See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: true
base_model: EleutherAI/pythia-70m-deduped
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 3d9626b8bae4d161_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/3d9626b8bae4d161_train_data.json
  type:
    field_instruction: prompt
    field_output: chosen
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
do_eval: true
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 50
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: false
group_by_length: true
hub_model_id: lesso16/f2a0d099-c01e-41d1-81bf-74ea1dadc49c
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.000216
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 10
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_grad_norm: 1.0
max_steps: 500
micro_batch_size: 4
mlflow_experiment_name: /tmp/3d9626b8bae4d161_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 1
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 50
saves_per_epoch: null
seed: 160
sequence_len: 512
special_tokens:
  pad_token: <|endoftext|>
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: a5b36458-e110-4d1e-bf33-31b108030990
wandb_project: 16a
wandb_run: your_name
wandb_runid: a5b36458-e110-4d1e-bf33-31b108030990
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

f2a0d099-c01e-41d1-81bf-74ea1dadc49c

This model is a fine-tuned version of EleutherAI/pythia-70m-deduped on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.2405

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: 0.000216
• train_batch_size: 4
• eval_batch_size: 4
• seed: 160
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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: 50
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0008	1	9.0526
41.5982	0.0403	50	7.6238
22.4897	0.0807	100	6.7522
18.4772	0.1210	150	4.9873
12.1999	0.1614	200	4.9691
10.8443	0.2017	250	4.3068
11.2256	0.2420	300	4.6625
11.3407	0.2824	350	4.1743
11.1298	0.3227	400	4.1506
10.6977	0.3630	450	4.2888
10.4879	0.4034	500	4.2405

Framework versions

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