Downloads · 30 days
0
SPIN_Qwen3-4B
HoangTran223/SPIN_Qwen3-4B
SPIN_Qwen3-4B is a machine learning model from HoangTran223. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as apache-2.0.
SFT init then SPIN self-play DPO on UltraChat50k, 4 iterations (2 epochs each).
Access
Public
Updated Sep 16, 2026
Repo size
89.4 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors40.2 GB · 82%
At a glance
- Library
- transformers
- License
- apache-2.0
- Access
- Public
- Created
- Aug 28, 2026
- Updated
- Sep 16, 2026
- SHA
- b2d230a5
Base models
Model card
From the Hugging Face model README
SPIN Qwen3-4B (UltraChat50k)
SFT init then SPIN self-play DPO on UltraChat50k, 4 iterations (2 epochs each).
LR 5e-7 on ite0–1, 1e-7 on ite2–3. beta=0.1, batch 2 × grad accum 2,
max_length=1024, max_prompt_length=512.
All 4 iterations finished. ite3/ is the final policy (not ite3/LATEST).
from transformers import AutoModelForCausalLM
m = AutoModelForCausalLM.from_pretrained("HoangTran223/SPIN_Qwen3-4B", subfolder="ite3")
- Library
- transformers
- License
- apache-2.0
- Created
- Aug 28, 2026
- Updated
- Sep 16, 2026