geo3k_vlm

VLM Single-Turn RL (FSDP & Megatron)

Training VLMs with FSDP or Megatron on single-turn reasoning task using GRPO on the GEO3K dataset. We used processed version here.

Supported models:

• Qwen2.5-VL
• Qwen3-VL (Dense and Moe)

Note: Please make sure the cudnn version in the environment is 9.16.0.29 to prevent severe performance regression in conv3d in torch 2.9 mentioned in pytorch/pytorch#168167. Otherwise, you can reinstall cudnn with:

pip install nvidia-cudnn-cu12==9.16.0.29

Data Preparation (For SFT Training)

The geo3k_imgurl dataset contains:

• problem: The math problem text (string)
• answer: The answer (string, e.g., "270")
• images: Image data (list)

For SFT training, we need to format the answer field for \boxed{} format and the messages. You can use the following script to format the answer field:

from datasets import load_dataset
import pandas as pd

ds = load_dataset("chenhegu/geo3k_imgurl", split="train")

def format_answer(answer: str) -> str:
    """Format answer to include \\boxed{} format."""
    return f"Answer: \\boxed{{{answer}}}"

def process_sample(sample):
    formatted_answer = f"Answer: \\boxed{{{sample['answer']}}}"
    
    sample["messages"] = [
        {"role": "user", "content": sample["problem"]},
        {"role": "assistant", "content": formatted_answer}
    ]
    return sample

ds = ds.map(process_sample)
ds.to_parquet("/root/datasets/geo3k_imgurl/train_formatted.parquet")

Reproduce

export WANDB_API_KEY=your_wandb_api_key

# Megatron backend (default -> Qwen3-VL-8B-Instruct + Megatron)
./examples/geo3k_vlm/run_geo3k_vlm.sh

# FSDP backend
SLIME_SCRIPT_TRAIN_BACKEND=fsdp ./examples/geo3k_vlm/run_geo3k_vlm.sh

# With different model
SLIME_SCRIPT_MODEL_NAME=Qwen3-VL-4B-Instruct ./examples/geo3k_vlm/run_geo3k_vlm.sh

# SFT
./examples/geo_3k_vlm/run_geo3k_vlm_sft.sh

Configuration

Environment Variable	Default	Description
SLIME_SCRIPT_TRAIN_BACKEND	megatron	Training backend (megatron or fsdp)
SLIME_SCRIPT_MODEL_NAME	Qwen3-VL-8B-Instruct	Model name
SLIME_SCRIPT_DATASET_NAME	chenhegu/geo3k_imgurl	HuggingFace dataset name
SLIME_SCRIPT_NUM_GPUS	8	Number of GPUs
SLIME_SCRIPT_EXTERNAL_RAY	0	Use external Ray cluster (1 to enable)

Supported Models

• Qwen3-VL-2B-Instruct
• Qwen3-VL-4B-Instruct
• Qwen3-VL-8B-Instruct
• Qwen3-VL-30B-A3B-Instruct
• Qwen3-VL-235B-A22B-Instruct
• Qwen3-VL-2B-Thinking
• Qwen3-VL-4B-Thinking
• Qwen3-VL-8B-Thinking
• Qwen3-VL-30B-A3B-Thinking
• Qwen3-VL-235B-A22B-Thinking

Notes

Reward Model Configuration

We experimented with three reward model configurations:

1. A geo3k-specific RM with tolerance=0.05 (to handle rounding in ground truth labels)
2. A geo3k-specific RM with tolerance=0.0 (strict matching)
3. The default math RM

All three performed similarly, so we use the default math RM for simplicity.

Numerical Precision with Non-Binary Rewards

Our initial geo3k-specific verifier produced "format scores" (0 and 0.9) instead of clean binary rewards. Under fp32, fractional values like 0.9 can't be exactly represented, so when all samples in a group have the same reward, reward - mean doesn't equal zero—creating spurious gradient signal.

We fixed this by switching to the default math RM with clean binary 0/1 rewards. If you encounter similar precision issues with non-binary rewards, you can change the reward tensor dtype from torch.float to torch.float16 in slime/ray/rollout.py (_post_process_rewards method) to truncate precision artifacts.

B200

Blackwell currently does not support fa3, we need to use --sglang-mm-attention-backend sdpa and --attn-implementation flash_attention_2