Qwen3-4B-Biology-GRPO-TR/README.md

license, library_name, pipeline_tag, tags, base_model

license	library_name	pipeline_tag	tags	base_model
apache-2.0	transformers	text-generation	qwen3 reinforcement-learning grpo trust-region text-generation	Qwen/Qwen3-4B

Qwen3-4B-Biology-GRPO-TR

This repository contains the Qwen3-4B Biology GRPO batch-size-32 run. The repository name uses the project GRPO-TR naming convention, but the actual training method for this checkpoint is GRPO.

The repository root contains the final checkpoint, global_step_100. The best validation checkpoint was global_step_90, but that step does not contain an actor model checkpoint in the local run directory, so only the final actor checkpoint can be exported. checkpoints/last/ also contains the same final checkpoint.

Performance

Dataset	Method	Base model	Train batch size	Uploaded checkpoint	Uploaded val mean@16	Best val mean@16	Best checkpoint	Best actor available
Biology	GRPO	Qwen3-4B	32	100	42.62%	46.75%	90	no

Validation Mean@16

step	val_mean16	percent
10	0.326250000000	32.62%
20	0.357500000000	35.75%
30	0.382500000000	38.25%
40	0.353750000000	35.38%
50	0.418750000000	41.88%
60	0.382500000000	38.25%
70	0.395000000000	39.50%
80	0.416250000000	41.62%
90	0.467500000000	46.75%
100	0.426250000000	42.62%

Detailed Training Hyperparameters

Section	Parameter	Value	Source
Run identity	Base model	Qwen/Qwen3-4B	queue/script override
Run identity	Dataset	Biology / SciKnowEval biology	run_qwen3_generalization.sh
Run identity	Method	GRPO	run_qwen3_generalization.sh
Run identity	Config	baseline_grpo	run_qwen3_generalization.sh
Run identity	Experiment	qwen3gen-biology-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8	run_qwen3_generalization.sh
Run identity	W&B run	run-20260702_102415-94h2xs6b	wandb
Data	Train file	datasets/sciknoweval/biology/train.parquet	script override
Data	Validation file	datasets/sciknoweval/biology/test.parquet	script override
Data	Train batch size	32	queue/script override
Data	Train max samples	3200	queue/script override
Data	Prompt key	prompt	legacy_data.yaml default
Data	Reward key	data_source	legacy_data.yaml default
Data	Shuffle train data	True	user.yaml / legacy_data.yaml
Data	Validation shuffle	False	legacy_data.yaml default
Data	Filter overlong prompts	True	user.yaml
Data	Prompt truncation	error	legacy_data.yaml default
Data	enable_thinking	false	script override
Schedule	Total training steps	100	queue/script override
Schedule	Total epochs	30	ppo_trainer/user.yaml default
Schedule	Validation before train	False	queue/script override
Schedule	Save frequency	10	queue/script override
Schedule	Validation frequency	10	queue/script override
Sequence	Max prompt length	2048	queue/script override
Sequence	Max response length	8192	queue/script override
Sequence	Max model length	10240	queue/script override
Sequence	Actor max token length per GPU	10240	queue/script override
Rollout	Rollout engine	vllm	user.yaml
Rollout	Rollout dtype	bfloat16	rollout.yaml default
Rollout	Train rollout n	8	queue/script override
Rollout	Train rollout temperature	1.0	script override
Rollout	Train rollout top_p	1.0	script override
Rollout	Train rollout do_sample	True	rollout.yaml default
Rollout	Calculate rollout log probs	True	baseline_grpo.yaml / script override
Rollout	Max num batched tokens	10240	queue/script override
Rollout	vLLM GPU memory utilization	0.8	queue/script override
Rollout	Tensor model parallel size	2	rollout.yaml default
Rollout	Free cache engine	True	rollout.yaml default
Validation	Validation rollout n	16	queue/script override
Validation	Validation temperature	0.6	queue/script override
Validation	Validation top_p	0.95	queue/script override
Validation	Validation do_sample	True	queue/script override
Optimization	Optimizer	AdamW	fsdp optimizer config
Optimization	Learning rate	1e-6	GRPO method override
Optimization	LR scheduler	constant	W&B config
Optimization	LR warmup steps	10	script override
Optimization	Weight decay	0.01	script override
Optimization	Betas	(0.9, 0.999)	W&B config
Optimization	Gradient clip	1.0	script override
PPO/GRPO	Policy loss mode	vanilla	method override
PPO/GRPO	Advantage estimator	grpo	baseline_grpo.yaml
PPO/GRPO	Normalize GRPO advantages by std	False	baseline_grpo.yaml / script override
PPO/GRPO	PPO epochs	1	W&B config
PPO/GRPO	PPO mini batch size	8	queue/script override
PPO/GRPO	PPO micro batch size per GPU	1	user.yaml
PPO/GRPO	Clip ratio low	0.2	script override
PPO/GRPO	Clip ratio high	0.28	script override
PPO/GRPO	Gamma	1.0	ppo_trainer.yaml default
PPO/GRPO	Lambda	1.0	ppo_trainer.yaml default
PPO/GRPO	Actor KL loss coef	0.0	method override
PPO/GRPO	Use KL in reward	False	ppo_trainer/user.yaml
Rollout correction	Importance sampling mode	token	script override
Rollout correction	IS threshold	2.0	script override
FSDP/System	Actor strategy	fsdp	dp_actor.yaml
FSDP/System	FSDP dtype	bfloat16	W&B config
FSDP/System	FSDP model dtype	fp32	W&B config
FSDP/System	Use torch compile	True	W&B config
FSDP/System	GPUs per node	8	queue/script override
FSDP/System	Nodes	1	user.yaml
FSDP/System	GPU type	NVIDIA H200	wandb-metadata
Checkpoint/Logging	Checkpoint root	checkpoints/datasets/sciknoweval/biology	script override
Checkpoint/Logging	Latest checkpointed iteration	100	latest_checkpointed_iteration.txt
Checkpoint/Logging	Max actor checkpoints to keep	1	user.yaml
Checkpoint/Logging	Logger	console, wandb	ppo_trainer.yaml
Checkpoint/Logging	W&B entity	seongryongjung-chung-ang-university	environment
Checkpoint/Logging	W&B project	SDPO-root	user.yaml project_name
Checkpoint/Logging	W&B group	QWEN3-GRPO-generalization	method override

Raw result and artifact files:

• results/validation_mean16.csv
• results/training_scores.csv
• results/hyperparameters.csv
• results/training_score.png
• results/training_score.svg
• artifacts/config.yaml
• artifacts/wandb-summary.json
• artifacts/wandb-metadata.json
• artifacts/output.log
• artifacts/queue.log

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

repo_id = "SeongryongJung/Qwen3-4B-Biology-GRPO-TR"
tokenizer = AutoTokenizer.from_pretrained(repo_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    repo_id,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True,
)

Source

• Checkpoint: checkpoints/datasets/sciknoweval/biology/qwen3gen-biology-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8
• Uploaded actor checkpoint: global_step_100/actor
• Best validation step: global_step_90; actor checkpoint unavailable locally
• W&B run: run-20260702_102415-94h2xs6b
• Queue log: artifacts/queue.log