syaffers/Atom-350M-GGUF
Text Generation • 0.4B • Updated • 225
Syafiq Kamarul Azman
syaffers
AI & ML interests
Inference and model servers
Recent Activity
updated a model 4 days ago
syaffers/Atom-350M-GGUF published a model 4 days ago
syaffers/Atom-350M-GGUF reacted to ManniX-ITA's post with 🚀 11 days ago
🚀 New release: Qwen3.6-27B-A3B-Coder
A code-specialized MoE carved out of Qwen3.6-35B-A3B by pure expert pruning — no fine-tuning, no distillation. I profiled all 256 experts on balanced corpora plus targeted code benchmarks (LiveCodeBench +
MultiPL-E), built a competence map with the code classes up-weighted 1.5×, and dropped the 72 weakest experts per layer (256→184, ~35B→27B). Router, attention, norms, the MTP head and the vision tower are all
preserved; active params stay at A3B and routing is baked to top-10 (revert to top-8 anytime).
📊 Benchmarks (Q6_K, temp 0.6):
• MultiPL-E 0.840
• HumanEval 0.970
• LiveCodeBench 0.688
• GSM8K 0.970 · ARC-C 0.944 · AIME 0.733
• GPQA-Diamond 0.773 · MATH-500 0.620 · IFEval 0.730
• Average 0.808
27B footprint, A3B speed, coding that punches well above its size — and the preserved MTP head gives you speculative decoding out of the box (text + vision).
🔗 Model: https://huggingface.co/ManniX-ITA/Qwen3.6-27B-A3B-Coder
📦 GGUF (+MTP): https://huggingface.co/ManniX-ITA/Qwen3.6-27B-A3B-Coder-MTP-GGUF
🦙 Ollama: https://ollama.com/mannix/qwen3.6-27b-a3b-coderOrganizations
None yet
updated a
model 4 days ago
published a
model 4 days ago
reacted to ManniX-ITA's post with 🚀 11 days ago
Post
2034
🚀 New release: Qwen3.6-27B-A3B-Coder
A code-specialized MoE carved out of Qwen3.6-35B-A3B by pure expert pruning — no fine-tuning, no distillation. I profiled all 256 experts on balanced corpora plus targeted code benchmarks (LiveCodeBench +
MultiPL-E), built a competence map with the code classes up-weighted 1.5×, and dropped the 72 weakest experts per layer (256→184, ~35B→27B). Router, attention, norms, the MTP head and the vision tower are all
preserved; active params stay at A3B and routing is baked to top-10 (revert to top-8 anytime).
📊 Benchmarks (Q6_K, temp 0.6):
• MultiPL-E 0.840
• HumanEval 0.970
• LiveCodeBench 0.688
• GSM8K 0.970 · ARC-C 0.944 · AIME 0.733
• GPQA-Diamond 0.773 · MATH-500 0.620 · IFEval 0.730
• Average 0.808
27B footprint, A3B speed, coding that punches well above its size — and the preserved MTP head gives you speculative decoding out of the box (text + vision).
🔗 Model: ManniX-ITA/Qwen3.6-27B-A3B-Coder
📦 GGUF (+MTP): ManniX-ITA/Qwen3.6-27B-A3B-Coder-MTP-GGUF
🦙 Ollama: https://ollama.com/mannix/qwen3.6-27b-a3b-coder
A code-specialized MoE carved out of Qwen3.6-35B-A3B by pure expert pruning — no fine-tuning, no distillation. I profiled all 256 experts on balanced corpora plus targeted code benchmarks (LiveCodeBench +
MultiPL-E), built a competence map with the code classes up-weighted 1.5×, and dropped the 72 weakest experts per layer (256→184, ~35B→27B). Router, attention, norms, the MTP head and the vision tower are all
preserved; active params stay at A3B and routing is baked to top-10 (revert to top-8 anytime).
📊 Benchmarks (Q6_K, temp 0.6):
• MultiPL-E 0.840
• HumanEval 0.970
• LiveCodeBench 0.688
• GSM8K 0.970 · ARC-C 0.944 · AIME 0.733
• GPQA-Diamond 0.773 · MATH-500 0.620 · IFEval 0.730
• Average 0.808
27B footprint, A3B speed, coding that punches well above its size — and the preserved MTP head gives you speculative decoding out of the box (text + vision).
🔗 Model: ManniX-ITA/Qwen3.6-27B-A3B-Coder
📦 GGUF (+MTP): ManniX-ITA/Qwen3.6-27B-A3B-Coder-MTP-GGUF
🦙 Ollama: https://ollama.com/mannix/qwen3.6-27b-a3b-coder
- 13 replies
updated a
model about 2 months ago
published a
model about 2 months ago
updated a
model about 2 months ago
published a
model about 2 months ago
liked 2
models 6 months ago
updated a
model 7 months ago
published a
model 7 months ago
reacted to cgeorgiaw's post with ❤️ 11 months ago
Post
6161
🚀🚀🚀 The largest ever dataset of co-folded 3D protein-ligand structures just dropped on HF!!
Meet SAIR (Structurally Augmented IC₅₀ Repository): 5M+ AI-generated complexes with experimentally measured drug potency data from SandboxAQ. 🚀🚀🚀
Check it out and explore here: SandboxAQ/SAIR
Meet SAIR (Structurally Augmented IC₅₀ Repository): 5M+ AI-generated complexes with experimentally measured drug potency data from SandboxAQ. 🚀🚀🚀
Check it out and explore here: SandboxAQ/SAIR
- 3 replies
published a
dataset 11 months ago
updated a
Space about 1 year ago
published a
Space about 1 year ago