See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/SmolLM2-135M
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 51027af41b24ecea_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/51027af41b24ecea_train_data.json
  type:
    field_instruction: bio
    field_output: reply
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
do_eval: true
early_stopping_patience: 3
eval_batch_size: 4
eval_max_new_tokens: 128
eval_steps: 500
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: true
hub_model_id: lesso05/ac7ffea7-4dad-4fe8-86c5-640a3e80935a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.000205
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 50
lora_alpha: 128
lora_dropout: 0.15
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 5000
micro_batch_size: 4
mlflow_experiment_name: /tmp/51027af41b24ecea_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
optimizer: adamw_torch_fused
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 500
saves_per_epoch: null
seed: 50
sequence_len: 1024
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 893384c3-fe6e-4406-a06d-f7b0b65693ce
wandb_project: 05a
wandb_run: your_name
wandb_runid: 893384c3-fe6e-4406-a06d-f7b0b65693ce
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

ac7ffea7-4dad-4fe8-86c5-640a3e80935a

This model is a fine-tuned version of unsloth/SmolLM2-135M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.1982

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: 0.000205
• train_batch_size: 4
• eval_batch_size: 4
• seed: 50
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 100
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0002	1	3.4703
3.2876	0.1208	500	3.2851
3.264	0.2417	1000	3.2576
3.2494	0.3625	1500	3.2411
3.2305	0.4834	2000	3.2264
3.2099	0.6042	2500	3.2156
3.2234	0.7251	3000	3.2079
3.2001	0.8459	3500	3.2008
3.1994	0.9668	4000	3.2003
3.1712	1.0876	4500	3.1987
3.1854	1.2085	5000	3.1982

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1