0xBitNet
โก
2
Generate AI text responses in your browser
poe
wop
AI & ML interests
AI research
AGI
Recent Activity
liked a Space 2 days ago
m96-chan/0xBitNet liked a Space 4 days ago
wop/Trillim new activity 4 days ago
lmsys/lmsys-chat-1m:Request: DOIOrganizations
liked a
Space 2 days ago
liked a
Space 4 days ago
Request: DOI
1
#23 opened about 1 month ago
by 
mehendiparvin2006
updated a
Space 4 days ago
published a
Space 4 days ago
reacted to imnotkitty's post with ๐ฅ 5 days ago
Post
3930
tencent/Hy3-preview is out: an open-weights MoE reasoning model.
โ 295B total / 21B active / 256K context
โ Fused fast-and-slow thinking in a single model
โ First model trained on Hunyuan's rebuilt pretraining + RL infra (Feb โ Apr)
Benchmarks:
๐ SWE-Bench Verified, Terminal-Bench 2.0, BrowseComp, WideSearch โ competitive results, particularly strong on agentic tool use
๐ Top score on Tsinghua's 2026 Spring math PhD qualifying exam
๐ Strong context-learning and instruction-following on Tencent's CL-bench / CL-bench-Life
More details can be found in my article: https://huggingface.co/blog/imnotkitty/hy3-preview
โ 295B total / 21B active / 256K context
โ Fused fast-and-slow thinking in a single model
โ First model trained on Hunyuan's rebuilt pretraining + RL infra (Feb โ Apr)
Benchmarks:
๐ SWE-Bench Verified, Terminal-Bench 2.0, BrowseComp, WideSearch โ competitive results, particularly strong on agentic tool use
๐ Top score on Tsinghua's 2026 Spring math PhD qualifying exam
๐ Strong context-learning and instruction-following on Tencent's CL-bench / CL-bench-Life
More details can be found in my article: https://huggingface.co/blog/imnotkitty/hy3-preview
- 2 replies
liked 3
Spaces 5 days ago
published a
model 5 days ago
liked a
Space 5 days ago
liked 2
Spaces 7 days ago
Error, Fp16 not available
#2 opened 7 days ago
by wop
reacted to SeanLee97's post with ๐ฅ 7 days ago
Post
8079
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: ShadowPEFT: Shadow Network for Parameter-Efficient Fine-Tuning (2604.19254)
- GitHub: https://github.com/ShadowLLM/shadow-peft
- HF Collection: https://huggingface.co/collections/shadow-llm/shadow-peft-models
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: ShadowPEFT: Shadow Network for Parameter-Efficient Fine-Tuning (2604.19254)
- GitHub: https://github.com/ShadowLLM/shadow-peft
- HF Collection: https://huggingface.co/collections/shadow-llm/shadow-peft-models
- 7 replies
liked a
model 10 days ago
liked a
model 11 days ago
liked 3
Spaces 11 days ago