Qwen3-4B-Instruct-2507-Fr

Qwen3-4B-Instruct-2507-Fr

This model is obtained by fine-tuning Qwen/Qwen3-4B-Instruct-2507 on the kurakurai/luth-sft dataset, specifically subsets luth_smoltalk2, luth_aya_dataset, luth_croissantllm and luth_tulu3_persona_instruct. The model is used in the experiments described in https://bknyaz.github.io/blog/2026/meta-merge/. Single A100 was used for fine-tuning and evaluation.

The following versions were used for train/eval:

• python >= 3.10
• torch : 2.9.0+cu128
• lm_eval : 0.4.9.1
• vllm : 0.11.1
• transformers : 4.57.6
• datasets : 3.2.0
• numpy : 2.2.6

Training

The TRL library was used with SFT/full-rank options:

python trl/scripts/sft.py --model_name_or_path Qwen/Qwen3-4B-Instruct-2507 --dataset_name kurakurai/luth-sft --dataset_config main --learning_rate 2e-5 \
--num_train_epochs 1 --per_device_train_batch_size 2 --gradient_accumulation_steps 8 --gradient_checkpointing --eos_token '<|im_end|>' --eval_strategy no \
--completion_only_loss True --report_to wandb --output_dir /path/to/the/finetuned/model

This is by far not the most compute and performance efficient fine-tuning, but it could be a good baseline.

The dataset was preprocessed to the conversational format:

# trl/scripts/sft.py

dataset = load_dataset(...)

trainer = SFTTrainer(
    model=model,
    args=training_args,
    train_dataset=concatenate_datasets([dataset['luth_smoltalk2'], dataset['luth_aya_dataset'], dataset['luth_croissantllm'], dataset['luth_tulu3_persona_instruct']]),
    eval_dataset=dataset[script_args.dataset_test_split] if training_args.eval_strategy != "no" else None,
    peft_config=get_peft_config(model_args),
)

Evaluation

Evaluation was done with lm_eval on the test split of gsm8k, french_bench (avg score) and gsm8k-fr:

python -m lm_eval --model vllm --model_args pretrained=${model},tensor_parallel_size=1,dtype=auto,gpu_memory_utilization=0.9,data_parallel_size=1 \
 --tasks gsm8k,french_bench,gsm8k-fr --batch_size 1 --apply_chat_template=True --confirm_run_unsafe_code --trust_remote_code

To evaluate on gsm8k-fr you can use my fork https://github.com/bknyaz/lm-evaluation-harness/tree/main/lm_eval/tasks/gsm8k.

Results

License

Please refer to the license of the original model Qwen/Qwen3-4B-Instruct-2507 and dataset kurakurai/luth-sft.