Shadowgraph
TrendingBuzzResearch
Chrome extension
Sign in

phanxuanphucnd/LLMs-SFT-DPO

Python

1

1 commits

updated Mar 1, 2024

See the code

README

Table of Contents

Structure Folders

  • data: Folder chứa data
    • data/testset: Folder chứa data test
  • models: Folder lưu trữ model sau khi fine-tuning (adapter)
  • src: Folder chứa code
  • storages: Folder lưu trữ model sử dụng để inference

Training

bash sft_run.sh

Inference

python inference.py

Similar repos

zhangzhenyu13/llm3s-conatiner

large language model training-3-stages+deployment

47

FDU-INC/SplitFM

SplitFM: A Split Parameter-Efficient Fine-Tuning and Inference Framework for Foundation Models

350

LLM-bot-ai/Finetune

0

IBM/llm-scoping

llm scoping source code

2

TIGER-AI-Lab/StructLM

Code and data for "StructLM: Towards Building Generalist Models for Structured Knowledge Grounding"…

76

nolanimage/llm_finetuning

0

PKU-RL/AdaRefiner

AdaRefiner: Refining Decisions of Language Models with Adaptive Feedback (NAACL 2024)

20

DLLXW/baby-llama2-chinese

用于从头预训练+SFT一个小参数量的中文LLaMa2的仓库；24G单卡即可运行得到一个具备简单中文问答能力的chat-llama2.

2,922

phanxuanphucnd/LLMs-SFT-DPO

Python

1

1 commits

updated Mar 1, 2024

See the code

README

Table of Contents

Structure Folders

  • data: Folder chứa data
    • data/testset: Folder chứa data test
  • models: Folder lưu trữ model sau khi fine-tuning (adapter)
  • src: Folder chứa code
  • storages: Folder lưu trữ model sử dụng để inference

Training

bash sft_run.sh

Inference

python inference.py

Similar repos

zhangzhenyu13/llm3s-conatiner

large language model training-3-stages+deployment

47

FDU-INC/SplitFM

SplitFM: A Split Parameter-Efficient Fine-Tuning and Inference Framework for Foundation Models

350

LLM-bot-ai/Finetune

0

IBM/llm-scoping

llm scoping source code

2

TIGER-AI-Lab/StructLM

Code and data for "StructLM: Towards Building Generalist Models for Structured Knowledge Grounding"…

76

nolanimage/llm_finetuning

0

PKU-RL/AdaRefiner

AdaRefiner: Refining Decisions of Language Models with Adaptive Feedback (NAACL 2024)

20

DLLXW/baby-llama2-chinese

用于从头预训练+SFT一个小参数量的中文LLaMa2的仓库；24G单卡即可运行得到一个具备简单中文问答能力的chat-llama2.

2,922