SeanLee97/angle-roberta-wwm-base-zhnli-v1
Feature Extraction • 0.1B • Updated • 68 • 3
SeanLee
SeanLee97
AI & ML interests
NLP; Information Retrieval; Information Extraction
Recent Activity
reacted to theirpost with 🤗 about 8 hours ago
Our lab recently released a paper where we introduce ShadowPEFT, a new Parameter-Efficient Fine-Tuning (PEFT) paradigm tailored for edge computing scenarios.
Unlike traditional approaches such as LoRA and its variants, which inject trainable parameters directly into the weights of Transformer, requiring tight coupling with the backbone.
ShadowPEFT instead enhances the frozen large base model by adding a lightweight, centralized, pretrainable, and detachable Shadow network.
This shadow network operates in parallel with the base model, delivering learned corrections to each decoder layer. Because the shadow module is architecturally decoupled from the backbone, it can be independently trained, stored, and deployed, benefiting edge computing scenarios and edge-cloud collaboration computing.
- HF Paper: https://huggingface.co/papers/2604.19254
- GitHub: https://github.com/ShadowLLM/shadow-peft
- HF Collection: https://huggingface.co/collections/shadow-llm/shadow-peft-models
reacted to theirpost with 🔥 about 10 hours ago
Our lab recently released a paper where we introduce ShadowPEFT, a new Parameter-Efficient Fine-Tuning (PEFT) paradigm tailored for edge computing scenarios.
Unlike traditional approaches such as LoRA and its variants, which inject trainable parameters directly into the weights of Transformer, requiring tight coupling with the backbone.
ShadowPEFT instead enhances the frozen large base model by adding a lightweight, centralized, pretrainable, and detachable Shadow network.
This shadow network operates in parallel with the base model, delivering learned corrections to each decoder layer. Because the shadow module is architecturally decoupled from the backbone, it can be independently trained, stored, and deployed, benefiting edge computing scenarios and edge-cloud collaboration computing.
- HF Paper: https://huggingface.co/papers/2604.19254
- GitHub: https://github.com/ShadowLLM/shadow-peft
- HF Collection: https://huggingface.co/collections/shadow-llm/shadow-peft-models
repliedto their post about 12 hours ago
Our lab recently released a paper where we introduce ShadowPEFT, a new Parameter-Efficient Fine-Tuning (PEFT) paradigm tailored for edge computing scenarios.
Unlike traditional approaches such as LoRA and its variants, which inject trainable parameters directly into the weights of Transformer, requiring tight coupling with the backbone.
ShadowPEFT instead enhances the frozen large base model by adding a lightweight, centralized, pretrainable, and detachable Shadow network.
This shadow network operates in parallel with the base model, delivering learned corrections to each decoder layer. Because the shadow module is architecturally decoupled from the backbone, it can be independently trained, stored, and deployed, benefiting edge computing scenarios and edge-cloud collaboration computing.
- HF Paper: https://huggingface.co/papers/2604.19254
- GitHub: https://github.com/ShadowLLM/shadow-peft
- HF Collection: https://huggingface.co/collections/shadow-llm/shadow-peft-models
Organizations
SeanLee97/bellm-llama-7b-nli
Updated • 105 • 2
SeanLee97/UAE-GIS-Large-V1
Feature Extraction • 0.3B • Updated • 8 • 2
SeanLee97/mxbai-embed-large-v1-nli-matryoshka
Sentence Similarity • 0.3B • Updated • 45 • 2
SeanLee97/mxbai-embed-large-v1-sts-matryoshka
Updated • 1
SeanLee97/distilbert-base-uncased-sts-matryoshka
Updated
SeanLee97/2dmse-bert-base-nli-v1
Feature Extraction • 0.1B • Updated • 15 • 2
SeanLee97/angle-llama-13b-nli
Updated • 6 • 7
SeanLee97/angle-llama-7b-nli-v2
Updated • 19 • 19
SeanLee97/angle-llama-7b-zhnli-v1
Updated • 2
SeanLee97/angle-llama-7b-nli-20231027
Updated • 8 • 3
SeanLee97/angle-bert-base-uncased-nli-en-v1
Feature Extraction • Updated • 273 • 5