HayatoHongo/lfm2-vl-ja-finetuned-jamt1ep

wandb: wandb: 🚀 View run deft-sunset-40 at: wandb: Find logs at: wandb/run-20251011_193153-ismjeihd/logs

torchrun --nproc_per_node=8 ja_train_multiturn_onlast.py
--model_id LiquidAI/LFM2-VL-450M
--chat_json /workspace/Japanese_multiturn_32154_kaken_962_shuf.jsonl
--output_dir /workspace/output
--epochs 1
--lr 2e-5
--per_device_train_batch_size 16
--gradient_accumulation_steps 1
--bf16
--tf32
--report_to wandb
--logging_steps 20
--seed 42

Model Card for output

This model is a fine-tuned version of LiquidAI/LFM2-VL-450M. It has been trained using TRL.

Quick start

from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="None", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with SFT.

Framework versions

• TRL: 0.22.2
• Transformers: 4.55.0
• Pytorch: 2.8.0+cu126
• Datasets: 4.0.0
• Tokenizers: 0.21.4

Citations

Cite TRL as:

@misc{vonwerra2022trl,
    title        = {{TRL: Transformer Reinforcement Learning}},
    author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
    year         = 2020,
    journal      = {GitHub repository},
    publisher    = {GitHub},
    howpublished = {\url{https://github.com/huggingface/trl}}
}