Downloads · 30 days
4
6% of all-time downloads
reward-model-uniform-260429
WooYoungSeok/reward-model-uniform-260429
reward-model-uniform-260429 is a machine learning model from WooYoungSeok. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Math error alignment classifier fine-tuned on the 6-category error taxonomy.
All-time downloads
62
Public
Parameters
7.6B
15.2 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors15.2 GB · 100%
At a glance
- License
- apache-2.0
- Model type
- qwen2
- Access
- Public
- Created
- Apr 30, 2026
- Updated
- Apr 30, 2026
- SHA
- df801084
Base models
Model card
From the Hugging Face model README
reward-model-uniform-260429
Math error alignment classifier fine-tuned on the 6-category error taxonomy.
Training Info
- Source checkpoint:
checkpoint-1089- Selected as best via
load_best_model_at_end=True(lowesteval_loss)
- Selected as best via
- Approximate epoch: ~4.48 / 8 (1089 / 243 steps per epoch)
- Total training steps: 1944 (8 epochs)
- Base model:
Qwen/Qwen2.5-Math-7B-Instruct - Training script:
finetuning/train_0428.py
Training Hyperparameters
- per_device_train_batch_size: 8
- gradient_accumulation_steps: 4
- learning_rate: 1e-5
- weight_decay: 0.01
- warmup_ratio: 0.1
- num_train_epochs: 8
- max_seq_length: 4096
- bf16: True
- gradient_checkpointing: True
- DeepSpeed ZeRO-2 with CPU optimizer offload
- Type
- qwen2
- License
- apache-2.0
- Created
- Apr 30, 2026
- Updated
- Apr 30, 2026