Model card
TimeLLM.
Tongji University / Ant Groupopen-source7B paramsLLM Reprogramming (Llama-7B backbone)
Reprograms frozen LLMs for time-series forecasting via learned input patches + natural language prompts.
§ 01 · Benchmarks
Every benchmark TimeLLM has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | ETTh1 | Time Series · Time Series Forecasting | mae | 0.5% | #2 | 2024-05-07 | source ↗ |
| 02 | ETTh1 | Time Series · Time Series Forecasting | mse | 0.4% | #3 | 2024-05-07 | source ↗ |
| 03 | ETTh2 | Time Series · Time Series Forecasting | mae | 0.4% | #3 | 2024-05-07 | source ↗ |
| 04 | ETTh2 | Time Series · Time Series Forecasting | mse | 0.4% | #3 | 2024-05-07 | source ↗ |
| 05 | ETTm1 | Time Series · Time Series Forecasting | mae | 0.4% | #3 | 2024-05-07 | source ↗ |
| 06 | ETTm1 | Time Series · Time Series Forecasting | mse | 0.4% | #3 | 2024-05-07 | source ↗ |
| 07 | ETTm2 | Time Series · Time Series Forecasting | mae | 0.3% | #3 | 2024-05-07 | source ↗ |
| 08 | ETTm2 | Time Series · Time Series Forecasting | mse | 0.3% | #3 | 2024-05-07 | source ↗ |
| 09 | Weather | Time Series · Time Series Forecasting | mae | 0.3% | #6 | 2024-05-07 | source ↗ |
| 10 | Weather | Time Series · Time Series Forecasting | mse | 0.2% | #6 | 2024-05-07 | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Papers
1 paper with results for TimeLLM.
- 2024-05-07· Time Series· 10 results
Time-LLM: Time Series Forecasting by Reprogramming Large Language Models
Ming Jin, Shiyu Wang, Lintao Ma, Zhixuan Chu et al.
§ 05 · Sources & freshness
Where these numbers come from.
TimeLLM Table 11 (avg)
10
results
10 of 10 rows marked verified.