shisa-chotto-lfm2-350m-sft

<!-- 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: 0.13.0.dev0

base_model: LiquidAI/LFM2-350M

chunked_cross_entropy: true

eot_tokens:
  - "<|im_end|>"
datasets:
  - path: chotto-20251007.sft.jsonl
    type: chat_template
    field_messages: conversations
    message_property_mappings:
      role: role
      content: content
    roles:
      system:
        - system
      assistant:
        - assistant
        - gpt
        - model
      user:
        - user
        - human
    roles_to_train: ["assistant"]
dataset_prepared_path: last_run_prepared
output_dir: ./outputs/out

sequence_len: 8192
sample_packing: true
flash_attention: true
pad_to_sequence_len: true

neftune_noise_alpha: 5

use_wandb: true
wandb_entity: augmxnt
wandb_project: lfm2
wandb_name: "LFM2 Test"

# GBS = 128 / 1 GPU / 32 MBS / 4 GAS
gradient_accumulation_steps: 4
micro_batch_size: 32
num_epochs: 3
optimizer: adamw_torch_4bit
lr_scheduler: cosine
learning_rate: 5e-5

train_on_inputs: false
group_by_length: false
bf16: true
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
logging_steps: 1

warmup_ratio: 0.03
saves_per_epoch: 1

deepspeed: zero3_bf16.json
# deepcompile: true
weight_decay: 1e-4


# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

</details><br>

outputs/out

This model is a fine-tuned version of LiquidAI/LFM2-350M on the chotto-20251007.sft.jsonl 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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH_4BIT 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: 124
• training_steps: 4137

Training results

Framework versions

• Transformers 4.56.1
• Pytorch 2.10.0.dev20251008+cu130
• Datasets 4.0.0
• Tokenizers 0.22.1