Downloads · 30 days
0
T-SPIN_Qwen3-4B
HoangTran223/T-SPIN_Qwen3-4B
T-SPIN_Qwen3-4B is a text generation model from HoangTran223. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
Self-play T-SPIN (Wang et al., NeurIPS 2025) run on Qwen3-4B SFT (HoangTran223/qwen34bsftultrachat200k20260818170154).
Access
Public
Updated Oct 6, 2026
Repo size
81.9 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors32.2 GB · 96%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- apache-2.0
- Access
- Public
- Created
- Aug 26, 2026
- Updated
- Oct 6, 2026
- SHA
- b4bb4dc5
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
T-SPIN Qwen3-4B (Ultrachat)
Self-play T-SPIN (Wang et al., NeurIPS 2025) run on Qwen3-4B SFT
(HoangTran223/qwen3_4b_sft_ultrachat200k_20260818_170154).
- Prompts: UltraChat 50k subset of
HuggingFaceH4/ultrachat_200k - Loss: T-SPIN (
alpha=1.0,beta=0.1), RMSProp, batch 2 × grad accum 2,max_length=1024 - LR:
5e-7(ite0–1),1e-7(ite2–3) proto/train.jsonl: proto responsesy_0from the initial policy
All 4 iterations finished. ite3/ is the final policy.
from transformers import AutoModelForCausalLM
m = AutoModelForCausalLM.from_pretrained("HoangTran223/T-SPIN_Qwen3-4B", subfolder="ite3")
Layout
ite0/,ite1/,ite2/,ite3/: finished checkpoints + generatedtrain.jsonlproto/train.jsonl: y_0 proto responses used by T-SPINdata/Ultrachat50k/train.jsonl: prompts used for generationscripts/T_SPIN_full_qwen3_4b.sh: launch scriptlogs/: wandb offline runs, Hydra configs, tmux stdout
- Task
- Text Generation
- Library
- transformers
- License
- apache-2.0
- Languages
- en
- Created
- Aug 26, 2026
- Updated
- Oct 6, 2026