See axolotl config

axolotl version: 0.4.1

adapter: qlora
base_model: unsloth/SmolLM-135M
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - d3e5bdd73347c26c_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/d3e5bdd73347c26c_train_data.json
  type:
    field_input: system_prompt
    field_instruction: problem
    field_output: solution
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: true
group_by_length: true
hub_model_id: error577/c06956a3-fe98-497b-b43e-8ab5d5a0d214
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
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
micro_batch_size: 3
mlflow_experiment_name: /tmp/d3e5bdd73347c26c_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_torch_fused
output_dir: miner_id_24
pad_to_sequence_len: true
restore_best_weights: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.005
wandb_entity: null
wandb_mode: online
wandb_name: 6f6675a6-cb87-49fd-824d-f8d0dfce8107
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 6f6675a6-cb87-49fd-824d-f8d0dfce8107
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

c06956a3-fe98-497b-b43e-8ab5d5a0d214

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

• Loss: 0.9896

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: 3
• eval_batch_size: 3
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 48
• 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: 10
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.2311	0.0010	1	1.7154
2.1415	0.0971	100	1.2307
1.9643	0.1941	200	1.1575
1.9656	0.2912	300	1.1358
1.8197	0.3883	400	1.1138
1.9307	0.4853	500	1.0955
1.7157	0.5824	600	1.0814
1.8345	0.6794	700	1.0735
1.8775	0.7765	800	1.0685
1.687	0.8736	900	1.0693
1.8377	0.9706	1000	1.0511
0.9808	1.0677	1100	1.0321
1.0324	1.1648	1200	1.0326
1.0425	1.2618	1300	1.0186
0.969	1.3589	1400	1.0205
0.9458	1.4560	1500	1.0165
0.9246	1.5530	1600	1.0105
0.8464	1.6501	1700	1.0113
0.9311	1.7471	1800	1.0011
0.9494	1.8442	1900	1.0005
0.879	1.9413	2000	1.0046
1.0447	2.0383	2100	0.9988
1.0148	2.1354	2200	0.9965
1.0102	2.2325	2300	0.9945
1.0027	2.3295	2400	0.9951
1.2094	2.4266	2500	0.9949
0.9836	2.5237	2600	0.9879
0.9018	2.6207	2700	0.9893
1.0633	2.7178	2800	0.9944
0.9116	2.8149	2900	0.9896

Framework versions

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