Qwen3-4B-Thinking-2507-StreetMath

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.8.1

adapter: lora
base_model: Qwen/Qwen3-4B-Thinking-2507
bf16: auto
dataset_processes: 32
datasets:
- type: alpaca
  message_property_mappings:
    instruction: instruction
    input: input
    output: output
  path: LuxMuseAI/StreetMath
  trust_remote_code: false
gradient_accumulation_steps: 1
gradient_checkpointing: false
learning_rate: 0.0002
lisa_layers_attribute: model.layers
load_best_model_at_end: false
load_in_4bit: false
load_in_8bit: true
lora_alpha: 16
lora_dropout: 0.05
lora_r: 8
lora_target_modules:
- q_proj
- v_proj
- k_proj
- o_proj
- gate_proj
- down_proj
- up_proj
loraplus_lr_embedding: 1.0e-06
lr_scheduler: cosine
max_prompt_len: 512
mean_resizing_embeddings: false
micro_batch_size: 16
num_epochs: 3.0
optimizer: adamw_bnb_8bit
output_dir: ./outputs/mymodel
pretrain_multipack_attn: true
pretrain_multipack_buffer_size: 10000
qlora_sharded_model_loading: false
ray_num_workers: 1
resources_per_worker:
  GPU: 1
sample_packing_bin_size: 200
sample_packing_group_size: 100000
save_only_model: false
save_safetensors: true
sequence_len: 4096
shuffle_merged_datasets: true
skip_prepare_dataset: false
strict: false
train_on_inputs: false
trl:
  log_completions: false
  ref_model_mixup_alpha: 0.9
  ref_model_sync_steps: 64
  sync_ref_model: false
  use_vllm: false
  vllm_device: auto
  vllm_dtype: auto
  vllm_gpu_memory_utilization: 0.9
use_ray: false
val_set_size: 0.0
weight_decay: 0.0

# Weights & Biases logging (optional)
wandb_project: StreetMath
wandb_entity: ctseng-  # Replace with your actual entity name
wandb_watch: all
wandb_name: StreetMath-4B-Thinking-2507
wandb_log_model: "end"

</details><br>

<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>

outputs/mymodel

This model is a fine-tuned version of Qwen/Qwen3-4B-Thinking-2507 on the LuxMuseAI/StreetMath dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 9
• num_epochs: 3.0

Training results

Framework versions

• PEFT 0.15.1
• Transformers 4.51.0
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.4