DaisyChain-Web
๐ผ
Train a shared model P2P by opening a browser tab
Dean Byrne PRO
Quazim0t0
AI & ML interests
DaisyChainAI๐ผ / SmallLM's / San Francisco / Open Source
Recent Activity
updated a Space about 4 hours ago
Quazim0t0/DaisyChain-Web published a Space about 4 hours ago
Quazim0t0/DaisyChain-Web updated a model about 18 hours ago
DaisyChainAI/DaisyChain-TrainOrganizations
updated a
Space about 4 hours ago
published a
Space about 4 hours ago
updated a
model about 18 hours ago
published a
model about 18 hours ago
updated a
Space about 22 hours ago
published a
Space about 22 hours ago
upvoted a paper about 22 hours ago
updated a
model about 22 hours ago
published a
model about 22 hours ago
reacted to pranavupadhyaya52's post with ๐ฅ about 24 hours ago
Post
431
Wiki-SmartBotLM Benchmarked. Here are the results:
Benchmark WikiSB | Gemma3_270M
BoolQ 37.8 | 61.4
ARC Challenge 24.2 | 28.2
HellaSwag 26.3 | 37.7
PIQA 53.5 | 66.2
Winogrande 50.2 | 52.3
Summary:
Closest benchmark: Winogrande (2.1-point gap)
Strong performance: ARC Challenge (4.0-point gap)
Moderate gap: HellaSwag and PIQA
Largest gap: BoolQ (Reading Comprehension)
It shows despite the training difference between Gemma3 270M and WikiSmartBotLM, It does come close to Gemma 3 in some benchmarks.
pranavupadhyaya52/Wiki-SmartBotLM-Instruct
Benchmark WikiSB | Gemma3_270M
BoolQ 37.8 | 61.4
ARC Challenge 24.2 | 28.2
HellaSwag 26.3 | 37.7
PIQA 53.5 | 66.2
Winogrande 50.2 | 52.3
Summary:
Closest benchmark: Winogrande (2.1-point gap)
Strong performance: ARC Challenge (4.0-point gap)
Moderate gap: HellaSwag and PIQA
Largest gap: BoolQ (Reading Comprehension)
It shows despite the training difference between Gemma3 270M and WikiSmartBotLM, It does come close to Gemma 3 in some benchmarks.
pranavupadhyaya52/Wiki-SmartBotLM-Instruct
- 3 replies
reacted to AbstractPhil's post with ๐ about 24 hours ago
Post
47
https://huggingface.co/blog/AbstractPhil/aleph-autoregressive-differentiation-ft1
After some analysis and a bit of research the upgraded aleph autoregression is capable as a prototype selection tool. I approached the direct aleph attention routing mechanism and formed a progression from it, which already provided the necessary footholds to continue into an upgraded core mechanism. The followup mechanisms show autoregression is very possible and will be simpler than expected.
The results are promising and the autoregression stable enough to scale up. Thanks to Claude Fable who is able to keep my entire research context window in scope, the progression was rapid and the results quick. The tests yielded improved accuracy over standard MLP in many cases. I believe the improvement is not topical and will scale with a bit of effort.
Fingers crossed my friends, the addressing is part of the distillation paradigm and it now learns directly without needing an expert controller. I'll be progressing the mechanism over the coming days. With enough effort and time I hope the standard mechanism becomes a universal improvement on autoregression.
After some analysis and a bit of research the upgraded aleph autoregression is capable as a prototype selection tool. I approached the direct aleph attention routing mechanism and formed a progression from it, which already provided the necessary footholds to continue into an upgraded core mechanism. The followup mechanisms show autoregression is very possible and will be simpler than expected.
The results are promising and the autoregression stable enough to scale up. Thanks to Claude Fable who is able to keep my entire research context window in scope, the progression was rapid and the results quick. The tests yielded improved accuracy over standard MLP in many cases. I believe the improvement is not topical and will scale with a bit of effort.
Fingers crossed my friends, the addressing is part of the distillation paradigm and it now learns directly without needing an expert controller. I'll be progressing the mechanism over the coming days. With enough effort and time I hope the standard mechanism becomes a universal improvement on autoregression.
- 1 reply
updated a
model about 24 hours ago
posted an update about 24 hours ago
Post
592
I never really posted about my DaisyChain project because it's still work in progress. I decided to post a small bit about it and the demo.
DaisyChain Genomics: four small DNA/RNA specialists chained behind a learned router that behave like one big genomics model, at ~7ร less active compute.
I built a modular genomics model chasing a 500M-parameter foundation model, then caught myself measuring it wrong. Here's the honest version.
DaisyChain is a different bet: instead of one monolithic DNA model, it's four ~74M specialists (eukaryote, prokaryote, mRNA, splice) chained behind a learned router, each distilled per-domain from HuggingFaceBio's Carbon-500M. Every specialist reports how surprised it is (bits/base) and the router hands each sequence to the link most at home with it. In lineage it's a cluster Branch-Train-Merge mixture of experts, so you can chain on a new domain without retraining the others.
The pitch: ~295M total params (under Carbon-500M), but only one ~74M specialist runs per query, so ~7ร cheaper per token, routing at 100% held-out.
The mistake: Carbon works in 6-mers, and I'd been scoring likelihood as 6-mer cross-entropy. By that number I was +0.043 bits/base behind, splice even "beating" Carbon. But Carbon scores at the base-pair level, which is harder and more honest. Re-run their way:
Real gap: 1.862 vs 1.787 bits/base, +0.089 behind, not +0.043
No domain actually beats Carbon; the "splice win" was an artifact
Seq recovery: euk 31.5% vs 38.9%, bacteria 40.9% vs 54.1%
DaisyChain is still behind Carbon-500M (itself a draft model, not built to top benchmarks), but by a number I can defend, and the gap closes with every per-domain pass. ๐ผ
DaisyChainAI
DaisyChainAI/Daisychain-Genomics-Demo
DaisyChainAI/daisychain-genomics
DaisyChain Genomics: four small DNA/RNA specialists chained behind a learned router that behave like one big genomics model, at ~7ร less active compute.
I built a modular genomics model chasing a 500M-parameter foundation model, then caught myself measuring it wrong. Here's the honest version.
DaisyChain is a different bet: instead of one monolithic DNA model, it's four ~74M specialists (eukaryote, prokaryote, mRNA, splice) chained behind a learned router, each distilled per-domain from HuggingFaceBio's Carbon-500M. Every specialist reports how surprised it is (bits/base) and the router hands each sequence to the link most at home with it. In lineage it's a cluster Branch-Train-Merge mixture of experts, so you can chain on a new domain without retraining the others.
The pitch: ~295M total params (under Carbon-500M), but only one ~74M specialist runs per query, so ~7ร cheaper per token, routing at 100% held-out.
The mistake: Carbon works in 6-mers, and I'd been scoring likelihood as 6-mer cross-entropy. By that number I was +0.043 bits/base behind, splice even "beating" Carbon. But Carbon scores at the base-pair level, which is harder and more honest. Re-run their way:
Real gap: 1.862 vs 1.787 bits/base, +0.089 behind, not +0.043
No domain actually beats Carbon; the "splice win" was an artifact
Seq recovery: euk 31.5% vs 38.9%, bacteria 40.9% vs 54.1%
DaisyChain is still behind Carbon-500M (itself a draft model, not built to top benchmarks), but by a number I can defend, and the gap closes with every per-domain pass. ๐ผ
DaisyChainAI/Daisychain-Genomics-Demo
DaisyChainAI/daisychain-genomics
reacted to ucr-max's post with ๐ฅ 1 day ago
Post
87
Published under Universal Computing Research: Atom2.7m, a 2.74M-parameter causal LM experiment with arithmetic-aware numeric representation.
It keeps regular BPE-style text continuation, while adding structure for numeric spans: digit ordering, place_ids, and role_ids.
Atom2.7m reaches 69.24% accuracy on ArithMark2.0.
The question behind the project is narrow: how much of small-model arithmetic performance depends on representation, rather than scale alone?
Model: UniversalComputingResearch/Atom2.7m
Article: https://huggingface.co/blog/ucr-max/atom2-7m-arithmetic-representation
Demo: UniversalComputingResearch/Atom2.7-Demo
It keeps regular BPE-style text continuation, while adding structure for numeric spans: digit ordering, place_ids, and role_ids.
Atom2.7m reaches 69.24% accuracy on ArithMark2.0.
The question behind the project is narrow: how much of small-model arithmetic performance depends on representation, rather than scale alone?
Model: UniversalComputingResearch/Atom2.7m
Article: https://huggingface.co/blog/ucr-max/atom2-7m-arithmetic-representation
Demo: UniversalComputingResearch/Atom2.7-Demo
- 1 reply
reacted to Banaxi-Tech's post with ๐ 1 day ago
Post
120
We're excited to release BananaMind 2 Mini the first model in our BananaMind 2 series!
BananaMind 2 Mini features a custom digit-aware BPE tokenizer that keeps every digit isolated, fixing the core arithmetic weakness of our previous models. It's trained on 30B tokens from FineWeb-Edu, DCLM, Cosmopedia-v2 and FineMath-4+, and already outperforms Pythia-31M despite having fewer parameters.
Check it out at
BananaMind/BananaMind-2-Mini
BananaMind 2 Pro is coming next. Apache 2.0.
BananaMind 2 Mini features a custom digit-aware BPE tokenizer that keeps every digit isolated, fixing the core arithmetic weakness of our previous models. It's trained on 30B tokens from FineWeb-Edu, DCLM, Cosmopedia-v2 and FineMath-4+, and already outperforms Pythia-31M despite having fewer parameters.
Check it out at
BananaMind/BananaMind-2-Mini
BananaMind 2 Pro is coming next. Apache 2.0.
- 1 reply
reacted to HannesVonEssen's post with ๐ฅ 1 day ago
Post
118
๐ฃ HF Viewer now supports Hugging Face login! ๐ค
โก Generate visualizations for all your models at once!
๐ Feature your models in the community showcase and set up your own profile/org collection page on hfviewer!
โ๏ธ Write your own model articles on hfviewer in the same novel interactive style as our "Gemma 4 family" article - with linking between the graph nodes and article text!
๐ We are also now rolling out support for tensor shapes, FLOPs and param counts per layer! :)
Thanks for all your positive feedback and suggestions! โค๏ธ
Try the logged in experience here: https://hfviewer.com/
Here are some really cool articles already written by our users:
๐ LFM2.5-Audio: edge-first speech inference by Anna Piunova at Liquid AI
https://hfviewer.com/LiquidAI/LFM2.5-Audio-1.5B
๐ DeepSeek V4 mHC Explained by Shakti Wadekar
https://hfviewer.com/deepseek-ai/DeepSeek-V4-Pro
๐ Borealis - open recipe for training Audio LLMs by Alex Wortega
https://hfviewer.com/Vikhrmodels/Borealis-5b-it
โก Generate visualizations for all your models at once!
๐ Feature your models in the community showcase and set up your own profile/org collection page on hfviewer!
โ๏ธ Write your own model articles on hfviewer in the same novel interactive style as our "Gemma 4 family" article - with linking between the graph nodes and article text!
๐ We are also now rolling out support for tensor shapes, FLOPs and param counts per layer! :)
Thanks for all your positive feedback and suggestions! โค๏ธ
Try the logged in experience here: https://hfviewer.com/
Here are some really cool articles already written by our users:
๐ LFM2.5-Audio: edge-first speech inference by Anna Piunova at Liquid AI
https://hfviewer.com/LiquidAI/LFM2.5-Audio-1.5B
๐ DeepSeek V4 mHC Explained by Shakti Wadekar
https://hfviewer.com/deepseek-ai/DeepSeek-V4-Pro
๐ Borealis - open recipe for training Audio LLMs by Alex Wortega
https://hfviewer.com/Vikhrmodels/Borealis-5b-it
- 1 reply
reacted to Hari5115's post with ๐ 1 day ago
Post
83
๐ Calling all space lovers โ every "Astronomy Picture of the Day" from NASA since 1995 is now an open dataset. For 30+ years, NASA has shared one amazing image of space every single day, colorful galaxies, bright stars, planets, and the sun, each with a short explanation written by a real astronomer.
It's now an open dataset that anyone can use. ๐
๐ฆ Hari5115/nasa-apod
Honestly, the pictures amazed me โ full credit to all the photographers and astronomers behind them. ๐ Whether you love space, enjoy building things, or just like looking at amazing pictures, this one's for you. If it gives you an idea, let's build it together. ๐
Feel free to use the dataset, a mention or credit is always appreciated. ๐
Data from NASA ยท public domain ยท not affiliated with NASA
#space #nasa #dataset #astronomy #opensource #photographers
It's now an open dataset that anyone can use. ๐
๐ฆ Hari5115/nasa-apod
Honestly, the pictures amazed me โ full credit to all the photographers and astronomers behind them. ๐ Whether you love space, enjoy building things, or just like looking at amazing pictures, this one's for you. If it gives you an idea, let's build it together. ๐
Feel free to use the dataset, a mention or credit is always appreciated. ๐
Data from NASA ยท public domain ยท not affiliated with NASA
#space #nasa #dataset #astronomy #opensource #photographers
reacted to razaali10's post with ๐ค 1 day ago
Post
52
I built a free, browser-based hydraulic simulation tool for water engineers โ no software to install.
๐ WNTR GIS Tool โ live on Hugging Face Spaces
https://lnkd.in/g3_fPbKN
Here's exactly how it works:
๐ญ. Upload your data
Drop in a shapefile ZIP (junctions, pipes, pumps, valves, reservoirs, tanks, customer points) or an EPANET .inp file directly. No Hugging Face account needed โ just open the link and go.
๐ฎ. Configure your simulation
Choose static or Extended Period Simulation (EPS up to 7 days), set demand per junction or distribute from customer-point shapefiles, and optionally apply a 24-hour diurnal demand pattern.
๐ฏ. Run
WNTR's pure-Python hydraulic solver runs on the server โ no EPANET binary, no local compute. Takes seconds on typical networks.
๐ฐ. Inspect results
โ Interactive map: pressure-coded nodes, diameter-scaled pipes, customer points coloured by junction pressure
โ Ontario MECP compliance gauge (275 kPa / 140 kPa thresholds)
โ EPS time-series pressure charts across any selected junctions
โ Pipe criticality analysis โ ranks which pipes cause the most failures when closed
โ Shapefile + CSV export ready for QGIS or ArcGIS
๐ฑ. Ask AI about your results
Open the AI Analysis tab, paste your own API key (Claude, GPT, Gemini, Groq, or Mistral), and ask questions directly about your network. The simulation results are sent automatically as context โ no copy-pasting.
Groq has a free tier, so users who don't want to pay anything can still run AI analysis at zero cost.
๐ No installation. Completely free to simulate.
Raza Ali | Water Infrastructure Engineer & Educator
hashtag#WaterEngineering hashtag#EPANET hashtag#HydraulicModeling hashtag#GIS hashtag#WNTR hashtag#DigitalWater hashtag#OpenSource hashtag#Streamlit hashtag#HuggingFace hashtag#Ontario hashtag#Infrastructure
๐ WNTR GIS Tool โ live on Hugging Face Spaces
https://lnkd.in/g3_fPbKN
Here's exactly how it works:
๐ญ. Upload your data
Drop in a shapefile ZIP (junctions, pipes, pumps, valves, reservoirs, tanks, customer points) or an EPANET .inp file directly. No Hugging Face account needed โ just open the link and go.
๐ฎ. Configure your simulation
Choose static or Extended Period Simulation (EPS up to 7 days), set demand per junction or distribute from customer-point shapefiles, and optionally apply a 24-hour diurnal demand pattern.
๐ฏ. Run
WNTR's pure-Python hydraulic solver runs on the server โ no EPANET binary, no local compute. Takes seconds on typical networks.
๐ฐ. Inspect results
โ Interactive map: pressure-coded nodes, diameter-scaled pipes, customer points coloured by junction pressure
โ Ontario MECP compliance gauge (275 kPa / 140 kPa thresholds)
โ EPS time-series pressure charts across any selected junctions
โ Pipe criticality analysis โ ranks which pipes cause the most failures when closed
โ Shapefile + CSV export ready for QGIS or ArcGIS
๐ฑ. Ask AI about your results
Open the AI Analysis tab, paste your own API key (Claude, GPT, Gemini, Groq, or Mistral), and ask questions directly about your network. The simulation results are sent automatically as context โ no copy-pasting.
Groq has a free tier, so users who don't want to pay anything can still run AI analysis at zero cost.
๐ No installation. Completely free to simulate.
Raza Ali | Water Infrastructure Engineer & Educator
hashtag#WaterEngineering hashtag#EPANET hashtag#HydraulicModeling hashtag#GIS hashtag#WNTR hashtag#DigitalWater hashtag#OpenSource hashtag#Streamlit hashtag#HuggingFace hashtag#Ontario hashtag#Infrastructure
reacted to ProCreations's post with ๐ค 1 day ago
Post
412
I won 4th place on the huggingface hackathon, genuinely so happy. Congrats to everyone that won, this has been super fun