GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
SeongryongJung
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
SeongryongJung
Model Tree
Qwen/Qwen3-8B-Base
Input Modalities
Output Modalities
Supported Functionality
rlrt.grpo in the trainer config.compute_score reward manager.| Field | Value |
|---|---|
| Base model | Qwen/Qwen3-8B-Base |
| Train file | /home1/irteam/SDPO/self-distillation-analysis/data/math/train.parquet |
| Validation file | /home1/irteam/SDPO/self-distillation-analysis/data/math/evaluation/aime24.parquet |
| Train max samples | 25600 |
| Train batch size | 256 |
| Rollouts per prompt | 8 |
| PPO mini batch size | 128 |
| PPO micro batch size per GPU | 1 |
| Optimizer | AdamW |
| Learning rate | 1e-06 |
| Weight decay | 0.01 |
| LR warmup steps | 10 |
| Total training steps | 100 |
| Save frequency | every 10 steps |
| Validation frequency | every 10 steps |
| Max prompt length | 2048 |
| Max response length | 20480 |
| Rollout backend | vllm |
| Rollout temperature | 1 |
| Rollout top_p | 1 |
| vLLM GPU memory utilization | 0.75 |
| Actor strategy | fsdp |
| Dtype | bfloat16 |
| Advantage estimator | grpo |
| Gamma / Lambda | 1 / 1 |
| KL loss enabled | False |
| KL loss coefficient | 0.001 |
| Checkpoint uploaded | math-RLRT-Qwen3-8B-Base-128-train256-rollout8-lr1e-6-vllm0.75-modelQwen-Qwen3-8B-Base/global_step_100 |
| W&B run id | xbqpa46q |
The plot below shows critic/score/mean logged during training.
CSV data is included in training_score.csv.
| Metric | Value |
|---|---|
| Final training step | 100 |
Final critic/score/mean | 0.144531 |
Final val-core/math_dapo/acc/mean@1 | 0.0333333 |
This model is intended for internal research and analysis of math-focused RL fine-tuning methods. It has not been broadly safety evaluated for production use.
The model was trained for 100 optimization steps on a local math dataset split. Reported scores are training-time reward/validation metrics from the same experiment setup and should not be treated as broad benchmark results.