TinyV is a reward system for efficient RL post-training that detects false negatives in current rule-based verifiers and provides more accurate reward signals via a small LLM during RL training. Experiments show that TinyV incurs only 6% additional computational cost while significantly increasing both RL efficiency and final model performance.

📄 Technical Report - Including false negative analysis and theotical insights behind TinyV
💾 Github Repo - Access the complete pipeline for more efficient RL training via TinyV
🤗 HF Collection - Training Data, Benchmarks, and Model Artifact

This model is a fine-tuned version of Qwen/Qwen3-1.7B on zhangchenxu/TinyV_Think_Training_Data_Qwen3_Balanced dataset.

Overview

How to use it?

Please refer to the codebase: https://github.com/uw-nsl/TinyV for details.

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 8
total_train_batch_size: 512
total_eval_batch_size: 64
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 2.0

Framework versions

Transformers 4.52.4
Pytorch 2.7.1+cu126
Datasets 3.6.0
Tokenizers 0.21.1

zhangchenxu/TinyV-Qwen3-1.7B-Think

Overview

How to use it?

Training hyperparameters

Framework versions

zhangchenxu/
TinyV-Qwen3-1.7B-Think