SYMBOL INDEX (376 symbols across 41 files)

FILE: src/timesfm/configs.py
  class ForecastConfig (line 22) | class ForecastConfig:
  class ResidualBlockConfig (line 64) | class ResidualBlockConfig:
  class RandomFourierFeaturesConfig (line 75) | class RandomFourierFeaturesConfig:
  class TransformerConfig (line 85) | class TransformerConfig:
  class StackedTransformersConfig (line 101) | class StackedTransformersConfig:

FILE: src/timesfm/flax/dense.py
  class ResidualBlock (line 34) | class ResidualBlock(nnx.Module):
    method __init__ (line 37) | def __init__(self, config: ResidualBlockConfig, *, rngs=nnx.Rngs(42)):
    method __call__ (line 66) | def __call__(self, x: Float[Array, "b ... i"]) -> Float[Array, "b ... ...
  class RandomFourierFeatures (line 72) | class RandomFourierFeatures(nnx.Module):
    method __init__ (line 77) | def __init__(self, config: RandomFourierFeaturesConfig, *, rngs=nnx.Rn...
    method __call__ (line 100) | def __call__(self, x: Float[Array, "b ... i"]) -> Float[Array, "b ... ...

FILE: src/timesfm/flax/normalization.py
  class RMSNorm (line 29) | class RMSNorm(nnx.Module):
    method __init__ (line 34) | def __init__(
    method __call__ (line 46) | def __call__(self, inputs: Float[Array, "b ... d"]) -> Float[Array, "b...
  class LayerNorm (line 53) | class LayerNorm(nnx.Module):
    method __init__ (line 58) | def __init__(self, num_features: int, *, epsilon: float = 1e-6, rngs=n...
    method __call__ (line 65) | def __call__(self, inputs: Float[Array, "b ... d"]) -> Float[Array, "b...

FILE: src/timesfm/flax/transformer.py
  function make_attn_mask (line 46) | def make_attn_mask(
  class RotaryPositionalEmbedding (line 67) | class RotaryPositionalEmbedding(nnx.Module):
    method __init__ (line 70) | def __init__(
    method __call__ (line 80) | def __call__(
  class PerDimScale (line 118) | class PerDimScale(nnx.Module):
    method __init__ (line 123) | def __init__(self, num_dims: int, *, rngs=nnx.Rngs(42)):
    method __call__ (line 128) | def __call__(self, x: Float[Array, "b ... d"]) -> Float[Array, "b ... ...
  class MultiHeadAttention (line 134) | class MultiHeadAttention(nnx.Module):
    method __init__ (line 137) | def __init__(
    method __call__ (line 207) | def __call__(
  class Transformer (line 291) | class Transformer(nnx.Module):
    method __init__ (line 294) | def __init__(self, config: TransformerConfig, *, rngs=nnx.Rngs(42)):
    method __call__ (line 338) | def __call__(

FILE: src/timesfm/flax/util.py
  class DecodeCache (line 33) | class DecodeCache:
  function update_running_stats (line 43) | def update_running_stats(
  function scan_along_axis (line 80) | def scan_along_axis(f, init, xs, axis: int, **kwargs):
  function revin (line 91) | def revin(

FILE: src/timesfm/timesfm_2p5/timesfm_2p5_base.py
  function strip_leading_nans (line 33) | def strip_leading_nans(arr):
  function linear_interpolation (line 49) | def linear_interpolation(arr):
  class TimesFM_2p5_200M_Definition (line 85) | class TimesFM_2p5_200M_Definition:
  class TimesFM_2p5 (line 134) | class TimesFM_2p5:
    method load_checkpoint (line 147) | def load_checkpoint(self, path: str):
    method compile (line 151) | def compile(self, forecast_config: ForecastConfig | None = None):
    method forecast (line 155) | def forecast(
    method forecast_with_covariates (line 198) | def forecast_with_covariates(

FILE: src/timesfm/timesfm_2p5/timesfm_2p5_flax.py
  function try_gc (line 48) | def try_gc():
  function _create_stacked_transformers (line 59) | def _create_stacked_transformers(
  function _scan_along_axis (line 65) | def _scan_along_axis(f, init, xs, axis: int, **kwargs):
  function _apply_stacked_transformers (line 76) | def _apply_stacked_transformers(
  class TimesFM_2p5_200M_flax_module (line 85) | class TimesFM_2p5_200M_flax_module(nnx.Module):  # pylint: disable=inval...
    method __init__ (line 96) | def __init__(self):
    method __call__ (line 126) | def __call__(
    method decode (line 149) | def decode(self, horizon: int, inputs, masks):
    method compile (line 237) | def compile(
  function _flip_quantile_fn (line 276) | def _flip_quantile_fn(x):
  function _force_flip_invariance_fn (line 284) | def _force_flip_invariance_fn(
  function _use_continuous_quantile_head_fn (line 309) | def _use_continuous_quantile_head_fn(full_forecast, quantile_spreads, ma...
  function _fix_quantile_crossing_fn (line 329) | def _fix_quantile_crossing_fn(full_forecast):
  function _before_model_decode (line 357) | def _before_model_decode(fc, inputs, masks):
  function _after_model_decode (line 385) | def _after_model_decode(
  class TimesFM_2p5_200M_flax (line 445) | class TimesFM_2p5_200M_flax(timesfm_2p5_base.TimesFM_2p5):
    method from_pretrained (line 451) | def from_pretrained(
    method compile (line 494) | def compile(

FILE: src/timesfm/timesfm_2p5/timesfm_2p5_torch.py
  class TimesFM_2p5_200M_torch_module (line 36) | class TimesFM_2p5_200M_torch_module(nn.Module):
    method __init__ (line 41) | def __init__(self):
    method load_checkpoint (line 79) | def load_checkpoint(self, path: str, **kwargs):
    method forward (line 93) | def forward(
    method decode (line 122) | def decode(self, horizon: int, inputs, masks):
    method forecast_naive (line 228) | def forecast_naive(
  class TimesFM_2p5_200M_torch (line 266) | class TimesFM_2p5_200M_torch(
    method __init__ (line 282) | def __init__(
    method _from_pretrained (line 293) | def _from_pretrained(
    method _save_pretrained (line 341) | def _save_pretrained(self, save_directory: Union[str, Path]):
    method compile (line 352) | def compile(self, forecast_config: configs.ForecastConfig, **kwargs) -...

FILE: src/timesfm/torch/dense.py
  class ResidualBlock (line 23) | class ResidualBlock(nn.Module):
    method __init__ (line 26) | def __init__(self, config: configs.ResidualBlockConfig):
    method forward (line 53) | def forward(self, x: torch.Tensor) -> torch.Tensor:
  class RandomFourierFeatures (line 59) | class RandomFourierFeatures(nn.Module):
    method __init__ (line 62) | def __init__(self, config: configs.RandomFourierFeaturesConfig):
    method forward (line 84) | def forward(self, x: torch.Tensor) -> torch.Tensor:

FILE: src/timesfm/torch/normalization.py
  class RMSNorm (line 21) | class RMSNorm(nn.Module):
    method __init__ (line 24) | def __init__(
    method forward (line 35) | def forward(self, inputs: torch.Tensor) -> torch.Tensor:

FILE: src/timesfm/torch/transformer.py
  function make_attn_mask (line 32) | def make_attn_mask(
  class RotaryPositionalEmbedding (line 56) | class RotaryPositionalEmbedding(nn.Module):
    method __init__ (line 59) | def __init__(
    method forward (line 70) | def forward(
  function _dot_product_attention (line 114) | def _dot_product_attention(
  function _torch_dot_product_attention (line 132) | def _torch_dot_product_attention(query, key, value, mask=None):
  class PerDimScale (line 154) | class PerDimScale(nn.Module):
    method __init__ (line 157) | def __init__(self, num_dims: int):
    method forward (line 162) | def forward(self, x: torch.Tensor) -> torch.Tensor:
  class MultiHeadAttention (line 169) | class MultiHeadAttention(nn.Module):
    method __init__ (line 172) | def __init__(
    method forward (line 224) | def forward(
  class Transformer (line 307) | class Transformer(nn.Module):
    method __init__ (line 310) | def __init__(self, config: configs.TransformerConfig):
    method forward (line 354) | def forward(

FILE: src/timesfm/torch/util.py
  class DecodeCache (line 24) | class DecodeCache:
  function update_running_stats (line 33) | def update_running_stats(
  function revin (line 77) | def revin(

FILE: src/timesfm/utils/xreg_lib.py
  function _unnest (line 36) | def _unnest(nested: Sequence[Sequence[Any]]) -> np.ndarray:
  function _repeat (line 40) | def _repeat(elements: Iterable[Any], counts: Iterable[int]) -> np.ndarray:
  function _to_padded_jax_array (line 46) | def _to_padded_jax_array(x: np.ndarray) -> jax.Array:
  function normalize (line 61) | def normalize(batch):
  function renormalize (line 68) | def renormalize(batch, stats):
  class BatchedInContextXRegBase (line 72) | class BatchedInContextXRegBase:
    method __init__ (line 97) | def __init__(
    method _assert_covariates (line 210) | def _assert_covariates(self, assert_covariate_shapes: bool = False) ->...
    method create_covariate_matrix (line 327) | def create_covariate_matrix(
    method fit (line 407) | def fit(self) -> Any:
  class BatchedInContextXRegLinear (line 411) | class BatchedInContextXRegLinear(BatchedInContextXRegBase):
    method fit (line 414) | def fit(

FILE: timesfm-forecasting/examples/anomaly-detection/detect_anomalies.py
  function detect_context_anomalies (line 50) | def detect_context_anomalies(
  function build_synthetic_future (line 98) | def build_synthetic_future(
  function detect_forecast_anomalies (line 121) | def detect_forecast_anomalies(
  function plot_results (line 172) | def plot_results(
  function main (line 391) | def main() -> None:

FILE: timesfm-forecasting/examples/covariates-forecasting/demo_covariates.py
  function generate_sales_data (line 49) | def generate_sales_data() -> dict:
  function create_visualization (line 132) | def create_visualization(data: dict) -> None:
  function demonstrate_api (line 405) | def demonstrate_api() -> None:
  function explain_xreg_modes (line 431) | def explain_xreg_modes() -> None:
  function main (line 450) | def main() -> None:

FILE: timesfm-forecasting/examples/global-temperature/generate_animation_data.py
  function main (line 30) | def main() -> None:

FILE: timesfm-forecasting/examples/global-temperature/generate_gif.py
  function create_frame (line 26) | def create_frame(
  function main (line 157) | def main() -> None:

FILE: timesfm-forecasting/examples/global-temperature/generate_html.py
  function main (line 521) | def main() -> None:

FILE: timesfm-forecasting/examples/global-temperature/visualize_forecast.py
  function main (line 30) | def main() -> None:

FILE: timesfm-forecasting/scripts/check_system.py
  class CheckResult (line 75) | class CheckResult:
    method icon (line 82) | def icon(self) -> str:
    method __str__ (line 85) | def __str__(self) -> str:
  class SystemReport (line 90) | class SystemReport:
    method passed (line 99) | def passed(self) -> bool:
    method to_dict (line 102) | def to_dict(self) -> dict[str, Any]:
  function _get_total_ram_gb (line 127) | def _get_total_ram_gb() -> float:
  function _get_available_ram_gb (line 174) | def _get_available_ram_gb() -> float:
  function check_ram (line 223) | def check_ram(profile: dict[str, Any]) -> CheckResult:
  function check_gpu (line 263) | def check_gpu() -> CheckResult:
  function check_disk (line 304) | def check_disk(profile: dict[str, Any]) -> CheckResult:
  function check_python (line 337) | def check_python() -> CheckResult:
  function check_package (line 358) | def check_package(pkg_name: str, import_name: str | None = None) -> Chec...
  function recommend_batch_size (line 384) | def recommend_batch_size(report: SystemReport) -> int:
  function estimate_memory_gb (line 428) | def estimate_memory_gb(
  function check_dataset_fit (line 481) | def check_dataset_fit(
  function print_memory_estimate (line 539) | def print_memory_estimate(
  function run_checks (line 595) | def run_checks(model_version: str = "v2.5") -> SystemReport:
  function print_report (line 637) | def print_report(report: SystemReport) -> None:
  function main (line 654) | def main() -> None:

FILE: timesfm-forecasting/scripts/forecast_csv.py
  function run_preflight (line 32) | def run_preflight() -> dict:
  function load_model (line 49) | def load_model(batch_size: int = 32):
  function load_csv (line 78) | def load_csv(
  function forecast_series (line 118) | def forecast_series(
  function write_csv_output (line 144) | def write_csv_output(
  function write_json_output (line 187) | def write_json_output(results: dict[str, dict], output_path: str) -> None:
  function main (line 194) | def main() -> None:

FILE: v1/experiments/baselines/timegpt_pipeline.py
  function get_seasonality (line 33) | def get_seasonality(freq: str) -> int:
  function maybe_convert_col_to_datetime (line 37) | def maybe_convert_col_to_datetime(
  function zero_pad_time_series (line 46) | def zero_pad_time_series(df, freq, min_length=36):
  class Forecaster (line 84) | class Forecaster:
    method forecast (line 90) | def forecast(
    method cross_validation (line 98) | def cross_validation(
  class TimeGPT (line 152) | class TimeGPT(Forecaster):
    method __init__ (line 159) | def __init__(
    method _get_client (line 173) | def _get_client(self) -> NixtlaClient:
    method forecast (line 184) | def forecast(
  function run_timegpt (line 225) | def run_timegpt(

FILE: v1/experiments/extended_benchmarks/run_timegpt.py
  function main (line 70) | def main():

FILE: v1/experiments/extended_benchmarks/run_timesfm.py
  function main (line 88) | def main():

FILE: v1/experiments/extended_benchmarks/utils.py
  function parallel_transform (line 36) | def parallel_transform(inp):
  function quantile_loss (line 41) | def quantile_loss(
  class ExperimentHandler (line 59) | class ExperimentHandler:
    method __init__ (line 61) | def __init__(
    method _maybe_download_m3_or_m5_file (line 104) | def _maybe_download_m3_or_m5_file(dataset: str):
    method _transform_quantiles_to_levels (line 124) | def _transform_quantiles_to_levels(quantiles: List[float]) -> List[int]:
    method _create_dir_if_not_exists (line 132) | def _create_dir_if_not_exists(directory: str):
    method _transform_gluonts_instance_to_df (line 136) | def _transform_gluonts_instance_to_df(
    method _transform_gluonts_dataset_to_df (line 151) | def _transform_gluonts_dataset_to_df(
    method train_df (line 164) | def train_df(self) -> pd.DataFrame:
    method test_df (line 169) | def test_df(self) -> pd.DataFrame:
    method save_dataframe (line 177) | def save_dataframe(self, df: pd.DataFrame, file_name: str):
    method save_results (line 180) | def save_results(
    method fcst_from_level_to_quantiles (line 193) | def fcst_from_level_to_quantiles(
    method evaluate_models (line 213) | def evaluate_models(self, models: List[str]) -> pd.DataFrame:
    method evaluate_from_predictions (line 232) | def evaluate_from_predictions(

FILE: v1/experiments/long_horizon_benchmarks/run_eval.py
  function get_forecasts (line 95) | def get_forecasts(model_path, model, past, freq, pred_len):
  function _mse (line 112) | def _mse(y_pred, y_true):
  function _mae (line 117) | def _mae(y_pred, y_true):
  function _smape (line 122) | def _smape(y_pred, y_true):
  function eval (line 131) | def eval():

FILE: v1/peft/finetune.py
  function finetune (line 63) | def finetune(

FILE: v1/src/adapter/dora_layers.py
  class DoraTheta (line 23) | class DoraTheta(base_layer.Theta):
    method __init__ (line 24) | def __init__(self, module):
    method _dora_initialized (line 27) | def _dora_initialized(self):
    method _dorafy_var (line 40) | def _dorafy_var(self, w):
    method __getattr__ (line 55) | def __getattr__(self, k):
    method __getitem__ (line 65) | def __getitem__(self, k):
  class DoraThetaDescriptor (line 76) | class DoraThetaDescriptor:
    method __get__ (line 79) | def __get__(self, obj, objtype=None):
  class DoraLinear (line 83) | class DoraLinear(linears.Linear):
    method setup (line 88) | def setup(self) -> None:
  class DoraAttentionProjection (line 121) | class DoraAttentionProjection(attentions.AttentionProjection):
    method setup (line 126) | def setup(self) -> None:
  class DoraCombinedQKVProjection (line 166) | class DoraCombinedQKVProjection(attentions.CombinedQKVProjectionLayer):
    method setup (line 171) | def setup(self) -> None:

FILE: v1/src/adapter/lora_layers.py
  class LoraTheta (line 23) | class LoraTheta(base_layer.Theta):
    method __init__ (line 24) | def __init__(self, module):
    method _lora_initialized (line 27) | def _lora_initialized(self):
    method _lorafy_var (line 38) | def _lorafy_var(self, w):
    method __getattr__ (line 46) | def __getattr__(self, k):
    method __getitem__ (line 56) | def __getitem__(self, k):
  class LoraThetaDescriptor (line 67) | class LoraThetaDescriptor:
    method __get__ (line 70) | def __get__(self, obj, objtype=None):
  class LoraLinear (line 74) | class LoraLinear(linears.Linear):
    method setup (line 79) | def setup(self) -> None:
  class LoraAttentionProjection (line 103) | class LoraAttentionProjection(attentions.AttentionProjection):
    method setup (line 108) | def setup(self) -> None:
  class LoraCombinedQKVProjection (line 139) | class LoraCombinedQKVProjection(attentions.CombinedQKVProjectionLayer):
    method setup (line 144) | def setup(self) -> None:

FILE: v1/src/adapter/utils.py
  function get_adapter_params (line 43) | def get_adapter_params(
  function load_adapter_checkpoint (line 101) | def load_adapter_checkpoint(
  function _merge_adapter_weights (line 200) | def _merge_adapter_weights(
  function _get_adapter_weight_params (line 281) | def _get_adapter_weight_params(
  function load_adapter_layer (line 334) | def load_adapter_layer(
  function _initialize_adapter_params (line 417) | def _initialize_adapter_params(

FILE: v1/src/finetuning/finetuning_example.py
  class TimeSeriesDataset (line 51) | class TimeSeriesDataset(Dataset):
    method __init__ (line 54) | def __init__(self,
    method _prepare_samples (line 77) | def _prepare_samples(self) -> None:
    method __len__ (line 88) | def __len__(self) -> int:
    method __getitem__ (line 91) | def __getitem__(
  function prepare_datasets (line 105) | def prepare_datasets(series: np.ndarray,
  function get_model (line 141) | def get_model(load_weights: bool = False):
  function plot_predictions (line 171) | def plot_predictions(
  function get_data (line 248) | def get_data(context_len: int,
  function single_gpu_example (line 269) | def single_gpu_example():
  function setup_process (line 300) | def setup_process(rank, world_size, model, config, train_dataset, val_da...
  function multi_gpu_example (line 335) | def multi_gpu_example():
  function main (line 372) | def main(argv):

FILE: v1/src/finetuning/finetuning_torch.py
  class MetricsLogger (line 21) | class MetricsLogger(ABC):
    method log_metrics (line 29) | def log_metrics(self,
    method close (line 41) | def close(self) -> None:
  class WandBLogger (line 46) | class WandBLogger(MetricsLogger):
    method __init__ (line 55) | def __init__(self, project: str, config: Dict[str, Any], rank: int = 0):
    method log_metrics (line 60) | def log_metrics(self,
    method close (line 72) | def close(self) -> None:
  class DistributedManager (line 78) | class DistributedManager:
    method __init__ (line 89) | def __init__(
    method setup (line 103) | def setup(self) -> None:
    method cleanup (line 113) | def cleanup(self) -> None:
  class FinetuningConfig (line 120) | class FinetuningConfig:
  class TimesFMFinetuner (line 162) | class TimesFMFinetuner:
    method __init__ (line 173) | def __init__(
    method _setup_distributed_model (line 203) | def _setup_distributed_model(self) -> nn.Module:
    method _create_dataloader (line 210) | def _create_dataloader(self, dataset: Dataset, is_train: bool) -> Data...
    method _quantile_loss (line 236) | def _quantile_loss(self, pred: torch.Tensor, actual: torch.Tensor,
    method _process_batch (line 251) | def _process_batch(self, batch: List[torch.Tensor]) -> tuple:
    method _train_epoch (line 279) | def _train_epoch(self, train_loader: DataLoader,
    method _validate (line 312) | def _validate(self, val_loader: DataLoader) -> float:
    method finetune (line 339) | def finetune(self, train_dataset: Dataset,

FILE: v1/src/timesfm/data_loader.py
  class TimeSeriesdata (line 27) | class TimeSeriesdata(object):
    method __init__ (line 30) | def __init__(
    method _get_cat_cols (line 120) | def _get_cat_cols(self, cat_cov_cols):
    method _normalize_data (line 131) | def _normalize_data(self):
    method train_gen (line 137) | def train_gen(self):
    method test_val_gen (line 178) | def test_val_gen(self, mode='val', shift=1):
    method _get_features_and_ts (line 220) | def _get_features_and_ts(self, dtimes, tsidx, hist_len=None):
    method tf_dataset (line 245) | def tf_dataset(self, mode='train', shift=1):

FILE: v1/src/timesfm/patched_decoder.py
  function _shift_padded_seq (line 61) | def _shift_padded_seq(mask: JTensor, seq: JTensor) -> JTensor:
  class ResidualBlock (line 84) | class ResidualBlock(base_layer.BaseLayer):
    method setup (line 107) | def setup(self):
    method __call__ (line 146) | def __call__(self, inputs: JTensor) -> JTensor:
  function _masked_mean_std (line 157) | def _masked_mean_std(inputs: JTensor,
  function _create_quantiles (line 206) | def _create_quantiles() -> list[float]:
  class PatchedTimeSeriesDecoder (line 211) | class PatchedTimeSeriesDecoder(base_layer.BaseLayer):
    method setup (line 242) | def setup(self) -> None:
    method transform_decode_state (line 288) | def transform_decode_state(
    method _forward_transform (line 293) | def _forward_transform(
    method _reverse_transform (line 305) | def _reverse_transform(self, outputs: JTensor,
    method _preprocess_input (line 311) | def _preprocess_input(
    method _postprocess_output (line 350) | def _postprocess_output(
    method __call__ (line 365) | def __call__(self, inputs: NestedMap) -> NestedMap:
    method decode (line 399) | def decode(
  class PatchedDecoderFinetuneModel (line 482) | class PatchedDecoderFinetuneModel(base_model.BaseModel):
    method setup (line 493) | def setup(self) -> None:
    method compute_predictions (line 496) | def compute_predictions(self, input_batch: NestedMap) -> NestedMap:
    method _quantile_loss (line 515) | def _quantile_loss(self, pred: JTensor, actual: JTensor,
    method compute_loss (line 532) | def compute_loss(self, prediction_output: NestedMap,

FILE: v1/src/timesfm/pytorch_patched_decoder.py
  function create_quantiles (line 24) | def create_quantiles() -> list[float]:
  class TimesFMConfig (line 29) | class TimesFMConfig:
  function _masked_mean_std (line 62) | def _masked_mean_std(
  function _shift_padded_seq (line 112) | def _shift_padded_seq(mask: torch.Tensor, seq: torch.Tensor) -> torch.Te...
  function get_large_negative_number (line 146) | def get_large_negative_number(dtype: torch.dtype) -> torch.Tensor:
  function apply_mask_to_logits (line 155) | def apply_mask_to_logits(logits: torch.Tensor,
  function convert_paddings_to_mask (line 173) | def convert_paddings_to_mask(
  function causal_mask (line 191) | def causal_mask(input_t: torch.Tensor) -> torch.Tensor:
  function merge_masks (line 211) | def merge_masks(a: torch.Tensor, b: torch.Tensor) -> torch.Tensor:
  class ResidualBlock (line 239) | class ResidualBlock(nn.Module):
    method __init__ (line 242) | def __init__(
    method forward (line 264) | def forward(self, x):
  class RMSNorm (line 271) | class RMSNorm(torch.nn.Module):
    method __init__ (line 274) | def __init__(
    method _norm (line 285) | def _norm(self, x):
    method forward (line 288) | def forward(self, x):
  class TransformerMLP (line 297) | class TransformerMLP(nn.Module):
    method __init__ (line 300) | def __init__(
    method forward (line 310) | def forward(self, x, paddings=None):
  class TimesFMAttention (line 320) | class TimesFMAttention(nn.Module):
    method __init__ (line 323) | def __init__(
    method _per_dim_scaling (line 352) | def _per_dim_scaling(self, query: torch.Tensor) -> torch.Tensor:
    method forward (line 360) | def forward(
  class TimesFMDecoderLayer (line 418) | class TimesFMDecoderLayer(nn.Module):
    method __init__ (line 421) | def __init__(
    method forward (line 443) | def forward(
  class StackedDecoder (line 468) | class StackedDecoder(nn.Module):
    method __init__ (line 471) | def __init__(
    method forward (line 495) | def forward(
  class PositionalEmbedding (line 518) | class PositionalEmbedding(torch.nn.Module):
    method __init__ (line 529) | def __init__(
    method forward (line 540) | def forward(self, seq_length=None, position=None):
  class PatchedTimeSeriesDecoder (line 574) | class PatchedTimeSeriesDecoder(nn.Module):
    method __init__ (line 577) | def __init__(self, config: TimesFMConfig):
    method _forward_transform (line 604) | def _forward_transform(
    method _reverse_transform (line 622) | def _reverse_transform(
    method _preprocess_input (line 629) | def _preprocess_input(
    method _postprocess_output (line 677) | def _postprocess_output(
    method forward (line 694) | def forward(
    method decode (line 712) | def decode(

FILE: v1/src/timesfm/time_features.py
  function _distance_to_holiday (line 45) | def _distance_to_holiday(holiday):
  class TimeCovariates (line 112) | class TimeCovariates(object):
    method __init__ (line 115) | def __init__(
    method _minute_of_hour (line 135) | def _minute_of_hour(self):
    method _hour_of_day (line 141) | def _hour_of_day(self):
    method _day_of_week (line 147) | def _day_of_week(self):
    method _day_of_month (line 153) | def _day_of_month(self):
    method _day_of_year (line 159) | def _day_of_year(self):
    method _month_of_year (line 165) | def _month_of_year(self):
    method _week_of_year (line 171) | def _week_of_year(self):
    method _get_holidays (line 177) | def _get_holidays(self):
    method get_covariates (line 186) | def get_covariates(self):

FILE: v1/src/timesfm/timesfm_base.py
  function process_group (line 39) | def process_group(key, group, value_name, forecast_context_len):
  function moving_average (line 44) | def moving_average(arr, window_size):
  function freq_map (line 53) | def freq_map(freq: str):
  function strip_leading_nans (line 77) | def strip_leading_nans(arr):
  function linear_interpolation (line 94) | def linear_interpolation(arr):
  function _normalize (line 131) | def _normalize(batch):
  function _renormalize (line 140) | def _renormalize(batch, stats):
  class TimesFmHparams (line 145) | class TimesFmHparams:
  class TimesFmCheckpoint (line 184) | class TimesFmCheckpoint:
  class TimesFmBase (line 205) | class TimesFmBase:
    method _logging (line 214) | def _logging(self, s):
    method __post_init__ (line 217) | def __post_init__(self) -> None:
    method __init__ (line 221) | def __init__(self, hparams: TimesFmHparams,
    method load_from_checkpoint (line 253) | def load_from_checkpoint(self, checkpoint: TimesFmCheckpoint) -> None:
    method _preprocess (line 257) | def _preprocess(
    method _forecast (line 314) | def _forecast(
    method forecast (line 347) | def forecast(
    method forecast_with_covariates (line 429) | def forecast_with_covariates(
    method forecast_on_df (line 644) | def forecast_on_df(

FILE: v1/src/timesfm/timesfm_jax.py
  class TimesFmJax (line 41) | class TimesFmJax(timesfm_base.TimesFmBase):
    method _get_sample_inputs (line 57) | def _get_sample_inputs(self):
    method __post_init__ (line 85) | def __post_init__(self):
    method load_from_checkpoint (line 94) | def load_from_checkpoint(
    method jit_decode (line 178) | def jit_decode(self):
    method _forecast (line 239) | def _forecast(

FILE: v1/src/timesfm/timesfm_torch.py
  class TimesFmTorch (line 30) | class TimesFmTorch(timesfm_base.TimesFmBase):
    method __post_init__ (line 33) | def __post_init__(self):
    method load_from_checkpoint (line 52) | def load_from_checkpoint(
    method _forecast (line 72) | def _forecast(

FILE: v1/src/timesfm/xreg_lib.py
  function _unnest (line 31) | def _unnest(nested: Sequence[Sequence[Any]]) -> np.ndarray:
  function _repeat (line 35) | def _repeat(elements: Iterable[Any], counts: Iterable[int]) -> np.ndarray:
  function _to_padded_jax_array (line 42) | def _to_padded_jax_array(x: np.ndarray) -> jax.Array:
  class BatchedInContextXRegBase (line 56) | class BatchedInContextXRegBase:
    method __init__ (line 81) | def __init__(
    method _assert_covariates (line 193) | def _assert_covariates(self, assert_covariate_shapes: bool = False) ->...
    method create_covariate_matrix (line 298) | def create_covariate_matrix(
    method fit (line 377) | def fit(self) -> Any:
  class BatchedInContextXRegLinear (line 381) | class BatchedInContextXRegLinear(BatchedInContextXRegBase):
    method fit (line 384) | def fit(

FILE: v1/tests/test_timesfm.py
  function create_sample_dataframe (line 25) | def create_sample_dataframe(
  function test_timesfm_forecast_on_df (line 48) | def test_timesfm_forecast_on_df(