Downloads · 30 days
4
15% of all-time downloads
Qwen3-4B-Instruct_rl_64_32
haoranli-ml/Qwen3-4B-Instruct_rl_64_32
Qwen3-4B-Instruct_rl_64_32 is a machine learning model from haoranli-ml. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
All-time downloads
27
Public
Parameters
4.4B
8.8 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors8.8 GB · 100%
At a glance
- Model type
- qwen3
- Access
- Public
- Created
- Nov 20, 2025
- Updated
- Nov 20, 2025
- SHA
- 6f86143b
Model card
From the Hugging Face model README
train_batch_size=64
ppo_mini_batch_size=32
rollout_n=8
learning_rate=1e-6
kl_loss_coef=0.001
entropy_coeff=0.001
temperature=0.8
clip_ratio_low=0.2
clip_ratio_high=0.26
epoch=10
# below is settings for Dr.GRPO
loss_agg_mode="seq-mean-token-sum-norm"
use_kl_loss=False
norm_adv_by_std_in_grpo=False
- Type
- qwen3
- Created
- Nov 20, 2025
- Updated
- Nov 20, 2025