 0.1811 23.2% 0.1m\n", + " 840 0.1794 23.2% 0.1m\n", + " 850 0.1801 23.2% 0.0m\n", + " 860 0.1628 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 96 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 96 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16646 โ†’ Val=0.16379\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 11m\n", + "\n", + "๐Ÿ”„ Epoch 97/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1647 23.2% 2.7m\n", + " 10 0.1905 23.2% 2.6m\n", + " 20 0.2009 23.2% 2.6m\n", + " 30 0.1909 23.2% 2.6m\n", + " 40 0.1880 23.2% 2.6m\n", + " 50 0.1614 23.2% 2.5m\n", + " 60 0.1261 23.2% 2.5m\n", + " 70 0.1235 23.2% 2.5m\n", + " 80 0.1391 23.2% 2.4m\n", + " 90 0.1507 23.2% 2.4m\n", + " 100 0.1659 23.2% 2.4m\n", + " 110 0.1889 23.2% 2.4m\n", + " 120 0.1652 23.2% 2.3m\n", + " 130 0.1825 23.2% 2.3m\n", + " 140 0.1372 23.2% 2.3m\n", + " 150 0.1249 23.2% 2.2m\n", + " 160 0.1861 23.2% 2.2m\n", + " 170 0.2181 23.2% 2.2m\n", + " 180 0.1380 23.2% 2.1m\n", + " 190 0.1772 23.2% 2.1m\n", + " 200 0.1646 23.2% 2.1m\n", + " 210 0.1834 23.2% 2.0m\n", + " 220 0.1730 23.2% 2.0m\n", + " 230 0.1534 23.2% 2.0m\n", + " 240 0.1204 23.2% 2.0m\n", + " 250 0.1681 23.2% 1.9m\n", + " 260 0.1724 23.2% 1.9m\n", + " 270 0.1896 23.2% 1.9m\n", + " 280 0.2123 23.2% 1.8m\n", + " 290 0.1613 23.2% 1.8m\n", + " 300 0.1795 23.2% 1.8m\n", + " 310 0.1674 23.2% 1.7m\n", + " 320 0.1456 23.2% 1.7m\n", + " 330 0.2152 23.2% 1.7m\n", + " 340 0.1621 23.2% 1.6m\n", + " 350 0.1442 23.2% 1.6m\n", + " 360 0.1346 23.2% 1.6m\n", + " 370 0.1959 23.2% 1.5m\n", + " 380 0.1573 23.2% 1.5m\n", + " 390 0.2487 23.2% 1.5m\n", + " 400 0.1892 23.2% 1.4m\n", + " 410 0.1747 23.2% 1.4m\n", + " 420 0.1614 23.2% 1.4m\n", + " 430 0.2166 23.2% 1.4m\n", + " 440 0.1686 23.2% 1.3m\n", + " 450 0.1519 23.2% 1.3m\n", + " 460 0.1021 23.2% 1.3m\n", + " 470 0.1364 23.2% 1.2m\n", + " 480 0.1410 23.2% 1.2m\n", + " 490 0.1271 23.2% 1.2m\n", + " 500 0.1420 23.2% 1.1m\n", + " 510 0.1684 23.2% 1.1m\n", + " 520 0.1987 23.2% 1.1m\n", + " 530 0.1727 23.2% 1.0m\n", + " 540 0.1820 23.2% 1.0m\n", + " 550 0.1655 23.2% 1.0m\n", + " 560 0.1759 23.2% 0.9m\n", + " 570 0.1975 23.2% 0.9m\n", + " 580 0.1829 23.2% 0.9m\n", + " 590 0.1898 23.2% 0.9m\n", + " 600 0.1891 23.2% 0.8m\n", + " 610 0.1620 23.2% 0.8m\n", + " 620 0.1686 23.2% 0.8m\n", + " 630 0.2096 23.2% 0.7m\n", + " 640 0.1533 23.2% 0.7m\n", + " 650 0.1964 23.2% 0.7m\n", + " 660 0.1882 23.2% 0.6m\n", + " 670 0.2046 23.2% 0.6m\n", + " 680 0.1454 23.2% 0.6m\n", + " 690 0.1718 23.2% 0.5m\n", + " 700 0.1329 23.2% 0.5m\n", + " 710 0.1530 23.2% 0.5m\n", + " 720 0.1843 23.2% 0.4m\n", + " 730 0.1684 23.2% 0.4m\n", + " 740 0.1963 23.2% 0.4m\n", + " 750 0.1658 23.2% 0.4m\n", + " 760 0.1495 23.2% 0.3m\n", + " 770 0.1484 23.2% 0.3m\n", + " 780 0.1316 23.2% 0.3m\n", + " 790 0.1516 23.2% 0.2m\n", + " 800 0.1753 23.2% 0.2m\n", + " 810 0.1489 23.2% 0.2m\n", + " 820 0.1598 23.2% 0.1m\n", + " 830 0.1787 23.2% 0.1m\n", + " 840 0.1328 23.2% 0.1m\n", + " 850 0.1288 23.2% 0.0m\n", + " 860 0.1638 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 97 COMPLETE in 162s\n", + "\n", + "โœ… EPOCH 97 SUMMARY\n", + " โฑ๏ธ Time: 170s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16648 โ†’ Val=0.16402\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.2%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 9m\n", + "\n", + "๐Ÿ”„ Epoch 98/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1779 23.2% 2.7m\n", + " 10 0.1657 23.2% 2.7m\n", + " 20 0.1455 23.2% 2.7m\n", + " 30 0.1772 23.2% 2.6m\n", + " 40 0.1553 23.2% 2.6m\n", + " 50 0.2412 23.2% 2.6m\n", + " 60 0.1309 23.2% 2.5m\n", + " 70 0.2068 23.2% 2.5m\n", + " 80 0.1673 23.2% 2.5m\n", + " 90 0.1614 23.2% 2.4m\n", + " 100 0.1553 23.2% 2.4m\n", + " 110 0.1685 23.2% 2.4m\n", + " 120 0.1585 23.2% 2.3m\n", + " 130 0.1824 23.2% 2.3m\n", + " 140 0.1356 23.2% 2.3m\n", + " 150 0.1533 23.2% 2.3m\n", + " 160 0.1642 23.2% 2.2m\n", + " 170 0.1463 23.2% 2.2m\n", + " 180 0.1536 23.2% 2.2m\n", + " 190 0.1737 23.2% 2.1m\n", + " 200 0.1489 23.2% 2.1m\n", + " 210 0.2103 23.2% 2.1m\n", + " 220 0.1454 23.2% 2.0m\n", + " 230 0.1809 23.2% 2.0m\n", + " 240 0.2231 23.2% 2.0m\n", + " 250 0.2044 23.2% 1.9m\n", + " 260 0.1612 23.2% 1.9m\n", + " 270 0.1627 23.2% 1.9m\n", + " 280 0.1996 23.2% 1.8m\n", + " 290 0.1794 23.2% 1.8m\n", + " 300 0.1611 23.2% 1.8m\n", + " 310 0.1655 23.2% 1.7m\n", + " 320 0.1713 23.2% 1.7m\n", + " 330 0.1700 23.2% 1.7m\n", + " 340 0.1797 23.2% 1.6m\n", + " 350 0.1886 23.2% 1.6m\n", + " 360 0.1797 23.2% 1.6m\n", + " 370 0.1366 23.2% 1.5m\n", + " 380 0.1149 23.2% 1.5m\n", + " 390 0.1819 23.2% 1.5m\n", + " 400 0.1397 23.2% 1.5m\n", + " 410 0.1863 23.2% 1.4m\n", + " 420 0.1635 23.2% 1.4m\n", + " 430 0.1533 23.2% 1.4m\n", + " 440 0.1663 23.2% 1.3m\n", + " 450 0.1695 23.2% 1.3m\n", + " 460 0.1863 23.2% 1.3m\n", + " 470 0.1573 23.2% 1.2m\n", + " 480 0.1397 23.2% 1.2m\n", + " 490 0.1731 23.2% 1.2m\n", + " 500 0.1874 23.2% 1.1m\n", + " 510 0.1463 23.2% 1.1m\n", + " 520 0.1682 23.2% 1.1m\n", + " 530 0.2214 23.2% 1.0m\n", + " 540 0.1866 23.2% 1.0m\n", + " 550 0.1634 23.2% 1.0m\n", + " 560 0.1945 23.2% 1.0m\n", + " 570 0.1781 23.2% 0.9m\n", + " 580 0.1758 23.2% 0.9m\n", + " 590 0.1734 23.2% 0.9m\n", + " 600 0.1825 23.2% 0.8m\n", + " 610 0.1805 23.2% 0.8m\n", + " 620 0.1415 23.2% 0.8m\n", + " 630 0.1418 23.2% 0.7m\n", + " 640 0.1548 23.2% 0.7m\n", + " 650 0.1823 23.2% 0.7m\n", + " 660 0.1523 23.2% 0.6m\n", + " 670 0.1756 23.2% 0.6m\n", + " 680 0.1460 23.2% 0.6m\n", + " 690 0.1849 23.2% 0.5m\n", + " 700 0.1636 23.2% 0.5m\n", + " 710 0.1959 23.2% 0.5m\n", + " 720 0.1691 23.2% 0.4m\n", + " 730 0.2137 23.2% 0.4m\n", + " 740 0.1711 23.2% 0.4m\n", + " 750 0.1636 23.2% 0.4m\n", + " 760 0.1710 23.2% 0.3m\n", + " 770 0.1810 23.2% 0.3m\n", + " 780 0.1812 23.2% 0.3m\n", + " 790 0.1612 23.2% 0.2m\n", + " 800 0.2138 23.2% 0.2m\n", + " 810 0.1857 23.2% 0.2m\n", + " 820 0.1932 23.2% 0.1m\n", + " 830 0.1617 23.2% 0.1m\n", + " 840 0.1774 23.2% 0.1m\n", + " 850 0.1579 23.2% 0.0m\n", + " 860 0.1719 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 98 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 98 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16634 โ†’ Val=0.16388\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โš ๏ธ No improvement for 10/30\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 6m\n", + "\n", + "๐Ÿ”„ Epoch 99/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1941 23.2% 2.7m\n", + " 10 0.1761 23.2% 2.7m\n", + " 20 0.1436 23.2% 2.7m\n", + " 30 0.1630 23.2% 2.7m\n", + " 40 0.1470 23.2% 2.6m\n", + " 50 0.1399 23.2% 2.6m\n", + " 60 0.1535 23.2% 2.5m\n", + " 70 0.1766 23.2% 2.5m\n", + " 80 0.2055 23.2% 2.5m\n", + " 90 0.1620 23.2% 2.4m\n", + " 100 0.0960 23.2% 2.4m\n", + " 110 0.1716 23.2% 2.4m\n", + " 120 0.1680 23.2% 2.3m\n", + " 130 0.1396 23.2% 2.3m\n", + " 140 0.1484 23.2% 2.3m\n", + " 150 0.1880 23.2% 2.2m\n", + " 160 0.1627 23.2% 2.2m\n", + " 170 0.1820 23.2% 2.2m\n", + " 180 0.1500 23.2% 2.1m\n", + " 190 0.2018 23.2% 2.1m\n", + " 200 0.1708 23.2% 2.1m\n", + " 210 0.1480 23.2% 2.1m\n", + " 220 0.2187 23.2% 2.0m\n", + " 230 0.1584 23.2% 2.0m\n", + " 240 0.1480 23.2% 2.0m\n", + " 250 0.1351 23.2% 1.9m\n", + " 260 0.1668 23.2% 1.9m\n", + " 270 0.1794 23.2% 1.9m\n", + " 280 0.1613 23.2% 1.8m\n", + " 290 0.1671 23.2% 1.8m\n", + " 300 0.1912 23.2% 1.8m\n", + " 310 0.1516 23.2% 1.7m\n", + " 320 0.1488 23.2% 1.7m\n", + " 330 0.1657 23.2% 1.7m\n", + " 340 0.1249 23.2% 1.6m\n", + " 350 0.1597 23.2% 1.6m\n", + " 360 0.1488 23.2% 1.6m\n", + " 370 0.1813 23.2% 1.6m\n", + " 380 0.1324 23.2% 1.5m\n", + " 390 0.1921 23.2% 1.5m\n", + " 400 0.1664 23.2% 1.5m\n", + " 410 0.1443 23.2% 1.4m\n", + " 420 0.1895 23.2% 1.4m\n", + " 430 0.1867 23.2% 1.4m\n", + " 440 0.2018 23.2% 1.3m\n", + " 450 0.1425 23.2% 1.3m\n", + " 460 0.1556 23.2% 1.3m\n", + " 470 0.1265 23.2% 1.2m\n", + " 480 0.1904 23.2% 1.2m\n", + " 490 0.1497 23.2% 1.2m\n", + " 500 0.1387 23.2% 1.1m\n", + " 510 0.2107 23.2% 1.1m\n", + " 520 0.1441 23.2% 1.1m\n", + " 530 0.1644 23.2% 1.0m\n", + " 540 0.1835 23.2% 1.0m\n", + " 550 0.1201 23.2% 1.0m\n", + " 560 0.1688 23.2% 1.0m\n", + " 570 0.1652 23.2% 0.9m\n", + " 580 0.1776 23.2% 0.9m\n", + " 590 0.1430 23.2% 0.9m\n", + " 600 0.1955 23.2% 0.8m\n", + " 610 0.1810 23.2% 0.8m\n", + " 620 0.1491 23.2% 0.8m\n", + " 630 0.1145 23.2% 0.7m\n", + " 640 0.1922 23.2% 0.7m\n", + " 650 0.2135 23.2% 0.7m\n", + " 660 0.1491 23.2% 0.6m\n", + " 670 0.1709 23.2% 0.6m\n", + " 680 0.1698 23.2% 0.6m\n", + " 690 0.1414 23.2% 0.5m\n", + " 700 0.1722 23.2% 0.5m\n", + " 710 0.1610 23.2% 0.5m\n", + " 720 0.1344 23.2% 0.4m\n", + " 730 0.1824 23.2% 0.4m\n", + " 740 0.1557 23.2% 0.4m\n", + " 750 0.1567 23.2% 0.4m\n", + " 760 0.1394 23.2% 0.3m\n", + " 770 0.1994 23.2% 0.3m\n", + " 780 0.1793 23.2% 0.3m\n", + " 790 0.1712 23.2% 0.2m\n", + " 800 0.1583 23.2% 0.2m\n", + " 810 0.1396 23.2% 0.2m\n", + " 820 0.1456 23.2% 0.1m\n", + " 830 0.1237 23.2% 0.1m\n", + " 840 0.1819 23.2% 0.1m\n", + " 850 0.1864 23.2% 0.0m\n", + " 860 0.1641 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 99 COMPLETE in 163s\n", + "\n", + "โœ… EPOCH 99 SUMMARY\n", + " โฑ๏ธ Time: 171s (2.8m)\n", + " ๐Ÿ“‰ Loss: Train=0.16643 โ†’ Val=0.16373\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 3m\n", + "\n", + "๐Ÿ”„ Epoch 100/100\n", + " Batch Loss GPU ETA\n", + " ----------------------------------------\n", + " 0 0.1864 23.2% 2.7m\n", + " 10 0.1964 23.2% 2.7m\n", + " 20 0.1488 23.2% 2.7m\n", + " 30 0.1378 23.2% 2.6m\n", + " 40 0.1516 23.2% 2.6m\n", + " 50 0.1569 23.2% 2.6m\n", + " 60 0.1726 23.2% 2.5m\n", + " 70 0.1556 23.2% 2.5m\n", + " 80 0.1906 23.2% 2.5m\n", + " 90 0.1568 23.2% 2.4m\n", + " 100 0.1593 23.2% 2.4m\n", + " 110 0.1425 23.2% 2.4m\n", + " 120 0.1597 23.2% 2.4m\n", + " 130 0.1718 23.2% 2.3m\n", + " 140 0.1701 23.2% 2.3m\n", + " 150 0.1744 23.2% 2.3m\n", + " 160 0.1710 23.2% 2.2m\n", + " 170 0.1492 23.2% 2.2m\n", + " 180 0.1485 23.2% 2.2m\n", + " 190 0.1578 23.2% 2.1m\n", + " 200 0.1245 23.2% 2.1m\n", + " 210 0.1356 23.2% 2.1m\n", + " 220 0.1461 23.2% 2.0m\n", + " 230 0.2050 23.2% 2.0m\n", + " 240 0.1628 23.2% 2.0m\n", + " 250 0.1375 23.2% 1.9m\n", + " 260 0.2023 23.2% 1.9m\n", + " 270 0.1679 23.2% 1.9m\n", + " 280 0.1630 23.2% 1.8m\n", + " 290 0.1582 23.2% 1.8m\n", + " 300 0.1727 23.2% 1.8m\n", + " 310 0.2252 23.2% 1.8m\n", + " 320 0.1409 23.2% 1.7m\n", + " 330 0.1716 23.2% 1.7m\n", + " 340 0.1645 23.2% 1.7m\n", + " 350 0.1837 23.2% 1.6m\n", + " 360 0.1705 23.2% 1.6m\n", + " 370 0.1386 23.2% 1.6m\n", + " 380 0.1629 23.2% 1.5m\n", + " 390 0.1754 23.2% 1.5m\n", + " 400 0.1479 23.2% 1.5m\n", + " 410 0.1306 23.2% 1.4m\n", + " 420 0.1792 23.2% 1.4m\n", + " 430 0.1969 23.2% 1.4m\n", + " 440 0.1549 23.2% 1.3m\n", + " 450 0.1668 23.2% 1.3m\n", + " 460 0.1810 23.2% 1.3m\n", + " 470 0.1804 23.2% 1.2m\n", + " 480 0.1374 23.2% 1.2m\n", + " 490 0.1800 23.2% 1.2m\n", + " 500 0.1226 23.2% 1.2m\n", + " 510 0.1769 23.2% 1.1m\n", + " 520 0.1580 23.2% 1.1m\n", + " 530 0.1520 23.2% 1.1m\n", + " 540 0.1351 23.2% 1.0m\n", + " 550 0.1477 23.2% 1.0m\n", + " 560 0.1933 23.2% 1.0m\n", + " 570 0.1746 23.2% 0.9m\n", + " 580 0.2185 23.2% 0.9m\n", + " 590 0.1226 23.2% 0.9m\n", + " 600 0.1981 23.2% 0.8m\n", + " 610 0.1304 23.2% 0.8m\n", + " 620 0.1474 23.2% 0.8m\n", + " 630 0.1835 23.2% 0.7m\n", + " 640 0.1468 23.2% 0.7m\n", + " 650 0.1605 23.2% 0.7m\n", + " 660 0.1678 23.2% 0.6m\n", + " 670 0.1721 23.2% 0.6m\n", + " 680 0.2143 23.2% 0.6m\n", + " 690 0.1345 23.2% 0.5m\n", + " 700 0.1555 23.2% 0.5m\n", + " 710 0.1372 23.2% 0.5m\n", + " 720 0.1445 23.2% 0.5m\n", + " 730 0.1967 23.2% 0.4m\n", + " 740 0.1439 23.2% 0.4m\n", + " 750 0.1417 23.2% 0.4m\n", + " 760 0.1757 23.2% 0.3m\n", + " 770 0.1839 23.2% 0.3m\n", + " 780 0.1469 23.2% 0.3m\n", + " 790 0.1524 23.2% 0.2m\n", + " 800 0.1305 23.2% 0.2m\n", + " 810 0.1927 23.2% 0.2m\n", + " 820 0.1818 23.2% 0.1m\n", + " 830 0.1653 23.2% 0.1m\n", + " 840 0.1829 23.2% 0.1m\n", + " 850 0.1795 23.2% 0.0m\n", + " 860 0.1750 23.2% 0.0m\n", + " ----------------------------------------\n", + " โœ… Epoch 100 COMPLETE in 164s\n", + "\n", + "โœ… EPOCH 100 SUMMARY\n", + " โฑ๏ธ Time: 173s (2.9m)\n", + " ๐Ÿ“‰ Loss: Train=0.16634 โ†’ Val=0.16393\n", + " ๐Ÿ“Š Acc: EC=81.6% EL=75.5% EJ=86.7% Overall=81.3%\n", + " โฑ๏ธ Remaining ETA โ‰ˆ 0m\n" + ] + } + ], + "source": [ + "import gc\n", + "import os\n", + "import time\n", + "import torch\n", + "import matplotlib.pyplot as plt\n", + "\n", + "# ================= VRAM RESET =================\n", + "torch.cuda.empty_cache()\n", + "gc.collect()\n", + "os.environ[\"PYTORCH_CUDA_ALLOC_CONF\"] = \"expandable_segments:True\"\n", + "\n", + "# AMP scaler (mixed precision)\n", + "scaler = torch.amp.GradScaler(\"cuda\")\n", + "\n", + "print(\"=\" * 80)\n", + "print(\"๐Ÿš€ STAGE 1 TRAINING STARTED - LIVE MONITORING MODE\")\n", + "print(\"=\" * 80)\n", + "\n", + "train_losses, val_losses, val_accuracies = [], [], []\n", + "best_val_loss, patience_counter = float('inf'), 0\n", + "\n", + "# GPU memory monitoring\n", + "if device.type == 'cuda':\n", + " initial_mem = torch.cuda.memory_allocated() / 1e9\n", + " total_vram = torch.cuda.get_device_properties(0).total_memory / 1e9\n", + " print(f\"๐Ÿ“Š Initial GPU memory: {initial_mem:.2f} GB / {total_vram:.1f} GB\")\n", + "else:\n", + " initial_mem, total_vram = 0, 0\n", + "\n", + "print(f\"๐Ÿ“ˆ Train batches: {len(train_loader)} | Val batches: {len(val_loader)}\")\n", + "print()\n", + "\n", + "# =====================================================\n", + "# TRAINING FUNCTION\n", + "# =====================================================\n", + "def train_one_epoch_live(model, loader, optimizer, device, epoch):\n", + " model.train()\n", + "\n", + " total_loss = 0.0\n", + " start_epoch = time.time()\n", + "\n", + " print(f\" {'Batch':<6}{'Loss':<10}{'GPU':<10}{'ETA'}\")\n", + " print(\" \" + \"-\" * 40)\n", + "\n", + " for batch_idx, (imgs, labels) in enumerate(loader):\n", + "\n", + " imgs = imgs.to(device, non_blocking=True)\n", + " labels = labels.to(device, non_blocking=True)\n", + "\n", + " optimizer.zero_grad(set_to_none=True)\n", + "\n", + " # ===== Mixed Precision Forward =====\n", + " with torch.amp.autocast(\"cuda\"):\n", + " preds = model(imgs)\n", + " loss = criterion(preds, labels)\n", + "\n", + " # ===== Backprop =====\n", + " scaler.scale(loss).backward()\n", + " scaler.step(optimizer)\n", + " scaler.update()\n", + "\n", + " total_loss += loss.item()\n", + "\n", + " # ===== progress printing =====\n", + " if batch_idx % 10 == 0:\n", + " elapsed = time.time() - start_epoch\n", + " batches_per_sec = (batch_idx + 1) / elapsed\n", + " eta = (len(loader) - batch_idx) / batches_per_sec\n", + "\n", + " gpu_mem = torch.cuda.memory_allocated() / 1e9 if device.type == \"cuda\" else 0\n", + " gpu_percent = (gpu_mem / total_vram * 100) if total_vram else 0\n", + "\n", + " print(\n", + " f\" {batch_idx:<6}{loss.item():<10.4f}\"\n", + " f\"{gpu_percent:>6.1f}% \"\n", + " f\"{eta/60:.1f}m\"\n", + " )\n", + "\n", + " avg_loss = total_loss / len(loader)\n", + "\n", + " print(\" \" + \"-\" * 40)\n", + " print(f\" โœ… Epoch {epoch} COMPLETE in {time.time()-start_epoch:.0f}s\")\n", + "\n", + " return avg_loss\n", + "\n", + "\n", + "# =====================================================\n", + "# MAIN TRAINING LOOP\n", + "# =====================================================\n", + "for epoch in range(1, NUM_EPOCHS_STAGE1 + 1):\n", + "\n", + " # ===== Unfreeze backbone =====\n", + " if epoch == 5:\n", + " print(\"๐Ÿ”“ Unfreezing backbone for fine-tuning\")\n", + "\n", + " for param in model.parameters():\n", + " param.requires_grad = True\n", + "\n", + " optimizer = torch.optim.AdamW(\n", + " model.parameters(),\n", + " lr=3e-5,\n", + " weight_decay=1e-4\n", + " )\n", + "\n", + " print(f\"\\n๐Ÿ”„ Epoch {epoch}/{NUM_EPOCHS_STAGE1}\")\n", + "\n", + " epoch_start = time.time()\n", + "\n", + " # ===== TRAIN =====\n", + " train_loss = train_one_epoch_live(model, train_loader, optimizer, device, epoch)\n", + " train_losses.append(train_loss)\n", + "\n", + " # ===== VALIDATE =====\n", + " val_loss, val_acc = validate(model, val_loader, device)\n", + "\n", + " val_losses.append(val_loss)\n", + " val_accuracies.append(val_acc)\n", + "\n", + " epoch_time = time.time() - epoch_start\n", + "\n", + " print(f\"\\nโœ… EPOCH {epoch} SUMMARY\")\n", + " print(f\" โฑ๏ธ Time: {epoch_time:.0f}s ({epoch_time/60:.1f}m)\")\n", + " print(f\" ๐Ÿ“‰ Loss: Train={train_loss:.5f} โ†’ Val={val_loss:.5f}\")\n", + " print(\n", + " f\" ๐Ÿ“Š Acc: EC={val_acc['EC_acc']:4.1f}% \"\n", + " f\"EL={val_acc['EL_acc']:4.1f}% \"\n", + " f\"EJ={val_acc['EJ_acc']:4.1f}% \"\n", + " f\"Overall={val_acc['overall_acc']:4.1f}%\"\n", + " )\n", + "\n", + " # ===== Scheduler step AFTER epoch =====\n", + " if scheduler:\n", + " scheduler.step()\n", + "\n", + " # ===== Save best model =====\n", + " if val_loss < best_val_loss:\n", + "\n", + " best_val_loss = val_loss\n", + " patience_counter = 0\n", + "\n", + " torch.save(\n", + " {\n", + " \"epoch\": epoch,\n", + " \"model_state_dict\": model.state_dict(),\n", + " \"optimizer_state_dict\": optimizer.state_dict(),\n", + " \"scheduler_state_dict\": scheduler.state_dict() if scheduler else None,\n", + " \"train_loss\": train_loss,\n", + " \"val_loss\": val_loss,\n", + " \"val_acc\": val_acc,\n", + " },\n", + " \"models/checkpoints/stage1_best.pt\",\n", + " )\n", + "\n", + " print(f\" โญ NEW BEST MODEL SAVED (val_loss={val_loss:.5f})\")\n", + "\n", + " else:\n", + " patience_counter += 1\n", + " if patience_counter % 5 == 0:\n", + " print(f\" โš ๏ธ No improvement for {patience_counter}/{PATIENCE}\")\n", + "\n", + " if patience_counter >= PATIENCE:\n", + " print(f\"\\n๐Ÿ EARLY STOPPING at epoch {epoch}\")\n", + " break\n", + "\n", + " remaining_epochs = NUM_EPOCHS_STAGE1 - epoch\n", + " eta_total = epoch_time * remaining_epochs / 60\n", + "\n", + " print(f\" โฑ๏ธ Remaining ETA โ‰ˆ {eta_total:.0f}m\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## 2.7 Plot Stage 1 Training Curves\n", + "\n", + "Visualize the training and validation loss over epochs to check for\n", + "convergence and overfitting." + ] + }, + { + "cell_type": "code", + "execution_count": 24, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Training curves saved to results/figures/stage1_training.png\n" + ] + } + ], + "source": [ + "fig, axes = plt.subplots(1, 2, figsize=(14, 5))\n", + "\n", + "# Loss curves\n", + "axes[0].plot(train_losses, label='Train Loss', alpha=0.8)\n", + "axes[0].plot(val_losses, label='Val Loss', alpha=0.8)\n", + "axes[0].set_xlabel('Epoch')\n", + "axes[0].set_ylabel('MSE Loss')\n", + "axes[0].set_title('Stage 1: Loss Curves')\n", + "axes[0].legend()\n", + "axes[0].set_yscale('log')\n", + "axes[0].grid(True, alpha=0.3)\n", + "\n", + "# Accuracy curves\n", + "epochs_range = range(1, len(val_accuracies) + 1)\n", + "axes[1].plot(epochs_range, [a['EC_acc'] for a in val_accuracies], label='EC', alpha=0.8)\n", + "axes[1].plot(epochs_range, [a['EL_acc'] for a in val_accuracies], label='EL', alpha=0.8)\n", + "axes[1].plot(epochs_range, [a['EJ_acc'] for a in val_accuracies], label='EJ', alpha=0.8)\n", + "axes[1].plot(epochs_range, [a['overall_acc'] for a in val_accuracies], label='Overall', \n", + " linewidth=2, color='black', linestyle='--')\n", + "axes[1].set_xlabel('Epoch')\n", + "axes[1].set_ylabel('Accuracy (%)')\n", + "axes[1].set_title('Stage 1: Validation Accuracy')\n", + "axes[1].legend()\n", + "axes[1].grid(True, alpha=0.3)\n", + "\n", + "plt.tight_layout()\n", + "plt.savefig('results/figures/stage1_training.png', dpi=150)\n", + "plt.show()\n", + "print(\"Training curves saved to results/figures/stage1_training.png\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "---\n", + "## โœ… Stage 1 Complete\n", + "\n", + "The pre-trained model checkpoint is saved at `models/checkpoints/stage1_best.pt`.\n", + "\n", + "**Next**: Open **Notebook 3** to fine-tune on the dispersive readout dataset (Stage 2)." + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "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": "3.13.2" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +}