Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
19.2
TFLOPS
1
16
Tomoya Sawada
STomoya
Follow
0 followers
·
18 following
https://stomoya.github.io/
STomoya0110
STomoya
AI & ML interests
CV, illustration.
Recent Activity
liked
a Space
27 days ago
llm-jp/open-japanese-llm-leaderboard-v2
reacted
to
SeaWolf-AI
's
post
with 👍
about 1 month ago
Why This Matters — David Defeats Goliath MODEL: https://huggingface.co/FINAL-Bench/Darwin-4B-David SPACE: https://huggingface.co/spaces/FINAL-Bench/Darwin-4B-david We're releasing Darwin-4B-David, the first second-generation model in the Darwin Opus family. By evolving an already-evolved model, it achieves 85.0% on GPQA Diamond — surpassing its 58.6% original ancestor and even gemma-4-31B (84.3%) — with just 4.5B parameters. Second-Generation Evolution Most merges start from a base model and produce a single offspring. Darwin-4B-David breaks this pattern. The Father (Darwin-4B-Opus) was already evolved from gemma-4-E4B-it with Claude Opus reasoning distillation — a Gen-1 model. The Mother (DavidAU's DECKARD-Expresso-Universe) brings Unsloth deep tuning across 5 in-house datasets with thinking mode by default. Crossbreeding these two produced the first Gen-2 Darwin model. Darwin V6's Model MRI scanned both parents across all 42 layers, assigning independent optimal ratios per layer. The Mother's creativity and Korean language hotspot (Layer 22-25, weight 0.95) was maximally absorbed, while the Father's reasoning core (Layer 30-40, weight 0.48) was preserved. This is "Merge = Evolve" applied recursively — evolution of evolution. Benchmarks Darwin-4B-David scores 85.0% on GPQA Diamond (+26.4%p over original 58.6%), evaluated generatively with maj@8 (8 generations per question, majority vote), Epoch AI prompt format, thinking mode enabled, 50 sampled questions. On ARC-Challenge (25-shot, loglikelihood), both score 64.93% — expected, as loglikelihood doesn't capture thinking-mode reasoning differences. Why This Matters gemma-4-31B (30.7B) scores 84.3%. Darwin-4B-David surpasses it at 1/7th the size — no training, no RL, just 45 minutes of MRI-guided DARE-TIES on one H100. The name "David" honors Mother creator DavidAU and evokes David vs. Goliath.
liked
a dataset
2 months ago
nvidia/Nemotron-Cascade-2-SFT-Data
View all activity
Organizations
None yet
STomoya
's models
51
Sort: Recently updated
STomoya/vit_small_patch16_224.st_mae_sb1k_ft_sb1k
Image Classification
•
Updated
Sep 1, 2024
•
2
STomoya/MambaPainter
Updated
Aug 30, 2024
•
5
•
1
STomoya/vit_tiny_patch16_224.st_mae_sb1k_ft_sb1k
Image Classification
•
Updated
Aug 29, 2024
•
2
STomoya/vit_base_patch16_224.st_mae_sb1k
Image Classification
•
Updated
Aug 20, 2024
•
5
STomoya/vit_small_patch16_224.st_mae_sb1k
Image Classification
•
Updated
Jul 27, 2024
•
6
STomoya/vit_tiny_patch16_224.st_mae_sb1k
Image Classification
•
Updated
Jul 16, 2024
•
9
STomoya/vit_base_patch16_reg1_224.st_safebooru_1k
Image Classification
•
Updated
Jun 11, 2024
•
6
STomoya/vit_base_patch32_reg1_224.st_safebooru_1k
Image Classification
•
Updated
Jun 3, 2024
•
1
STomoya/vit_small_patch16_reg4_224.st_safebooru_1k
Image Classification
•
Updated
May 29, 2024
•
4
STomoya/vit_small_patch16_reg1_224.st_safebooru_1k
Image Classification
•
Updated
May 27, 2024
•
15
STomoya/vit_tiny_patch16_reg4_224.st_safebooru_1k
Image Classification
•
Updated
May 24, 2024
•
10
STomoya/vit_tiny_patch16_reg1_224.st_safebooru_1k
Image Classification
•
Updated
May 23, 2024
•
7
STomoya/caformer_m36.st_safebooru_1k
Image Classification
•
Updated
May 19, 2024
•
4
STomoya/caformer_s36.st_safebooru_1k
Image Classification
•
Updated
May 2, 2024
•
2
STomoya/caformer_s18.st_safebooru_1k
Image Classification
•
Updated
Apr 25, 2024
•
8
STomoya/convformer_m36.st_safebooru_1k
Image Classification
•
Updated
Apr 18, 2024
•
2
STomoya/convformer_s36.st_safebooru_1k
Image Classification
•
Updated
Mar 31, 2024
•
3
STomoya/convformer_s18.st_safebooru_1k
Image Classification
•
Updated
Mar 22, 2024
•
4
STomoya/poolformerv2_m48.st_safebooru_1k
Image Classification
•
Updated
Mar 18, 2024
•
7
STomoya/poolformerv2_m36.st_safebooru_1k
Image Classification
•
Updated
Mar 2, 2024
•
3
STomoya/poolformerv2_s36.st_safebooru_1k
Image Classification
•
Updated
Feb 22, 2024
•
2
STomoya/poolformerv2_s24.st_safebooru_1k
Image Classification
•
Updated
Feb 12, 2024
•
4
STomoya/poolformerv2_s12.st_safebooru_1k
Image Classification
•
Updated
Feb 7, 2024
•
8
STomoya/poolformer_m48.st_safebooru_1k
Image Classification
•
Updated
Feb 3, 2024
•
6
STomoya/poolformer_m36.st_safebooru_1k
Image Classification
•
Updated
Jan 24, 2024
•
2
STomoya/poolformer_s36.st_safebooru_1k
Image Classification
•
Updated
Jan 18, 2024
•
4
STomoya/poolformer_s24.st_safebooru_1k
Image Classification
•
Updated
Jan 13, 2024
•
7
STomoya/poolformer_s12.st_safebooru_1k
Image Classification
•
Updated
Jan 9, 2024
STomoya/efficientnetv2_m.st_safebooru_1k
Image Classification
•
Updated
Jan 6, 2024
•
7
STomoya/efficientnetv2_s.st_safebooru_1k
Image Classification
•
Updated
Jan 1, 2024
Previous
1
2
Next