iterative/vscode-dvc

{
  "DVCLive Catalyst": {
    "prefix": "dvclive-catalyst",
    "body": [
      "from dvclive.catalyst import DVCLiveCallback",
      "",
      "runner.train(",
      "    model=${1:model}, criterion=${2:criterion}, optimizer=${3:optimizer}, loaders=${4:loaders},",
      "    callbacks=[DVCLiveCallback()])"
    ],
    "description": "DVCLive Catalyst callback"
  },
  "DVCLive Fast.ai": {
    "prefix": "dvclive-fastai",
    "body": [
      "from dvclive.fastai import DVCLiveCallback",
      "",
      "learn.fit_one_cycle(",
      "    n_epoch=${3:2},",
      "    cbs=[DVCLiveCallback()])"
    ],
    "description": "DVCLive Fast.ai callback"
  },
  "DVCLive Hugging Face": {
    "prefix": "dvclive-huggingface",
    "body": [
      "from dvclive.huggingface import DVCLiveCallback",
      "",
      "trainer = Trainer(",
      "    ${1:model}, ${2:args},",
      "    train_dataset=${3:train_data},",
      "    eval_dataset=${4:eval_data},",
      "    tokenizer=${5:tokenizer},",
      "    compute_metrics=${6:compute_metrics},",
      ")",
      "trainer.add_callback(DVCLiveCallback(save_dvc_exp=True))",
      "trainer.train()"
    ],
    "description": "DVCLive Hugging Face callback"
  },
  "DVCLive Keras": {
    "prefix": "dvclive-keras",
    "body": [
      "from dvclive.keras import DVCLiveCallback",
      "",
      "model.fit(",
      "    ${1:train_dataset}, epochs=${2:num_epochs}, validation_data=${3:validation_dataset},",
      "    callbacks=[DVCLiveCallback(save_dvc_exp=True)])"
    ],
    "description": "DVCLive Keras callback"
  },
  "DVCLive LightGBM": {
    "prefix": "dvclive-lightgbm",
    "body": [
      "from dvclive.lgbm import DVCLiveCallback",
      "",
      "lightgbm.train(",
      "  ${1:param}, ${2:train_data}, valid_sets=[${3:validation_data}], num_round=${4:5},",
      "  callbacks=[DVCLiveCallback(save_dvc_exp=True)])"
    ],
    "description": "DVCLive LightGBM callback"
  },
  "DVCLive Optuna": {
    "prefix": "dvclive-optuna",
    "body": [
      "from dvclive.optuna import DVCLiveCallback",
      "",
      "study.optimize(",
      "    ${1:objective}, n_trials=${2:7}, callbacks=[DVCLiveCallback()])"
    ],
    "description": "DVCLive Optuna callback"
  },
  "DVCLive Pytorch Lightning": {
    "prefix": "dvclive-pytorch-lightning",
    "body": [
      "import lightning.pytorch as pl",
      "from dvclive.lightning import DVCLiveLogger",
      "",
      "class LitModule(pl.LightningModule):",
      "    def __init__(self, layer_1_dim=${1:128}, learning_rate=${2:1e-2}):",
      "        super().__init__()",
      "        # layer_1_dim and learning_rate will be logged by DVCLive",
      "        self.save_hyperparameters()",
      "",
      "    def training_step(self, batch, batch_idx):",
      "        metric = ${3:...}",
      "        # See Output Format bellow",
      "        self.log(${4:\"train_metric\"}, ${5:metric}, on_step=${6:False}, on_epoch=${7:True})",
      "",
      "dvclive_logger = DVCLiveLogger(save_dvc_exp=True)",
      "",
      "model = LitModule()",
      "trainer = pl.Trainer(logger=dvclive_logger)",
      "trainer.fit(model)"
    ],
    "description": "DVCLive Pytorch Lightning example"
  },
  "DVCLive XGBoost": {
    "prefix": "dvclive-xgboost",
    "body": [
      "from dvclive.xgb import DVCLiveCallback",
      "",
      "model = xgb.XGBClassifier(",
      "    n_estimators=${1:100},",
      "    early_stopping_rounds=${2:5},",
      "    eval_metric=[${3:\"merror\", \"mlogloss\"}],",
      "    callbacks=[DVCLiveCallback()]",
      ")",
      "",
      "model.fit(",
      "    X_train,",
      "    y_train,",
      "    eval_set=[(X_test, y_test)]",
      ")"
    ],
    "description": "DVCLive XGBoost callback"
  }
}