Training Domain Draft Models for Speculative Decoding: Best Practices and Insights
Paper
•
2503.07807
•
Published
SambaNova
company
Verified
Inference Provider
VERIFIED
192,127 monthly requests
AI & ML interests
None defined yet.
Recent Activity
View all activity
hongfenglu
authored
4
papers
about 2 months ago
On the Tool Manipulation Capability of Open-source Large Language Models
Paper
•
2305.16504
•
Published
•
2
Synthetic Document Question Answering in Hungarian
Paper
•
2505.23008
•
Published
Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models
Paper
•
2510.04618
•
Published
•
123
Interesting use case for using the crew platform
3
#1 opened 6 months ago
by
llmchamp
Interesting use case for using the crew platform
3
#1 opened 6 months ago
by
llmchamp
Amitabhab
updated
a
Space
7 months ago
varunbk
updated
a
dataset
8 months ago
varunbk
published
a
dataset
8 months ago
zolicsaki
updated
a
Space
8 months ago
bol20162021
updated
a
model
11 months ago
kz919
updated
a
model
11 months ago
Post
1702
Mini-QwQ an edge device friendly reasoning model distilled from QwQ-32B
🤗: kz919/QwQ-0.5B-Distilled-SFT
🇬 🇬 🇺 🇫: kz919/QwQ-0.5B-Distilled-SFT-gguf
🤖: kz919/Mini-QwQ
🤗: kz919/QwQ-0.5B-Distilled-SFT
🇬 🇬 🇺 🇫: kz919/QwQ-0.5B-Distilled-SFT-gguf
🤖: kz919/Mini-QwQ
kz919
updated
a
Space
12 months ago
kz919
authored
a
paper
about 1 year ago
Post
1645
Just for the meme.
But the clear lesson I learnt from building these demos are, the more powerful the underlying base model is, the closer you will get to GPT4o1. CoT is nothing more than simply inducing the latent reasoning capability from the model.
kz919/GPT4-O1-Proximas
But the clear lesson I learnt from building these demos are, the more powerful the underlying base model is, the closer you will get to GPT4o1. CoT is nothing more than simply inducing the latent reasoning capability from the model.
kz919/GPT4-O1-Proximas
Post
2464
"It's Sunday night, fancy a game?"
https://kz919-can-you-beat-405b-in-chess.hf.space/
built with the one and only SN fast API:
https://sambanova.ai/fast-api?api_ref=907266
https://kz919-can-you-beat-405b-in-chess.hf.space/
built with the one and only SN fast API:
https://sambanova.ai/fast-api?api_ref=907266
- 7 replies
Post
648
Good lord... Spent almost a day debugging this and it turns out it was an issue of gradio update incompatible with the new fastapi.
https://discuss.huggingface.co/t/huggingface-space-failed-after-working-initially/105514/8
Finally got it back online! Come chat with your favorite anime characters here:
kz919/Persona-AI
https://discuss.huggingface.co/t/huggingface-space-failed-after-working-initially/105514/8
Finally got it back online! Come chat with your favorite anime characters here:
kz919/Persona-AI
Post
1609
Spent a few minutes to build an alternative to Character AI on top of llama3.1 405B through SambaNova's super fast inference API
Space: kz919/Persona-AI
API referral link: https://sambanova.ai/fast-api?api_ref=907266
Space: kz919/Persona-AI
API referral link: https://sambanova.ai/fast-api?api_ref=907266
- 3 replies