See axolotl config

axolotl version: 0.6.0

base_model: unsloth/SmolLM-360M
batch_size: 92
bf16: true
chat_template: tokenizer_default_fallback_alpaca
datasets:
- format: custom
  path: argilla/databricks-dolly-15k-curated-en
  type:
    field_input: original-instruction
    field_instruction: original-instruction
    field_output: original-response
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
device_map: auto
eval_sample_packing: false
eval_steps: 20
flash_attention: true
gradient_checkpointing: true
group_by_length: true
hub_model_id: SystemAdmin123/SmolLM-360M
hub_strategy: checkpoint
learning_rate: 0.0002
logging_steps: 10
lr_scheduler: cosine
max_steps: 10000
micro_batch_size: 23
model_type: AutoModelForCausalLM
num_epochs: 100
optimizer: adamw_bnb_8bit
output_dir: /root/.sn56/axolotl/tmp/SmolLM-360M
pad_to_sequence_len: true
resize_token_embeddings_to_32x: false
sample_packing: true
save_steps: 20
save_total_limit: 1
sequence_len: 2048
tokenizer_type: GPT2TokenizerFast
torch_dtype: bf16
training_args_kwargs:
  hub_private_repo: true
trust_remote_code: true
val_set_size: 0.1
wandb_entity: ''
wandb_mode: online
wandb_name: unsloth/SmolLM-360M-argilla/databricks-dolly-15k-curated-en
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: default
warmup_ratio: 0.05

SmolLM-360M

This model is a fine-tuned version of unsloth/SmolLM-360M on the argilla/databricks-dolly-15k-curated-en dataset. It achieves the following results on the evaluation set:

• Loss: 2.0673

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: 23
• eval_batch_size: 23
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 92
• total_eval_batch_size: 92
• 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: 10
• training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.125	1	2.5584
2.2406	2.5	20	2.1562
2.136	5.0	40	2.0829
2.0938	7.5	60	2.0711
2.0632	10.0	80	2.0679
2.0298	12.5	100	2.0621
2.0168	15.0	120	2.0567
2.0188	17.5	140	2.0686
2.0108	20.0	160	2.0701
2.0169	22.5	180	2.0683
2.0109	25.0	200	2.0673

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0