ITERTL: An Iterative Framework for Fine-tuning LLMs for RTL Code Generation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908333642874880 |
|---|---|
| author | Wu, Peiyang Guo, Nan Xiao, Xiao Li, Wenming Ye, Xiaochun Fan, Dongrui |
| author_facet | Wu, Peiyang Guo, Nan Xiao, Xiao Li, Wenming Ye, Xiaochun Fan, Dongrui |
| contents | Recently, large language models (LLMs) have demonstrated excellent performance, inspiring researchers to explore their use in automating register transfer level (RTL) code generation and improving hardware design efficiency. However, the existing approaches to fine-tune LLMs for RTL generation typically are conducted on fixed datasets, which do not fully stimulate the capability of LLMs and require large amounts of reference data, which are costly to acquire. To mitigate these issues, we innovatively introduce an iterative training paradigm named ITERTL. During each iteration, samples are drawn from the model trained in the previous cycle. Then these new samples are employed for training in current loop. Furthermore, we introduce a plug-and-play data filtering strategy, thereby encouraging the model to generate high-quality, self-contained code. Our model outperforms GPT4 and state-of-the-art (SOTA) open-source models, achieving remarkable 53.8% pass@1 rate on VerilogEval-human benchmark. Under similar conditions of data quantity and quality, our approach significantly outperforms the baseline. Extensive experiments validate the effectiveness of the proposed method. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_12022 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | ITERTL: An Iterative Framework for Fine-tuning LLMs for RTL Code Generation Wu, Peiyang Guo, Nan Xiao, Xiao Li, Wenming Ye, Xiaochun Fan, Dongrui Computation and Language Artificial Intelligence Recently, large language models (LLMs) have demonstrated excellent performance, inspiring researchers to explore their use in automating register transfer level (RTL) code generation and improving hardware design efficiency. However, the existing approaches to fine-tune LLMs for RTL generation typically are conducted on fixed datasets, which do not fully stimulate the capability of LLMs and require large amounts of reference data, which are costly to acquire. To mitigate these issues, we innovatively introduce an iterative training paradigm named ITERTL. During each iteration, samples are drawn from the model trained in the previous cycle. Then these new samples are employed for training in current loop. Furthermore, we introduce a plug-and-play data filtering strategy, thereby encouraging the model to generate high-quality, self-contained code. Our model outperforms GPT4 and state-of-the-art (SOTA) open-source models, achieving remarkable 53.8% pass@1 rate on VerilogEval-human benchmark. Under similar conditions of data quantity and quality, our approach significantly outperforms the baseline. Extensive experiments validate the effectiveness of the proposed method. |
| title | ITERTL: An Iterative Framework for Fine-tuning LLMs for RTL Code Generation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2407.12022 |