> ## Documentation Index
> Fetch the complete documentation index at: https://nixtlaverse.nixtla.io/llms.txt
> Use this file to discover all available pages before exploring further.

> Time-LLM: Reprograms large language models for time series forecasting. Transforms forecasting tasks into language tasks using off-the-shelf LLM backbones.

# Time-LLM

Time-LLM is a reprogramming framework to repurpose LLMs for general time
series forecasting with the backbone language models kept intact. In
other words, it transforms a forecasting task into a “language task”
that can be tackled by an off-the-shelf LLM.

**References**

* [Ming Jin, Shiyu Wang, Lintao Ma, Zhixuan Chu,
  James Y. Zhang, Xiaoming Shi, Pin-Yu Chen, Yuxuan Liang, Yuan-Fang Li,
  Shirui Pan, Qingsong Wen. “Time-LLM: Time Series Forecasting by
  Reprogramming Large Language
  Models”](https://arxiv.org/abs/2310.01728)

<img src="https://mintcdn.com/nixtla/wOkzptAA8LlzXeB0/neuralforecast/imgs_models/timellm.png?fit=max&auto=format&n=wOkzptAA8LlzXeB0&q=85&s=dadfbafee5ced5b903842e207a48a4d0" alt="Figure 1. Time-LLM Architecture." width="612" height="372" data-path="neuralforecast/imgs_models/timellm.png" />

*Figure 1. Time-LLM Architecture.*

## 1. Time-LLM

### Usage example

```python theme={null}
import pandas as pd
import matplotlib.pyplot as plt

from neuralforecast import NeuralForecast
from neuralforecast.models import TimeLLM
from neuralforecast.utils import AirPassengersPanel

Y_train_df = AirPassengersPanel[AirPassengersPanel.ds<AirPassengersPanel['ds'].values[-12]] # 132 train
Y_test_df = AirPassengersPanel[AirPassengersPanel.ds>=AirPassengersPanel['ds'].values[-12]].reset_index(drop=True) # 12 test

prompt_prefix = "The dataset contains data on monthly air passengers. There is a yearly seasonality"

timellm = TimeLLM(h=12,
                 input_size=36,
                 llm='openai-community/gpt2',
                 prompt_prefix=prompt_prefix,
                 batch_size=16,
                 valid_batch_size=16,
                 windows_batch_size=16)

nf = NeuralForecast(
    models=[timellm],
    freq='ME'
)

nf.fit(df=Y_train_df, val_size=12)
forecasts = nf.predict(futr_df=Y_test_df)
```

## 2. Auxiliary Functions

### `ReprogrammingLayer`

```python theme={null}
ReprogrammingLayer(
    d_model, n_heads, d_keys=None, d_llm=None, attention_dropout=0.1
)
```

Bases: <code>[Module](#torch.nn.Module)</code>

ReprogrammingLayer

### `FlattenHead`

```python theme={null}
FlattenHead(n_vars, nf, target_window, head_dropout=0)
```

Bases: <code>[Module](#torch.nn.Module)</code>

FlattenHead

### `PatchEmbedding`

```python theme={null}
PatchEmbedding(d_model, patch_len, stride, dropout)
```

Bases: <code>[Module](#torch.nn.Module)</code>

PatchEmbedding

### `TokenEmbedding`

```python theme={null}
TokenEmbedding(c_in, d_model)
```

Bases: <code>[Module](#torch.nn.Module)</code>

TokenEmbedding

### `ReplicationPad1d`

```python theme={null}
ReplicationPad1d(padding)
```

Bases: <code>[Module](#torch.nn.Module)</code>

ReplicationPad1d