yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1-GGUF
Text Generation • 12B • Updated • 729k • 2.7k
🏝️ On Vacation
Croc-Prog-HF
Croc-Prog-HF
AI & ML interests
| I HATE MUST PAY AN BASIC-CPU SPACE FOR USE GRADIO OR DOKER (9/7/2026) |
Unbiased and not-persuasive LLM & Datasets, ethical & responsible use: never against people, AI can sostituite Human if we lost ability of learn and apply
Recent Activity
liked a model about 5 hours ago
yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1-GGUF updated a collection about 5 hours ago
Deepfake & AI content detection liked a Space about 12 hours ago
Quazim0t0/DaisyChain-WebOrganizations
None yet
liked a
model about 5 hours ago
updated a
collection about 5 hours ago
liked a
Space about 12 hours ago
reacted to Quazim0t0's post with 🤯 about 12 hours ago
Post
2041
🌼 DaisyChain-Web: train a language model with friends or by yourself with multiple devices, in the browser, no install
Open a webpage, share a room link, and every device that joins becomes part of the training cluster. Phones, laptops, old PCs: they connect peer-to-peer over WebRTC and train one shared transformer together, entirely in the browser.
What's actually happening under the hood:
🧠 A mini transformer LM trains on FineWeb-Edu, streamed live from the HuggingFace Hub. Each device pulls its own slice (data parallelism), tokenized with our 16.5k-token Spikewhale tokenizer
⚡ Every single multiply runs through verified INT8 neural units, no float fallback. On WebGPU browsers it uses the GPU's DP4A integer dot-product hardware, admitted only after proving bit-identical results against the verified units, with a 3×INT8 fast-accurate scheme (CUTLASS's 3xTF32 trick, ported to 8-bit)
🔒 Devices average gradients every step under a sync guard: a per-step roster protocol plus weight-hash verification keeps every device's model bit-identical. If anything drifts, training stops instead of silently forking
📊 Live logs show exactly what every device contributes, step by step
💾 When you're done: test generations right on the page, download a checkpoint, or grab the inference kit, a single self-contained HTML file with the weights baked in that runs generations offline, anywhere
Works solo too. Every extra device just grows the effective batch.
👉 Try it: Quazim0t0/DaisyChain-Web
🛠 Training framework: DaisyChainAI/DaisyChain-Train
Proof of concept: only train with devices you trust. Feedback welcome!
Open a webpage, share a room link, and every device that joins becomes part of the training cluster. Phones, laptops, old PCs: they connect peer-to-peer over WebRTC and train one shared transformer together, entirely in the browser.
What's actually happening under the hood:
🧠 A mini transformer LM trains on FineWeb-Edu, streamed live from the HuggingFace Hub. Each device pulls its own slice (data parallelism), tokenized with our 16.5k-token Spikewhale tokenizer
⚡ Every single multiply runs through verified INT8 neural units, no float fallback. On WebGPU browsers it uses the GPU's DP4A integer dot-product hardware, admitted only after proving bit-identical results against the verified units, with a 3×INT8 fast-accurate scheme (CUTLASS's 3xTF32 trick, ported to 8-bit)
🔒 Devices average gradients every step under a sync guard: a per-step roster protocol plus weight-hash verification keeps every device's model bit-identical. If anything drifts, training stops instead of silently forking
📊 Live logs show exactly what every device contributes, step by step
💾 When you're done: test generations right on the page, download a checkpoint, or grab the inference kit, a single self-contained HTML file with the weights baked in that runs generations offline, anywhere
Works solo too. Every extra device just grows the effective batch.
👉 Try it: Quazim0t0/DaisyChain-Web
🛠 Training framework: DaisyChainAI/DaisyChain-Train
Proof of concept: only train with devices you trust. Feedback welcome!
liked a
model 3 days ago
liked a
Space 4 days ago
updated a
dataset 4 days ago
upvoted 2
collections 4 days ago
published a
dataset 5 days ago
upvoted a collection 5 days ago
upvoted a changelog 6 days ago
Hugging Face Changelog
Filter Models page by Hardware
• 178
updated 2
collections 7 days ago
liked a
dataset 7 days ago
liked a
model 7 days ago
