See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: fxmarty/tiny-llama-fast-tokenizer
bf16: true
chat_template: llama3
datasets:
- data_files:
  - b0e38ca3be1f31fc_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/b0e38ca3be1f31fc_train_data.json
  type:
    field_instruction: Question
    field_output: Hints
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: false
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: false
hub_model_id: lesso11/35835bce-bfe6-4d5d-b612-86073eb06e46
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: cosine
max_memory:
  0: 77GiB
max_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/b0e38ca3be1f31fc_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 25
save_strategy: steps
sequence_len: 1024
special_tokens:
  pad_token: </s>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 35835bce-bfe6-4d5d-b612-86073eb06e46
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 35835bce-bfe6-4d5d-b612-86073eb06e46
warmup_steps: 10
weight_decay: 0.01
xformers_attention: false

35835bce-bfe6-4d5d-b612-86073eb06e46

This model is a fine-tuned version of fxmarty/tiny-llama-fast-tokenizer on the None dataset. It achieves the following results on the evaluation set:

• Loss: 10.2993

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss
10.3528	0.0000	1	10.3501
10.3473	0.0002	9	10.3472
10.3424	0.0003	18	10.3402
10.3326	0.0005	27	10.3326
10.3244	0.0007	36	10.3249
10.3156	0.0008	45	10.3175
10.3108	0.0010	54	10.3109
10.3026	0.0012	63	10.3056
10.3006	0.0013	72	10.3021
10.2972	0.0015	81	10.3002
10.2971	0.0017	90	10.2995
10.2984	0.0019	99	10.2993

Framework versions

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